Mathematics of Ideas
Uploaded by
WillMathematics of Ideas
Uploaded by
WillMath Without Numbers
The Mathematics of Ideas
Vol. 1 Foundations
William S. Veatch
Copyright 2016, by William S. Veatch.
All Rights Reserved. Moral rights asserted.
ISBN-13: 978-1539503392
ISBN-10: 1539503399
DEDICATION
This book is dedicated to the memory of my parents, Richard and Lorraine,
and my brother David; and to my wife Debbie, and my three children
Christina, Will, and Margaret.
CONTENTS
Preface xvii
VOLUME 1 FOUNDATIONS
1. Overview of the Mathematics of Ideas 1
1.1. Why study the Mathematics of Ideas? 2
1.2. What is an Idea? 3
1.3. What is the Mathematics of Ideas? 5
1.4. Foundations: Atoms, Compound Ideas, and Power 9
Sets
1.5. Critical Thinking and Logic: Propositions, Syllogisms, 11
and the Logic of Lattices
1.6. Applied Math Without Numbers (MWN): Arguments, 18
Debates, and Policy Making
1.7. Update to the Definition of Idea 20
1.8. References, Historical Notes, and Further Reading 20
2. How Do We Know What an Idea Means or Represents? The 22
Dual Nature of Ideas as Sets of Objects and Sets of Attributes
2.1. An Object is an Example of an Idea 24
2.2. An Attribute is a Property or Characteristic that We 27
Use to Identify an Idea
2.3. Objects and Attributes are Inversely Related 30
2.4. Using Power Sets of Atoms and Coatoms to Represent 32
Objects and Attributes
2.5. Cover Plus a Difference: A Third Way of Defining an 45
Idea
WILLIAM S. VEATCH
2.6. The Importance of Specifying the Domain 46
2.7. Working with Sets of Atoms (Coatoms) v. Working 47
with Power Sets of Atoms (Coatoms)
2.8. Three Worlds: Physical v. Ideas v. Abstract Sets 50
2.9. Update to the Definition of Idea 51
2.10. References, Historical Notes, and Further Reading 52
3. Set Theory as Applied to Sets of Ideas 54
3.1. The Definition of a Set 55
3.2. The Universe of Ideas or Universal Set 59
3.3. The Empty Set 61
3.4. Finite v. Infinite 62
3.5. Discrete v. Continuous 62
3.6. Subdivision v. Consolidation 63
3.7. Atoms v. Power Sets 66
3.8. Simplifying Assumption: Finite Universe of Discourse 69
with Discrete Atoms
3.9. Update to the Definition of Idea 70
3.10. References, Historical Notes, and Further Reading 70
4. How Object Sets of Ideas are Related to Each Other 71
4.1. Ideas are Either Atoms or Compounds Made Up of 71
Multiple Atoms
4.2. Object Power Set: All Possible Combinations of 73
Atoms - Forms a Lattice Ordered by Inclusion
4.3. The Five Ways in Which Two Object Sets May be 74
Related to Each Other
vi
MATH WITHOUT NUMBERS
4.3.1. Identical Sets: the Same Set of Objects (and 77
Attributes)
4.3.2. Disjoint Sets 78
4.3.3. Partially Overlapping Sets 79
4.3.4. Subset-Superset and Superset- Subset 80
4.4. The Five Principal Set Operations Applied to Object 81
Sets: Union, Intersection, Symmetric Difference, Set
Subtraction, and Complement
4.4.1. Union 81
4.4.2. Intersection 81
4.4.3. Symmetric Difference 83
4.4.4. Set Subtraction 83
4.4.5. Complement 84
4.4.5.1. Absolute Complement 85
4.4.5.2. Relative Complement 85
4.5. The Distributive Property of Union and Intersection 85
4.6. Update to the Definition of Idea 87
4.7. References, Historical Notes, and Further Reading 87
5. How Attribute Sets of Ideas are Related to Each Other 88
5.1. Definition of a Coatom 88
5.2. Tracking Attributes in a Universe of Discourse 90
5.3. Attributes are Either Coatoms or Compounds Made 93
Up of Multiple Coatoms
5.4. Attribute Power Set: All Possible Combinations of 93
Coatoms - Form a Lattice Ordered by Inclusion
vii
WILLIAM S. VEATCH
5.5. The Five Ways in Which Two Attribute Sets May be 94
Related to Each Other
5.5.1. Identical Sets: The Same Set of Attributes (and 94
Objects)
5.5.2. Disjoint Sets 94
5.5.3. Partially Overlapping Sets 95
5.5.4. Subset-Superset and Superset-Subset 99
5.6. The Five Principal Set Operations Applied to Attribute 101
Sets: Union, Intersection, Symmetric Difference, Set
Subtraction, and Complement
5.6.1. Union 102
5.6.2. Intersection 102
5.6.3. Symmetric Difference 103
5.6.4. Set Subtraction 104
5.6.5. Complement 104
5.6.5.1. Absolute Complement 105
5.6.5.2. Relative Complement 105
5.7. How Many Attributes Can an Idea Have? 105
5.8. Update to the Definition of Idea 108
5.9. References, Historical Notes, and Further Reading 109
6. Summary of the Ways In Which Objects and Attributes are 110
Similar to or Different from Each Other
6.1. Similarities Between Objects and Attributes 110
6.1.1. An Idea Can Be Described Completely By a List 110
of Objects (or Attributes)
6.1.2. Each of Objects and Attributes Can Constitute 111
viii
MATH WITHOUT NUMBERS
the Elements of Sets
6.1.3. The Five Possible Set Relationships Between any 112
Two Sets Govern the Relationships Among
Objects As Well As Attributes
6.1.4. The Five Principal Set Operations Apply to Both 112
Sets of Objects and Sets of Attributes
6.1.5. Specifying An Object Set or An Attribute Set 112
Creates a Partition of the Universe of Ideas
6.1.6. One Fully Determines the Other: the Idea 113
Signature
6.1.7. Objects and Attributes in the World of Ideas 113
Map to Power Sets of Atoms and Coatoms in the
World of Abstract Sets
6.2. Differences Between Objects and Attributes 114
6.2.1. Examples v. Properties 114
6.2.2. Equivalence Classes v. Partitions 114
6.2.3. Inverse Relationship 115
6.3. Duality 116
6.4. Do We Need Both Objects and Attributes? 119
6.5. Update to the Definition of Idea 120
6.6. References, Historical Notes, and Further Reading 120
7. A More Detailed Look at Power Sets: the Building Blocks of 122
Ideas
7.1. Why study Power Sets? 123
7.2. What is a Power Set? 124
7.3. How Power Sets Grow in Size 126
ix
WILLIAM S. VEATCH
7.4. Combinatorics of Power Sets the Binomial Theorem 127
7.5. A Brief Look at Notation 129
7.6. Boolean Algebras 131
7.7. Boolean Rings 132
7.8. General Methodology for Performing Operations on 134
Power Sets
7.9. Adding Power Sets 135
7.10. Multiplying Power Sets 140
7.11. Power Set Complements 142
7.11.1. Power Set Complement: Domain = Power Set 143
P(X)
7.11.2. Power Set Complement: Domain = Power Set 145
of Power Sets P(P(X))
7.12. Power Sets Viewed as a Sum of Chains; or a Sum of 149
Antichains/Partitions
7.13. Techniques to Simplify Working with Power Sets 151
7.14. Power Set Expansion and Contraction of a Set of 151
Atoms
7.15. Partition Equations 153
7.16. Nested Partition Equations 154
7.17. The Substitution Principle 157
7.18. Selecting the Order of the Nested Partition Sets 159
7.19. Multi-Dimensional Power Sets 164
7.20. Update to the Definition of Idea 166
7.21. References, Historical Notes, and Further Reading 167
x
MATH WITHOUT NUMBERS
8. Chains: Ideas that Form Ordered Sets of Attributes 169
8.1. Creating a Valid Chain; Examples of Chains 169
8.2. Each Attribute of an Idea Can Be Interpreted as a 170
Level in a Chain
8.3. Level of Detail in a Chain 172
8.4. Creating a Chain of Objects 173
8.5. Creating a Dual Chain of Attributes; Inheritance of 176
Attributes
8.6. Using a Chain to Sort Attributes from the World of 177
Ideas
8.7. Impact of Changes in Level of Detail on the Sorting of 178
Attributes
8.8. Partition Sets in a Partition Equation Create Levels in a 179
Chain; Datasets
8.9. Comparing Chains in the Physical World, the World of 181
Ideas, and the World of Abstract Sets
8.10. Three Principal Types of Chains: Classification, Value 183
System, and Action
8.11. Simplifying Assumption: Using a Five Level Chain 186
8.12. Displaying a Chain as a Nested Lattice 186
8.13. Update to the Definition of Idea 189
8.14. References, Historical Notes, and Further Reading 189
9. Partition Sets / Antichains: Ideas that Form Disjoint Sets of 190
Objects
9.1. Creating a valid Partition Set; Examples of Partitions 191
9.2. Each Object in An Idea List Set Can Be Interpreted as 193
an Equivalence Class in a Partition or Antichain
xi
WILLIAM S. VEATCH
9.3. Level of Detail in a Partition 193
9.4. Creating a Partition (or Antichain) of Attributes 195
9.5. Using a Partition Set to Sort Objects from the World 197
of Ideas
9.6. Impact of Changes in Level of Detail on the Sorting of 200
Objects
9.7. Partition Sets in a Partition Equation Create Levels in a 200
Chain; Datasets
9.8. Displaying a Partition as a Nested Lattice 200
9.9. Update to the Definition of Idea 200
9.10. References, Historical Notes, and Further Reading 201
10. Partition Equations: A Practical Tool for Organizing Ideas 202
10.1. Definition of a Partition Equation 202
10.2. Creating Nested Partition Equations with Two 204
Partition Sets
10.3. Creating Nested Partition Equations with Three 209
Partition Sets
10.4. Adding a Partition Set Creates a New Level in a Chain 212
10.5. Lattice Representations of a Nested Partition Equation 216
10.6. The Sorting Room: a Laboratory for Experiments 218
with Ideas
10.7. Update to the Definition of Idea 222
10.8. References, Historical Notes, and Further Reading 222
11. Universes of Discourse 223
11.1. What is a Universe of Discourse? 224
11.2. How Do We Combine Two or More Universes of 224
xii
MATH WITHOUT NUMBERS
Discourse?
11.3. Combining Universes of Discourse with One and 225
Three Atoms, Respectively
11.3.1. The Same Identical Sets of Atoms 226
11.3.2. Disjoint Sets of Atoms 226
11.3.3. Subset-Superset of Atoms 228
11.3.4. Superset-Subset of Atoms 230
11.3.5. Partially Overlapping (POL) Sets of Atoms 231
11.4. How Do We Combine More Complicated Universes 235
of Discourse?
11.4.1. The Same, Identical Universes of Discourse 235
11.4.2. Disjoint Universes of Discourse 237
11.4.3. Subset-Superset Universes of Discourse 240
11.4.4. Superset-Subset Universes of Discourse 242
11.4.5. Partially Overlapping (POL) Universes of 245
Discourse
11.5. Summary of the Rules for Combining Universes of 245
Discourse
11.6. Update to the Definition of Idea 247
11.7. References, Historical Notes, and Further Reading 247
12. Axioms of Set Theory 248
12.1. What are the Differences Between Traditional Set 249
Theory and MWN?
12.2. The Is a Relation: Belongs to ( ) or Is a Subset 252
of ( )?
12.3. The Importance of Specifying the Domain 254
xiii
WILLIAM S. VEATCH
12.4. Axioms of Set Theory 256
12.5. Domain : Atoms(X); Coatoms (Y) 261
12.5.1. The Universal Set 261
12.5.2. Subdivision and Consolidation 261
12.5.3. Partitions and Equivalence Classes 264
12.5.4. Coatoms 264
12.6. Domain: Power Set of Atoms P(X); Power Set of 266
Coatoms P(Y)
12.6.1. The Universal Set and the Empty Set in the 266
Power Set Domain
12.6.2. Subdivision and Consolidation in the Power Set 267
Domain
12.6.3. Set Operations in the Power Set Domain 269
12.6.4. Coatoms in the Power Set Domain 270
12.7. Domain: Power Set of Power Sets of Atoms P(P(X)); 272
Power Set Power Sets of Coatoms P(P(Y))
12.7.1. The Universal Set and the Empty Set in the 272
P(P(X)) Domain
12.7.2. Subdivision and Consolidation in the P(P(X)) 272
Domain
12.7.3. Set Operations in the P(P(X)) Domain 273
12.7.4. Coatoms in the P(P(X)) Domain 273
12.7.5. Converting Sets and Partition Equations from 273
One Domain to Another Within the World of
Abstract Sets
12.8. Mappings Across the Three Worlds: Physical, Ideas, 274
and Abstract Sets
xiv
MATH WITHOUT NUMBERS
12.9. What Exactly is an Object? 276
12.10. What Exactly is an Attribute?; Demystifying the 276
Concept of a Property
12.11. Russells Paradox 277
12.12. What is a Number? 281
12.13. Can a Number Also Be an Attribute? 283
12.14. Update to the Definition of Idea 285
12.15. References, Historical Notes, and Further Reading 285
13. Conclusion 287
13.1. Goals and Objectives in Developing a Mathematics of 287
Ideas
13.2. Did We Succeed in Developing a Mathematics of 288
Ideas?
13.3. Next Steps: Where Do We Go From Here? 290
Appendix A Definition of Idea 292
Appendix B - How the Mathematics of Ideas Relates to the 298
Study of Abstract Algebra and Other Fields of Mathematics
Bibliography 320
Index 330
xv
WILLIAM S. VEATCH
xvi
PREFACE
This Volume 1 Foundations is the first in a series of books that discuss
different aspects of the Mathematics of Ideas. The goal of this series of
books is to provide an overview that, hopefully, is accessible to everyone,
particularly those who have no formal training in mathematics beyond high
school math. Admittedly, some of the mathematics is complicated in places,
but the intent is that those parts could be skimmed over without losing the
overall message of the Book. The remaining books will explore the
Mathematics of Ideas in more detail, and can be used as a reference source if
the reader wants to explore particular areas in more detail.
Book One of the Math Without Numbers series comes in three
Volumes:
Volume 1 Foundations (i.e., this book),
Volume 2 Critical Thinking and Logic; the Logic of Lattices
(expected in 2017), and
Volume 3 Applied Math Without Numbers (MWN)
Argumentation, Debate, and Policy Making (expected in 2017).
The central theme of this Volume 1 is that we can apply the mathematics
of Set Theory to explain how we think; we cannot, however, use
mathematics to determine what to think. We use Set Theory to explain
not only the form in which we make valid arguments, which we refer to as
Logic, but also to explain the way in which we create, store, and retrieve
knowledge, which we refer to as Knowledge Representation. As a result,
our discussion of Math Without Numbers, or MWN for short, is much
more than just a study of Logic, which is an enormous undertaking in itself,
but it also provides a methodology for us to map the knowledge of the World
of Ideas in a systematic way. At least, we show how such a mapping could
be accomplished.
Another goal is to develop a number of practical tools that we can use in
any subject area, to improve the clarity of our thinking and communication,
particularly in areas like education, translation to and from a foreign language,
or computer science and artificial intelligence. The tools that we will develop
include:
Sets of Ideas,
Power Sets,
Chains,
Partitions,
xvii
WILLIAM S. VEATCH
Nested Partition Equations,
Hierarchy Trees, and
Lattices.
Chapter One of this book provides an overview of what we hope to
accomplish in the Math Without Numbers series, including Volume 1: the
foundations in Set Theory of the Mathematics of Ideas; Volume 2: critical
thinking and logic, including the logic of lattices, and Volume 3:
argumentation, debate, and policy making. (Note: this book is Volume 1;
Volumes 2 and 3 are expected later in 2017.)
In Chapter Two, we explore how we think of Ideas at an intuitive or
philosophical level as Objects and Attributes, and then translate that intuitive
sense into the mathematical concepts of dual Sets and Power Sets of Atoms
and Coatoms.
In Chapter Three, we explore the notion of Ideas as Sets in more
detail, given that this is central to our entire Mathematics of Ideas. We
explore a number of dual properties of ideas, including the Universal v.
Empty Set, Finite v. Infinite, Discrete v. Continuous, Subdivision v.
Consolidation, and Atoms v. Power Sets.
In Chapters Four, Five, and Six, we examine the five ways in which any
two sets may be related to one another (identical, disjoint, subset-superset,
superset-subset, and partially overlapping), and the five principal Operations
(union, intersection, symmetric difference, set subtraction, and complement).
The key point to note here is that the same five set relationships and five set
operations apply to both Sets of Idea-Objects and Sets of Idea-Attributes,
but the set operations for Objects and Attributes are inverse to one another.
In Chapter Seven, we look at the mathematics of Power Sets. Some
authors focus more on subsets of Power Sets, but we embrace Power Sets
full on, and develop a mathematics of Power Sets. In particular, we see that
we can view regular Sets and Power Sets as co-existing in parallel worlds or
domains. Power Sets are appealing in that they represent the full Universe
of Ideas; but calculations are easier to perform on regular Sets. The good
news is that we develop rules for moving back and forth between regular Sets
and Power Sets. In this way, we can simplify calculations involving Power
Sets by performing Operations on the related regular Sets, and then
converting the results to Power Sets.
In Chapter Eight, we look in more detail at Chains of Ideas, which will
become one of our primary tools for working with Ideas. Chains are
xviii
MATH WITHOUT NUMBERS
particularly important because an Idea inherits the Attributes of every Idea
above it in the Chain.
In Chapter Nine, we look at Partitions, which represent a breakdown of
an Idea into its component Object parts.
In Chapter Ten, we explore Partition Equations, which are a means of
combining Chains and Partitions in a practical manner that allows us to
organize our thoughts and analysis on practically any subject.
In Chapter Eleven examines Universes of Discourse, and in particular
how to combine two to more Universes of Discourse. While not a trivial
task, we see that we can combine Universes of Discourse by following a set
of specific rules. This means that we can study specific subsets of the
Universe of Ideas separately by creating smaller, discrete Universes of
Discourse, and then later combine the results. This provides the foundation
for creating a Knowledge Representation map of the entire Universe of Ideas.
In Chapter Twelve, we examine the Axioms of Classical Set Theory and
how they relate to our study of the Mathematics of Ideas. The principal
difference between Classical Set Theory and the Set Theory we develop to
support the Mathematics of Ideas, is that in the Mathematics of Ideas we
emphasize the dual Power Sets of Atoms and Coatoms. We can map Objects
to elements of Power Sets of Atoms, and Attributes to elements of Power
sets of Coatoms. The dual Power Sets work, in mathematical terms, to
represent how we think of Objects (Extension) and Attributes (Intension)
from a philosophical perspective.
In Chapter Thirteen, the Conclusion, we touch on where we have been,
and we look forward to Volumes 2 and 3 of Book One where we will explore
logic and critical thinking, and public policy argumentation and debate,
respectively.
In Appendix A, we review the definition of Idea as it has evolved over
the course of this book.
In Appendix B, we briefly consider the key areas of mathematics that
we have incorporated into our Mathematics of Ideas, including Group
Theory, Boolean Algebras, Ring Theory, Lattice Theory, Topology, and
Formal Concept Analysis. By demonstrating that Sets of Ideas form
mathematical structures called Groups, Boolean Algebras, Boolean Rings,
Boolean Lattices, and Topologies, we allow all of the knowledge and rules
from these areas of mathematics to be applied to Ideas, not just numbers.
xix
WILLIAM S. VEATCH
The three Volumes of Math Without Numbers cover a wide range of
topics, hopefully one or more of which will be of interest to any reader,
including:
Mathematics, particulary Set Theory,
Classification Theory,
Philosophy of Language,
Grammar and Linguistics,
Philosophy of Numbers,
Logic,
Computer Science and Artificial Intelligence,
Argumentation and Debate, and
Current Public Policy Issues.
For those with an interest in mathematics, Book Two (expected in 2018-
2019) will connect the World of Ideas to the World of Mathematics in more
detail. My professional training is as a lawyer, however, not a mathematician,
so I apologize in advance for any inaccuracies or errors when viewed through
the eyes of a true mathematician. Still, if we are to bridge the gap between
the World of Ideas and the World of Mathematics and numbers, then
someone must take the first step outside of his or her field of expertise and
risk offending the true experts.
Where we make statements without stating, or referring to, a formal
proof, we use terms such as Observation or Proposition, rather than
Theorem or Proof. Over time, the goal will be to convert Observations
and Propositions to Theorems and Proofs.
One way to use the books is to use this Book One as the main text for a
class, with an emphasis on topics of interest to the particular teacher or class
of students. For example, the focus could be on Logic, Abstract Algebra,
Formal Concept Analysis, Set Theory, analyzing Current Events, etc.
Hopefully, Math Without Numbers (MWN) techniques will prove to be
useful in a wide range of disciplines.
For those readers less interested in mathematics, it is okay to skip any
particular Chapter if the mathematics is too hard to follow. This Book One
is designed to allow the reader to take different paths, rather than having to
read from cover to cover. As a result, however, there is some repetition of
concepts from chapter to chapter.
xx
MATH WITHOUT NUMBERS
Anyone wishing to contact the author may send a message to
[email protected].
xxi
1. OVERVIEW OF THE MATHEMATICS OF IDEAS
In this first Chapter, we provide a high level overview of the path that we will
take on our journey to study and explore the Mathematics of Ideas. We will
demonstrate that how we think is governed by the rules of Mathematics. By
how, we mean both (i) the rules of Logic that determine how Ideas are
related to one another and how we can draw valid conclusions from true
premises using sound arguments, and (ii) Knowledge Representation, which
involves creating a map of the Universe of Ideas. Although the same rules
of thought apply to all Ideas, the subject matter of Ideas that we choose to
study initially breaks down into three main areas:
Classifications: descriptions and classifications of what we can see,
touch, experience, or measure in the Physical World, whether
tangible or intangible, including cause and effect as determined
through scientific experimentation,
Value Systems: a list and categorization of values, morals, and
beliefs that guide us when making judgments about how we as
humans should behave, and
Actions: an analysis of proposed actions that we can take, and how
to determine if the proposed actions are consistent with our Value
System. This will be our main focus later in Volume 3 of the series
when we examine the public policy issues raised in the 2016
Presidential election campaign in the United States, although the
issues themselves are timeless.
We will see that the rules of Set Theory and Mathematics work well to
describe how we think in all of these areas.
1
WILLIAM S. VEATCH
On the other hand, what we think about or what we decide to do is not
determined by Mathematics. To think clearly and communicate effectively,
however, it helps to understand the mathematical rules that govern how we
think. By the rules of Mathematics, we mean primarily Set Theory. For those
readers who have a deeper interest in Mathematics, we will discuss the
Mathematics of Power Sets, including Group Theory, Boolean Algebras,
Ring Theory, Lattice Theory, Topology, and Formal Concept Analysis. Do
not panic if you are not familiar with these areas of Mathematics. The point
here is simply that once we demonstrate that Sets of Ideas do in fact form
Groups, Boolean Algebras, Rings, Lattices, Topologies, and Formal
Concepts, then the entire wealth of knowledge in these areas of Mathematics
can be brought to bear on the World of Ideas. We will explain step-by-step
what you need to know in order to understand how these areas of
Mathematics impact our study of Ideas.
Note, however, that our Mathematics of Ideas is focused on Ideas
generally, rather than numbers. As a result, we sometimes refer to our field
of study as Math Without Numbers, or MWN for short.
1.1. Why study the Mathematics of Ideas?
We will revisit this question at the end of the Book, but our hope is to show
that by studying the Mathematics of Ideas, we can communicate more clearly
and with greater precision of meaning. If successful, the benefits are far
reaching, and could be applied in virtually any field of study. A few examples
of the applications include:
Education (in virtually any field of study),
Lexicography (the study of writing, editing, and compiling
dictionaries),
Translation to and from a foreign language,
Computer science and artificial intelligence,
Analysis of data collected over the internet,
Logic, and
Argumentation and debate.
In addition to the benefits to communication generally, we explore the
concept of Knowledge Representation, including the possibility of
mapping the knowledge base of mankind in a mathematically precise, on-line
dictionary or encyclopedia. Admittedly, this is a lofty goal, but as with any
great journey, we must start with a single step!
2
MATH WITHOUT NUMBERS
1.2. What is an Idea?
Lets start with a definition of what we mean by an Idea, and then illustrate
how the Mathematics of Ideas would be extremely helpful to us in many
walks of life. We start with the following characteristics of Ideas, taken from
various general definitions of the word Idea:
any picture, thought, conception, or notion;
of anything;
conceived, developed, or existing in the mind;
as a result of mental understanding, awareness, or activity;
including:
o an observation, or description;
o an impression;
o an opinion, view, or belief;
o a plan of action; an intention;
o a groundless supposition;
o a fantasy; and
o a conception of what is desirable or ought to be; what
is ideal.
Although there are many possible definitions of Idea to choose from,
we choose the following as the most suitable definition for our purposes: any
thought that can be conceived of in the mind. Throughout the course of this
book, however, we will come to learn in greater detail what we mean by an
Idea, how Ideas are related to one another, and how Ideas behave in different
circumstances. By the end of the Book, we will have a much different and
more precise definition and understanding of what an Idea is.
Definition of Idea: Any thought that can be conceived of in the mind.
Ideas distinguished from Words
Note that we speak of Ideas, not Words. The reason for this is that
we use words to attempt to capture the essence of an Idea, but we often fail
miserably. For example, an Idea may not be capable of expression in a single
word. In fact, sometimes it takes a sentence, a paragraph, or perhaps even
3
WILLIAM S. VEATCH
an entire book to convey a single Idea. In this sense, a new, complex Idea
may not have a single word associated with it.
Another issue with words, is that words are ambiguous, that is, the same
word is often used to mean several different things. This can be easily
illustrated by looking up words in a dictionary, and noting the numerous
definitions and meanings of a single word. For example, when hiking in the
woods, a sign at a fork in the path reads: Bear to the right. Does this sign
mean Keep to the right, or Watch out, grizzly bears have been spotted to
the right? Here, the word bear is ambiguous. See [McInerny 2004] at 31.
Ideas on the other hand, as we will soon see, are unique and
unambiguous if used properly. The ambiguity is in the words that we use to
describe Ideas, not in the Ideas themselves.
Sometimes, two words can have the same meaning. One example of
this is when we use a thesaurus to search for words with the same meaning
in order to add variety, diversity, or variation to our writing (although in fact
the Words listed as synonyms usually have slightly different meanings). A
better example might be words in different languages that have the same
meaning, and therefore represent the same Idea. The word cat in English
has the same meaning as chat in French. This illustrates the point that two
different words may illustrate the same picture or notion conceived of in
the mind, or Idea.
Another example of the limitation of words arises when we study a
foreign language and are dismayed to learn that there is no exact translation
of a particular foreign word into English. (Language professors always seem
to take delight when this happens, as it demonstrates the beauty and
importance of the foreign language, and the limitations of English!)
Conceived of in the mind as distinguished from appearing in the
real world
When we speak of Ideas, we mean as conceived of in the mind, such
that the Idea may or may not have any resemblance to anything in the real
world. As children, we start by learning words that relate to things that we
that we see around us. We learn the colors of the rainbow:
Rainbow Colors = { red, orange, yellow, green, blue, indigo, violet};
4
MATH WITHOUT NUMBERS
and we learn to identify farm animals:
Farm Animals = { horse, pig, sheep, cow }.
In our minds eye, however, we can see the Idea of a purple cow, even
though no such thing exists in the real world. We start learning words by
associating them with real world objects, so that we can communicate with
one another by using the same word to refer to the same object. Once the
meaning of a word has been established, however, it is released from the
shackles of its origin and free to combine with other words in new and
creative ways. Our imagination is in no way limited by our real world
experiences.
Ideas are defined in terms of other Ideas
Typically, we define Ideas in terms of other Ideas. Indeed, if the only
way to define an object was to point at it and say its name, we would be very
limited in our vocabulary. If we wanted to convey the Idea of a horse, we
would either have to find a horse, find a picture of a horse, or we would have
to improve our drawing skills. In many cases, however, an Idea is an abstract
concept that has no picture associated with it. For example, concepts like
love, hate, friendship, fairness, justice, etc., do not have a single picture
associated with them.
As a practical matter, we must rely upon words, along with their inherent
limitations, to define Ideas. Hence, we turn to the dictionary for definitions
of Ideas expressed in words that in turn are defined by other related Ideas.
1.3. What is the Mathematics of Ideas?
Now we turn to examining how the Mathematics of Ideas could be useful to
us, but first we should define what we mean by Mathematics. Once again
turning to the dictionary, we find the following general definition:
the abstract deductive science,
of space, number, quantity, and arrangement,
studied in its own right;
including, geometry, arithmetic, algebra, abstract algebra (including
group theory and ring theory), trigonometry, calculus, set theory,
topology, etc.
5
WILLIAM S. VEATCH
For the moment, we must accept on faith that Ideas can be described, as least
in part, by familiar concepts of geometry, arithmetic, and algebra. In fact,
one of the goals of this book is to show, or at least begin to show, the extent
to which the familiar rules of the mathematics of numbers apply to Ideas
generally. (After all, one interpretation of numbers is that numbers
themselves are simply an Idea, governed by the same rules as all other Ideas.)
Also, the use of etc. saves us here, and allows us to define a new branch of
mathematics relating specifically to the world of Ideas, in case we object to
saying that Ideas are merely a part of geometry, arithmetic, or algebra.
"Studied in its own right"
We can study the Mathematics of Ideas in its own right, regardless of
what we do in our personal or professional life, because it helps us
understand the way we think, and organize our Ideas in a systematic way.
This is not to say that the Mathematics of Ideas cannot be applied to other
subject areas. To the contrary, every subject area, particularly those that do
not involve counting numbers, could benefit from, and should be reanalyzed
and studied using, the principles of the Mathematics of Ideas.
At this point, it may seem like a bit of a stretch to say that Ideas are
governed by a branch of mathematics, but that view may change as you
proceed through this book. Some readers, particularly those who did not
enjoy math in school, may be disturbed by the concept of turning the world
of Ideas into a branch of mathematics. The good news for those readers,
however, is that by mathematics we mean an abstract deductive science,
but not one that is limited to numbers in the sense that we learned in grade
school. In fact, it will become apparent that numbers in the sense of
counting numbers are actually a small subset of the vast Universe of Ideas.
Rather than the study of Ideas being a branch of mathematics, mathematics
can be viewed as a branch of the study of the much broader world of abstract
Ideas.
As illustrated in the following Table, we break down our study of Math
Without Numbers (MWN for short) into three broad topics: Foundations,
Critical Thinking and Logic, and Applied MWN.
6
MATH WITHOUT NUMBERS
At the first stage, Foundations, we study Atoms, Compound Ideas, and
Dual Power Sets. Atoms are the simplest Ideas that constitute the building
blocks from which we build more complex Compound Ideas. Compound
Ideas are Sets made up of Atoms. For example, if Dog and Cat are
Atoms, then Mammal would be a Compound Idea that includes Dogs and
Cats. In fact, every Idea is either an Atom or a Compound Idea. We
introduce the concept of a Power Set next, which is defined as the Set of all
subsets of a set of Atoms. The Power Set will prove to be of critical
importance, because once we specify a Set of Atoms, the Power Set of those
Atoms will contain all possible Compound Ideas that can be made from those
Atoms. (As we will soon see, there are in fact two types of Power Sets that
are linked together in an inverse relationship: Power Sets of Objects and
Power Sets of Attributes - more on this later.)
At the second stage, Critical Thinking and Logic, we study Logic and
Knowledge Representation, including Propositions, Syllogisms, and the
Logic of Lattices. A Proposition is a special kind of sentence which is either
true or false, and is made up of two Ideas, a Subject and a Predicate,
connected by a Copula, which is simply a form of the verb to be. For
example, Socrates is human is an example of a Proposition. Whereas a
Proposition is a combination of two Ideas, a Syllogism is a combination of
three Ideas using three Propositions, as illustrated in the following example
showing how the three Ideas in the Set X = { Socrates, humans, mortal } are
related:
All humans are mortal.
Socrates is human.
Therefore, Socrates is mortal.
7
WILLIAM S. VEATCH
Next, we will generalize the relationships of Ideas from two Ideas
(Proposition) and three Ideas (Syllogism), to all possible combinations of
Atomic Ideas (Lattice). See Fig. 1-1. The first diagram illustrates the
specification of a single Idea of Socrates. The second diagram illustrates a
Proposition: Socrates is human, consisting of two Ideas. The third diagram
illustrates the interaction of three Ideas in a Syllogism. The advantage to the
fourth and final Lattice diagram is that it shows all possible combinations of
our Idea Atoms: Socrates, Other Humans, Other Mortals, Other Ideas.
At the third stage, Applied MWN, we study Complex Arguments,
8
MATH WITHOUT NUMBERS
Debates, and Policy Making. In its simplest form, an Argument may be made
up of a single Syllogism. (We call a Syllogism an explanation if it is
generally agreed upon, and an argument if it is open to debate.) Often,
Arguments consist of a string of Syllogisms, often referred to as a Sorites,
which is the term that Aristotle used as far back as 322 BCE in ancient
Greece. We look at other forms of Arguments as well. In debating and
policy making, we have two or more competing Arguments, and we need to
decide which we believe to be the better Argument.
Now, we look at each of the three stages in a little more detail.
1.4. Foundations: Atoms, Compound Ideas, and Power Sets
In the Physical World, tangible objects are made up of atoms of elements set
forth in the Periodic Table of Elements; in our World of Ideas we have a
similar concept of Atoms. In the Physical World, an atom is the smallest
piece of matter that still exhibits all of the properties of the element in
question; and compounds are made up of multiple atoms of different types.
For example, hydrogen and oxygen are types of atoms, and together two
atoms of Hydrogen and one atom of oxygen form H2O or water. Similarly,
in the world of Ideas, we define an Atom to be an Idea that cannot be
broken down into smaller component Ideas. (Often, for convenience we will
deem Ideas to be Atoms, even though in theory they could be broken down
into smaller Ideas.) Compound Ideas are in turn composed of
combinations of Atoms. For example, if we assume that the Ideas in the Set
S = { Dog, Cat, Human, Cow, Horse} are Atoms, then together they form
the Compound Idea Mammal. Given a set of Atoms X = { a, b, c, },
we can form a Power Set which is defined as the Set of all subsets of the set
of Atoms, including the Empty Set denoted by the symbol . For
example, if X = { a, b }, then the Power Set of X = P(X) = (, a, b, ab ). If
X = { a, b, c }, then the Power Set of X = P(X) = ( ,a, b, c, ab, ac, bc, abc),
i.e., the Set of all possible Subsets. Power Sets will play an important role in
our Mathematics of Ideas. For now, just note that every possible Compound
Idea that can be formed from a particular Set of Atoms will appear
somewhere in the Power Set for that Set of Atoms. This is true because by
definition, every possible combination appears somewhere in the Power Set.
In fact, there are two related Power Sets, one for a Set of Atoms, and
one for a Set of Coatoms. See Fig. 1-2. A Coatom is defined as the Set
containing all Atoms except one. The Atoms are found in the second row
from the bottom of the Power Set of Atoms/Objects, just above the 0 or
Element; the Coatoms are found in the second row from the top, just
below the Universe Element. Notice how when we relabel the Coatoms from
9
WILLIAM S. VEATCH
ab, ac, bc to 3, 2, 1, the inverse (upside-down) Power Set structure
becomes readily apparent. These two Power Sets relate to the same Objects
or Atoms; we have simply relabeled the Coatoms to show the inverse Power
Set structure more clearly. We sometimes refer to the Power Set of Atoms
as an Object Power Set and the Power Set of Coatoms as an Attribute
Power Set, and together we may refer to them as Dual Object-Attribute
Power Sets. (As we will soon see, we can map all Attributes to the Attribute
Power Set, which explains why we give it that name.)
Whereas an Atom represents the presence of a single Object, a Coatom
represents the absence of a single Object. For example, if we take the Set
X Atoms = { a, b, c } (in the second row from the bottom of the Power Set
of Objects/Atoms), then the related Set of Coatoms would be X Coatoms =
{ ab, ac, bc } = { ~c, ~b, ~a } = { 3, 2, 1} (in the second row from the top
of the Power Set). Fig. 1-2 illustrates how the Power Set of Atoms is related
to the Power Set of Coatoms. Note that we have three equivalent ways of
representing Coatoms: (i) as all Atoms except one (e.g., ab), (ii) using the not
symbol ~ (e.g., ~c), or (iii) using numbers (e.g., 3). Later, we will
demonstrate how what we commonly think of as Objects (sometimes
referred to as the extension of an Idea) corresponds to the Power Set of
Atoms; and what we commonly think of as Attributes (sometimes referred
to as the comprehension or intension of an Idea) corresponds to the
Power Set of Coatoms.
The second and fourth Power Set diagrams in Fig. 1-2 illustrate the
inverse relationship between the Power Set of Objects (Atoms) and the
10
MATH WITHOUT NUMBERS
Power Set of Attributes (Coatoms); in the Power Set of Objects we
traditionally draw the 0 Element or at the bottom, in which case the
0 Element is at the top of the related Power Set of Attributes.
Note that there is a rich history of books on Logic and critical thinking
that discuss the dual nature of Ideas as Objects (extension) and Attributes
(intension). See the list of references at the end of this Chapter. Our
treatment here of Attributes as a dual Power Set to the Power Set of Objects,
and our emphasis on Power Sets generally, is, however, different from the
traditional view.
1.5. Critical Thinking and Logic: Propositions, Syllogisms, and the
Logic of Lattices
As mentioned above, a Proposition is a special kind of sentence which is
either true or false, and is made up of two Ideas, a Subject and a Predicate,
connected by a Copula, which is simply a form of the verb to be. We also
know from our prior discussion that we can express every Idea as a Set. A
Proposition, therefore, is really a sentence that is telling us the relationship
between two Ideas or Sets. As illustrated in Fig. 1-3 there are only five ways
in which two Sets can be related to one another. (See if you can find a sixth!)
Note that that the Superset / Subset and Subset / Superset relationships
are arguably the same, but books on Logic often treat them as separate on
the basis that a particular Idea, Animal in this case, may be the Superset in
one situation or the Subset in another.
11
WILLIAM S. VEATCH
From the Set relationships that are illustrated in Fig. 1-3, we can derive a
number of Propositions, including, among others, the following:
Red in English is Rouge in French.
The color Black is not a Dog.
Some Dogs are Black.
Some Dogs are not Black.
All Dogs are Animals.
All Animals are Living Things.
Note that all of the above examples of Propositions follow the basic
format of Subject + Copula (i.e., a form of the verb to be) + Predicate.
In addition, there are some key words not, some, and all that we will
discuss further in the Chapter on Logic. For now the point that we want to
make is that we can view Ideas as Sets, and Propositions are statements about
the relationship between two Sets.
The above Propositions are all true, but we can also create Propositions
that are false, such as the following:
All Dogs are Black.
Some Animals are not Living Things.
12
MATH WITHOUT NUMBERS
You might ask: why do we give examples of false Propositions? The
reason is that it is always important to understand whether Propositions are
true or false, or whether their truth is undetermined, when using them in a
logical proof. Otherwise, we may come to a false conclusion. For example,
if we assume that all immigrants are criminals is a true Proposition, then
we may come to a false conclusion if we use that Proposition as a Premise in
a Syllogism.
Whereas a Proposition is a combination of two Ideas, a Syllogism is a
combination of three Ideas using three Propositions, as illustrated in the
following example showing how the three Ideas { Socrates, Humans, Mortal}
are related:
All Humans are Mortal.
Socrates is Human.
Therefore, Socrates is Mortal.
Viewing each of the three Ideas as Sets, we can create the diagram in
Fig. 1-4 to illustrate the relationships among the three Sets.
The innermost circle in Fig. 1.4 encloses the Object a representing
Socrates; the middle circle encloses the Objects a and b representing
Humans; and the outermost encloses the Objects a, b, and c
representing Mortals/Living Things generally.
While diagrams such as the one in Fig. 1-4 are helpful to visualize Set
13
WILLIAM S. VEATCH
relationships, they become cumbersome to work with when we have more
than two Ideas. As we saw earlier, with two Ideas, there were only five
possible ways that the two Sets could be related to one another. With three
Sets, however, the number of possible relationships increases rapidly.
Assuming that all three Sets are different from each other, Fig. 1-5 uses what
are called Euler Diagrams to show 11 other additional ways in which three
Sets could possibly be related to one another.
14
MATH WITHOUT NUMBERS
To give some specific examples of how we can use the Euler Diagrams
to represent Syllogisms, we look at three sample Syllogisms labelled Cesare,
Camestres, and Festino. See, Table 1-4. If you have not studied Logic before,
then these names may sound strange to you, but in fact they have been used
15
WILLIAM S. VEATCH
for many, many years by logicians. (More on that in Volume 2.) For now,
the significance of the names is not important. Rather, we just want a basic
understanding of what a Syllogism is, and how we can interpret it as the
relationships among three Sets of Ideas.
Each of these three Syllogisms can be represented using a Euler
Diagram, as the relationships among three Sets. See Fig. 1-6.
Later, when we study Classical Logic in more detail, we will see that
because of certain ambiguities inherent in the way Propositions are
constructed, more than one Euler Diagram is possible for many Syllogisms,
further adding to the limitation of such circle diagrams. (For example, in the
16
MATH WITHOUT NUMBERS
Festino Syllogism above, it is unclear from the Syllogism itself whether some
politicians are criminals or no politicians are criminals. The Euler Diagram
suggests that no politicians are criminals, but it is possible to draw the Euler
Diagram for the Festino Syllogism in other ways.) As we can see, while
helpful for comparing two or three Sets, the circle diagrams are difficult
enough with three Idea Sets, and will quickly become unmanageable as we
start to work with more than three Ideas.
We will use such circle diagrams extensively in connection with our
study of Syllogisms, but we need a more manageable way to represent
combinations of more than three Sets. Our solution to the problem is to use
Lattices, along with some special rules about how to simplify Lattice
diagrams.
With Lattices, we can generalize the relationships of Ideas from two
Ideas (Proposition) and three Ideas (Syllogism), to all possible combinations
of four, five, or more Atomic Ideas (Lattice). Note the Chain in the Lattice
in Fig. 1-7: ( Dog, Mammal, Vertebrate, Animal ) = ( a, ab, abc, abcd ).
17
WILLIAM S. VEATCH
(Larger lattices become cumbersome to work with, but fortunately, we can
use another tool, Partition Equations, to make it easier to work with
larger Power Sets.)
1.6. Applied Math Without Numbers (MWN): Arguments,
Debates, and Policy Making
We have seen some examples of how we can apply the concepts of the
Mathematics of Ideas (e.g., Atoms, Coatoms, Chains, Power Sets, and
Lattices) to objects in the Physical World, but the same rules also apply to
abstract Ideas, including how we form Arguments. We will explore in detail
the concept of an Action Chain or Argument Chain, to show how to
develop sound, persuasive Arguments. Fig. 1-8 shows a sample Argument
Chain.
18
MATH WITHOUT NUMBERS
In the Argument Chain example, we have a Set of possible Actions a
through g that could be taken to solve a particular issue or problem, such
as how best to combat terrorism. For each Action, starting at the top of the
Argument Chain, we first ask whether the Action addresses the issue at hand,
i.e., in this case, does the Action combat terrorism? If true, we move on to
the next element of the test (letters a through f in the sample Argument
Chain); if false, the inquiry ends there (letter g in the sample Argument
Chain). Then, since we are talking about actions that the federal government
might take, we ask whether the proposed Action is within the power and
authority of the federal government as set forth in the US Constitution.
19
WILLIAM S. VEATCH
Again, if true, we move to the next text; if false, the inquiry stops there.
We study this Argument Chain in more detail in Chapter 8, but for
now, just note the fact that the Test Elements for a sound policy decision
take the form of a Chain: ( a, ab, abc, abcd, abcde, abcdef, abcdefg ).
Notice the similarity to the Dog-Animal Chain = ( Dog, Mammal, Vertebrate,
Animal ) = ( a, ab, abc, abcd ).
In our study of debating and policy making, we will see how we can use
the Mathematics of Ideas to compare and contrast two competing
Arguments, and apply Values, in order to decide which is the better
Argument, and ultimately, to decide how we should act in our lives. We will
see, however, that while mathematics can tell us how we think, it cannot tell
us what to think or how to act. Mathematics can, however, help us
communicate in a clear and precise manner, and can highlight for us how two
competing Arguments differ from one another.
1.7. Update to The Definition of Idea
At the end of each Chapter of this Book, we will update our definition of an
Idea to reflect what was discussed in the course of that Chapter. So far, we
have the following:
Definition Idea: An Idea is any thought that can be conceived of in
the mind.
1.8. References, Historical Notes, and Further Reading
Mathematical Philosophy
[Russell 1920]: Bertrand Russells Introduction to Mathematical Philosophy is
perhaps the best place for the reader to start, if looking for more background
on the philosophical issues that we explore in our study of the Mathematics
of Ideas.
Logic
[Kreeft 2010]: In Peter Kreefts Socratic Logic at page 43 et seq., the author
discusses the concepts of extension and comprehension, which are
similar to our Math Without Numbers concepts of Objects and Attributes,
respectively. This book is an excellent contemporary introduction to classical
Socratic Logic.
20
MATH WITHOUT NUMBERS
[Minto 1893]: Although written over a hundred years ago, in Logic
Inductive and Deductive, William Minto provides a very readable overview of the
classical logic of Aristotle.
[Smullyan 2014]: Raymond Smullyan provides an excellent introduction
to mathematical logic in his book A Begninners Guide to Mathematical Logic.
[Tall 2014]: In From Mathematics in Logic to Logic in Mathematics, Aliou Tall
provides an important historical perspective on the evolution of thinking in
the field of mathematical logic.
[Shenefelt and White 2013]: In If A, Then B- How the World Discovered
Logic, the authors Michael Shenefelt and Heidi White provide another
excellent, contemporary review of the development of Logic throughout
history.
Formal Concept Analysis
[Ganter and Wille 1999]: For those readers with a deeper background
in mathematics, this book, Formal Concept Analysis, discusses intent and
extent, which are similar to our Math Without Numbers (MWN) concepts
of Attributes and Objects, respectively. At page 22, the authors discuss dual
concept lattices, which are similar to our Dual Power Set Lattices in MWN.
One of the principal differences between FCA and MWN is that in MWN
we focus on Power Sets, which also constitute Boolean Algebras, Rings,
Lattices, and Topologies; whereas, FCA focuses more on certain subsets of
Boolean Lattices known as Complete Lattices.
21
WILLIAM S. VEATCH
2. HOW DO WE KNOW WHAT AN IDEA MEANS
OR REPRESENTS? THE DUAL NATURE OF
IDEAS AS SETS OF OBJECTS AND ATTRIBUTES
How do we know what an Idea means, or what it represents? As we will
see in this Chapter, our understanding of the meaning of an Idea comes about
in two distinct ways: (1) by listing examples of what the Idea is (an Object
Set or List Set), and (2) by making a list of the attributes or elements of
a test for determining whether or not something is represented by a particular
Idea (an Attribute Set or Test Set). This dual nature of Ideas as both
sets of Objects, and sets of Attributes is one of the most important concepts
to understand as a foundation for the study of the Mathematics of Ideas.
Also, as we will see later, in mathematical terms we can interpret Attributes
to be what are known as Equivalence Relations that form Partitions;
and Objects are the Equivalence Classes formed by the Partitions. To
illustrate the concept, we start with a diagram showing how the Ideas in the
Set X = { Animal, Vertebrate, Mammal, Dog } are related to one another.
See Fig. 2-1.
22
MATH WITHOUT NUMBERS
The letters represent Objects. The letter a represents a Dog. The
letter b represents Mammals other than Dogs, so a+b = all Mammals. The
letter c represents other Vertebrates, and a+b+c = all Vertebrates.
Similarly, the letter d represents Animals other than Vertebrates, and
a+b+c+d = all Animals. The letter e represents all other Ideas in the
Universe that are not Animals.
The circles themselves are numbered, and each circle (and the area
inside the circle) represents an Attribute or Set of Attributes. The circle
labeled 5432 represents all Attributes of Dogs, and 2 represents all of
the Attributes that make a Dog a Dog, and not another type of Mammal.
Similarly, the circle labeled 543 represents all Attributes of Mammals, and
3 represents the Attributes of Mammals that distinguish Mammals from
other Vertebrates. The circle labeled 54 represents all Attributes of
Vertebrates, and the number 4 represents the Attributes of Vertebrates that
distinguish Vertebrates from other Animals. The circle labeled 5
represents all Attributes of Animals that distinguish Animals from other
Ideas in the Universe of Ideas.
The Object a inherits all of the Attributes of each circle, or Partition,
that encloses it, in this case the circles labeled 5432, 543, 54 and 5. For
example, Dog is within the Mammal-543 circle. As a result, since female
Mammals nurse their young and all Mammals have fur/hair, these Attributes
23
WILLIAM S. VEATCH
of Mammals are true of Dogs as well. We can summarize the Objects and
Attributes in the following Table 2-1.
The above table illustrates a number of important concepts, including a
Chain of Objects: X = (, a, ab, abc, abcd, abcde ), where each Element of
the Set is a Subset of the succeeding Element; and a dual inverse Chain of
Attributes: Y = [ 54321, 5432, 543, 54, 5, ], where each Element of the Set
is a Superset of the succeeding Element. Now, we look at Objects and
Attributes in more detail.
2.1. An Object is an Example of an Idea
When someone asks us what we mean by the Idea of Color, we may start
by listing out a variety of colors, such as the colors of the rainbow: Red,
Orange, Yellow, Green, Blue, Indigo, and Violet. Each of these colors is an
example of the Idea of Color.
In the notation of Set Theory, we write Color = { Red, Orange, Yellow,
Green, Blue, Indigo, Violet } to indicate that Color is a Set made up of the
elements Red through Violet. When we create a list of Objects in this way,
we sometimes refer to it as a List Set. Other examples of List Sets of
Objects include the following:
Animal = { Dog, Cat, Horse, Pig, Other Animals }. We use the
term Other Animals to encompass all of the other animals that
are too numerous to list. This is a common device that we will
use when we have particularly long List Sets.
Herbs = { Basil, Parsley, Rosemary, Thyme, Chervil, Other
Herbs }.
Type of Government = { Democracy, Dictatorship, Theocracy,
Other Government Type }.
Plants = { Flowering Plants, Conifers, Other Plants }.
Vehicle = { Car, Boat, Plane, Other Vehicles }.
24
MATH WITHOUT NUMBERS
Sports = { Soccer, American Football, Basketball, Volleyball,
Other Sports}.
Possible Immigration Policies = { build a wall, grant amnesty to
undocumented workers, set up drone surveillance at the border,
other immigration policies }.
Note that when we write an Object Set, the Elements may or may not be
mutually exclusive of one another. For example, in the Set X = { Dog, Cat,
Mammal }, the Elements are not Mutually Exclusive, because a Dog is
both a Dog and a Mammal, and a Cat is both a Cat and a
Mammal. This is easily illustrated using a Euler Diagram, as set forth in
Fig. 2-2.
Another tool we can use to identify Mutually Exclusive Sets, i.e., Sets
where each pair of Elements is Disjoint, is what we call an Is a Relation
Table. Fig. 2-3 illustrates how we can use an Is a Relation Table to
determine whether or not two Sets are Mutually Exclusive. We draw a line
from the left column to the right column for any Element in the left column
that is a something in the right column. A Dog is a Dog, so we draw a
line from Dog in the left column to Dog in the right column. A Dog
is a Mammal, so we draw a line there as well.
25
WILLIAM S. VEATCH
The Elements in the table on the left ( Dog and Cat ), are Mutually
Exclusive, as verified by only straight lines across the table. On the other
hand, the Elements in the table on the right ( Dog, Cat, and Mammal ), are
not Mutually Exclusive, as indicated by the diagonal lines. The advantage to
the Is A Relation Table over Euler Diagrams, is that we can list a dozen or
more elements and the resulting diagram is much less confusing than trying
to draw a dozen or more overlapping circles.
As we will see later, Mutually Exclusive Sets, where each pair of Elements
is Disjoint, play a special role in our Mathematics of Ideas. In particular, any
time we refer to a Partition, it is understood that the Elements in the
Partition Set must be Mutually Exclusive. Sometimes we will write the
subscript ME to indicate that the Elements of a Set are Mutually Exclusive,
although in the case of Partition Sets it is understood that the Elements must
be Mutually Exclusive.
In the example of the Set of Colors = { Red, Orange, Yellow }ME, if an
Object is Red, it cannot be Orange, as the two colors are Mutually Exclusive.
Similarly, in the Set Government = { Democracy, Dictatorship }ME, a
Government cannot be both a Democracy and a Dictatorship at the same
time.
To give another example of a set that is not a valid Mutually Exclusive
Object Set under our rules, consider Living Thing = {Animal, Dog, Plant,
Cat}. The Objects of this set are not Mutually Exclusive because an Object
can be both an Animal and a Dog, or an Animal and a Cat, at the same time.
You might ask: why do we care if the elements of a set are mutually exclusive?
The answer will become apparent later, as sets of mutually exclusive elements
behave in an interesting way in mathematical terms. If it is not obvious that
we intend to have Sets with mutually exclusive elements, then we may use
the subscript ME to emphasize the point. (This illustrates a difficulty in
how we use words and language, as it is not always obvious whether or not
two Ideas are Mutually Exclusive.)
26
MATH WITHOUT NUMBERS
Observation: Every Idea can be described by a Set consisting of a list of mutually
exclusive Objects that are examples of that Idea.
2.2. An Attribute is a Property or Characteristic that We Use to
Identify an Idea
When someone asks us what we mean by the Idea of Democracy, we might
start with a definition such as the following generic dictionary definition:
A form of government,
By the people,
Where the majority rules,
Supreme power is vested in the people,
Power is exercised by the people directly or indirectly through a
system of representation, and
There are periodic elections.
This definition is not a list of governments that are democracies, but rather
the definition is a list of the Attributes of a form of government that make
that form of government a democracy. Stated another way, the definition is
comprised of a list of the elements of a test for determining whether or not
an Object is a democracy or not. The elements form a Set where
DemocracyAtt = { form of government, by the people, the majority rules,
supreme power is vested in the people, exercised directly or by means of
elected representatives, periodic elections }. When we list out Attributes in
this way, we sometimes refer to the Set as a Test Set. We use the subscript
Att or Test to identify Attribute Sets and to distinguish them from
Object Sets with the Obj or List subscript.
Alternatively, we could describe a democracy in terms of countries in
the world that constitute democracies. In this case, we would write:
Democracy Obj = {United States, Canada, England, France, Germany, Other
Democracies}. (As noted above, we use the subscript Obj to identify
Object Sets.) Note that each Object in the Object Set has each of the
Attributes of the related Attribute Set. Of course, people can disagree on
what exactly the Object Set or Attribute Set for a democracy should be, but
the techniques involved in analyzing an Idea in terms of Objects and
27
WILLIAM S. VEATCH
Attributes remain valid.
Note that the same Idea, Democracy, can mean both a list of
Attributes of what makes a democracy a democracy; and a list of Objects, in
this case countries where the type of government meets the test for a
democracy. We will talk more about this dual nature of Ideas as we proceed.
Going back to our earlier example of the Idea Color, an alternative
way of defining Color would be to list the Elements of what makes a color a
color, i.e., Hue, Brightness/Lightness, and Saturation. In set theoretic
notation, we write: Color Att = {Hue, Brightness/Lightness, and Saturation}.
More often, however, in everyday life we use a List Set of Objects to describe
colors.
As was the case with a List Set of Objects, the Elements of a Test Set
of Attributes may or may not be mutually exclusive of one another. For
example, in the Set Y Att = { Vertebrate, Mammal, Dog }, the Elements are
not Mutually Exclusive, because Vertebrate is both an element of the test
for Vertebrate as well as for Mammal, and Dog; and Mammal is
both an element of the test for Mammal and for Dog. Note that as with
Objects, the is a test can also be used for Attributes to specify a Mutually
Exclusive Set of Attributes, but in a slightly different way. For Attributes, we
say that each element of the Attribute Set is an attribute of, or element of
the test for the specified Idea. Fig. 2-4 illustrates how we can use an Is a
Relation Table, applied to the same Sets we used in Fig. 2-1, to determine
whether or not two Sets of Attributes are Mutually Exclusive. We draw a
line from the left column to the right column for any Element in the left
column that is an element of the test for something in the right column.
28
MATH WITHOUT NUMBERS
In Fig. 2.4, the left column shows the is a relation table for Objects;
and the right column shows the is an element of the test for relation table
for Attributes. Note how the diagonal lines go in opposite directions, which
is what you would expect given the inverse relationship of Objects and
Attributes. (In set theoretic terms, the Element on the left side is equal to or
a subset of the Element to which it is connected on the right side.)
Observation: Every Idea can be described in terms of both a list of Objects, which
are examples of the Idea and which may or may not be mutually exclusive (a List
Set); and a list of Attributes consisting of test elements or characteristics of the Idea,
which may or may not be mutually exclusive (a Test Set).
Now, we examine more closely the concept of a Partition. Whereas in
the case of Objects we interpret each Object as an example of an Idea, in the
case of Attributes, we interpret each Attribute to be a Partition, where the
Attribute divides the Universe of Ideas into two classes of Objects: either an
Object exhibits the Attribute or it does not. In other words, each Attribute
divides the Universe of Ideas into two sections that we refer to as
Equivalence Classes in mathematics. Fig. 2-5 illustrates some sample
Attributes, where each Attribute can be seen to divide the Universe of Ideas
into two parts in which the Attribute is either True or False. The complexity
of Attributes arises by virtue of the way in which different Attributes interact
with each other. Each pair of Attributes, however, will be related to one
another in one of the five ways mentioned earlier in Section 1.5: the Same,
Disjoint, Partially Overlapping (POL), Subset-Superset, or Superset-Subset.
29
WILLIAM S. VEATCH
2.3. Objects and Attributes are Inversely Related
Interestingly, when we look at certain Sets that have overlapping Objects and
Attributes, we find that there is an inverse relationship between the number
of Objects and the number of Attributes. Consider the Sets in Table 2-2.
30
MATH WITHOUT NUMBERS
The fewer the number of Attributes we require Objects to have, the more
Objects there are that meet the test. If the only required Attribute is that the
Object is alive, then Dogs, Other Mammals, Other Vertebrates, Other
Living Things all qualify. If, however, we require that Objects also Have a
Backbone, then Other Living Things, such as invertebrates, drop out.
The relationship between an Object-Set and Attribute-Set for a particular
Idea can be visualized in a number of ways. We started by looking at a Euler
Diagram (see, Fig. 2-1), and now we look briefly at what we call a Chain.
(We will explore Chains in detail in Chapter 8.)
The following Table illustrates a Chain of Objects and an inverse Chain
of Attributes. By Chain, we mean that each Element is a subset (or superset
if going the other direction) of the next Element in the Chain.
On the one hand, this inverse relationship may seem like simply common
sense: given a set of Objects, the more Attributes you require an Object to
have, the fewer the number of Objects that will meet the test. This
observation, however, has great significance when it comes to understanding
the Mathematics of Ideas, as we will see later.
The Object Set and Attribute Set exist for each Idea, like two sides of the
same coin. We can prove, in fact, that once we specify an Object Chain, the
Attribute Chain can be calculated; similarly, once we specify an Attribute
Chain, the Object Chain can be calculated. Can we say that one or the other,
i.e., Objects or Attributes, is a better way to define an Idea? In practice, we
use both methods. For example, when defining the Idea color, we tend to
list the Objects, such as red, yellow, green, etc., rather than the Attributes,
such as Hue, Saturation, and Brightness/Lightness. On the other, with an
Idea such as Democracy, we tend to list the Attributes, such as a
Government by the people, where the majority rules, and where periodic
elections are held. The best answer is to recognize that we use both Objects
31
WILLIAM S. VEATCH
and Attributes to define Ideas, and that the two concepts are inextricably
linked.
2.4. Using Power Sets of Atoms and Coatoms to Represent Objects
and Attributes
In this Subchapter, we will explore in more detail how we can use a Power
Set of Atoms to represent Objects, and a Power Set of Coatoms to represent
Attributes. It is easy to understand the Power Set of Objects. We are
accustomed to thinking of tangible Objects as forming Sets, and the Power
Set of Atoms is simply the Set of all possible subsets of a Set of Objects.
Admittedly, however, it is much more difficult to understand the Power Set
of Coatoms and how it relates to Attributes. For this reason, we take some
time now to explore this concept in more detail.
Hypothesis: The Power Set of Atoms represents the Set of all possible Ideas
expressed as Objects (Atomic or Compound); the Power Set of Coatoms represents the
Set of all possible Ideas expressed as Attributes (Atomic or Compound).
The key to proving this hypothesis is to start with the assumption that
the Universe of all Ideas exists. With this assumption in hand, we can
demonstrate that an Attribute actually does two things: the Attribute divides
the Universe into (i) a class of Objects that exhibit the Attribute, and (ii) a
class of Objects that do not exhibit the Attribute.
Step 1: Each Attribute Creates a Partition of the Universe of Ideas into
Two Equivalence Classes
As noted earlier, an Attribute partitions the Objects in the Universe of
Ideas into two Equivalence Classes. See Fig. 2-6. By Equivalence Class,
we just mean that all of the Elements in a particular class or Set are the same
as each other when it comes to this particular Attribute. For example, if the
Attribute in question is Red Things, then either an Object is a Red Thing
or it is not a Red Thing. If the Attribute is Mammal, then either an Object
is a Mammal or it is not a Mammal.
32
MATH WITHOUT NUMBERS
We use xs to represent Objects labelled a, b, c, ; the circle itself
and the area within the circle represent the Attribute labeled 2 in this case.
The area outside the circle represents the Object (b) and the Attribute
~(a)=[1]. We say that the Attribute [2] is True for Objects inside the circle,
and False for Objects outside the circle. The key point here is that, so long
as we assume that a Universe of all Ideas exists, a single Attribute creates two
Equivalence Classes.
Step 2: There are two related Power Sets
We can also use Power Set Lattice notation to illustrate that a single
Attribute partitions the Universe of Ideas into two Equivalence Classes. See
Fig. 2-7. The Power Set Lattice notation clarifies how we arrive at our
numbering system. If we have a single Attribute, it will divide the Universe
of Ideas into two parts: Object (a) and Object (b), where (a) and (b) may
themselves be Sets. Given a two-Element Universe with Elements (a) and
(b), writing a is equivalent to writing not-b or ~b. Substituting
numbers 1, 2, 3, for the letters ~a, ~b, ~c, gives us a dual, inverse
Power Set Lattice of Coatoms. You might ask: why do we bother to use
numbers 1, 2, 3, instead of saying ~a, ~b, ~c, ? The reason is that we
will use numbers in square brackets to represent Attributes, and letters in
parentheses to represent Objects. Using different notation helps to
emphasize the difference between Objects and Attributes, so that we do not
confuse the two.
33
WILLIAM S. VEATCH
Fig. 2-8 gives an example of dual Object - Attribute Power Sets with
four Atoms. The important thing to understand here is that the two Power
Sets represent exactly the same Ideas, but just expressed in different ways.
34
MATH WITHOUT NUMBERS
We can demonstrate that the two Power Sets are really just two ways of
writing the same thing. For example, the Element (abc) in the Object Power
Set is the same as ~(d)=[4] in the Attribute Power Set. Therefore, you might
ask: do we really need both Power Sets, if one can be calculated from the
other? There are several reasons why the dual Power Sets are helpful:
In practice, we think of Ideas as having two characteristics: (i) an
Idea can be described as a list of examples of the Idea, which we
refer as a Set of Objects and (ii) an Idea can be described as a list of
required properties, which we refer to as a Set of Attributes. As a
result, it helps to separate out the two concepts.
Just as we observe Objects and Attributes to have an inverse
relationship, the Power Sets of Atoms and Coatoms have a
corresponding inverse relationship.
The Power Set of Coatoms will be helpful to show Inheritance of
Attributes. (We will discuss this more in Step 4.)
Often either the Set of Objects or the Set of Attributes will be
significantly smaller for any given Idea, so it helps to use whichever
is smaller. (Later, we will show that as we approach an infinite
number of Objects, we approach a single Attribute Coatom; and
conversely, as we approach an infinite number of Attributes, we
approach a single Object Atom.)
35
WILLIAM S. VEATCH
New areas of Mathematics, such as Formal Concept Analysis, are
exploring the idea of dual Power Sets of Objects and Attributes.
Step 3: The Equivalence Class where the Attribute is True will appear
somewhere in the Power Set of Atoms
The Equivalence Class consisting of Atoms where the Attribute is True
will appear somewhere in the Power Set of Atoms. In fact, each of the two
Equivalence Classes (True and False) will appear somewhere in the Power
Set of Atoms, because all possible subsets are in the Power Set by definition.
For example, suppose we have four Atoms in our Universe X = { a,b,c,d },
and two Atoms are Red and two are not-Red: Red = ( a, b ) and ~Red =
( c, d ). The Attribute Red created a Partition of the Universe into two
Sets, and each Set appears in the Power Set of X. See Fig. 2-9.
In this example, the pair of Sets created by the Partition were ( a, b ) and
( c, d ), but in total there are 16 possible Partitions of a four Atom Set into
two parts where one part is True and the other part is False, as follows:
36
MATH WITHOUT NUMBERS
In other words, depending upon the Attribute in question, there could
be 0, 1, 2, 3, or 4 Atoms that exhibit that particular Attribute. Table 2-4 lists
all of the possible ways in which a four Atom Set could exhibit a single
Attribute. In our example of Red and ~Red in Fig. 2-9, the Partition
corresponds to the line Red: (ab)=[43] and ~Red: (cd)=[21], because a and
b are Red, and c and d are ~Red.
Note that the True and False Sets are Complements of one another,
meaning that they are Disjoint, but together they constitute the entire
Universe of Ideas. Our focus, however, is on the Elements where the
Attribute is True. The Power Set of Objects is made up of the Elements
where the Attribute is True; and the Power Set of Attributes is made up of
the negative of the Complement Set. For example, the Element (a) is
equivalent to the negative of the Complement Set (bcd), which we write as
~(bcd) = [432].
Step 4: We can map each Attribute of an Idea to one of the entries in
the Power Set of Coatoms
Now, we come to the critical step. Can we show that the Power Set of
Coatoms adequately represents Attributes? Conveniently, the Power Set of
Coatoms provides us with a means of representing Attributes, because it is
possible to map all possible Attributes to the Elements in the Power Set of
37
WILLIAM S. VEATCH
Coatoms. (As we will see shortly, in everyday applications it is a many-to-
one mapping. In theory, we could make this relationship one-to-one by
adding more Atoms, but many-to-one is adequate for our purposes.) In other
words, each Attribute will map to a single entry in the Power Set of Coatoms.
Intuitively, this is clear, because every Attribute creates a Partition of the
Universe of Ideas into two complementary Sets, and the True Equivalence
Class must map to one of the entries in the Power Set, since by definition the
Power Set includes all possible subsets. To give a simple example, we assume
that we are working with a Universe with four Atoms. Any given Attribute
will divide the Universe into two classes as set forth in Table 2-4 above.
Suppose we take the four Atom Power Set of Objects, where the four
Atoms are Black Dog, Brown Cat, Black Snake, and Brown Crab. Fig. 2-10
illustrates the various Set relationships.
Now, we list all of the Attributes of the Atoms in the Power Set.
38
MATH WITHOUT NUMBERS
For any particular Attribute listed in the left hand column, the Atoms
listed across the top row that are marked with an X exhibit that Attribute.
By looking which Atoms have an X for a particular Attribute, we can map
that Attribute to the dual Power Sets. For example, for the Attribute has
four legs, the Atoms (a) and (b) each have four legs, so we map has four
legs to (ab) in the Object Power Set of Atoms, which corresponds to [43] in
the Attribute Power Set of Coatoms. The column at the far right shows the
Element in the Power Set of Coatoms to which each Attribute maps.
We say that the mapping is many-to-one, because there may be many
Attributes that map to the same Element of the Power Set of Coatoms. For
example, has hair/fur, nurses young, has four legs, and is a Mammal
all map to [43] in the Power Set of Coatoms. The reader might remark at
this point: but not all Mammals have four legs, do they? This is true. The
mapping of the Attribute has four legs is a result of using a Universe of
Atoms that includes Dogs and Cats, but not other Mammals such as Humans
or Whales. If we expand our Universe of Atoms to include other types of
Mammals, then we would find that there would be at least three subcategories
of Mammals: those that walk on two legs, those that walk on four legs, and
those that live and swim in water. (Later, we will see that a many-to-one
mapping, as opposed to a one-to-one mapping, is a sign that we could expand
39
WILLIAM S. VEATCH
our Base Set of Atoms.)
Now, we can see from Table 2-5 the enormous benefit to mapping
Attributes to the Elements of the Power Set of Coatoms: we now have a clear
record of the Inheritance of Attributes. For example, we can see that
Dog maps to [431] and Cat maps to [432]. Any single number or
combination of numbers that Dog and Cat have in common represents
Attributes that Dogs and Cats have in common. In this case, the possible
common Elements are [3], [4], and [43]. The number [3] does not have any
particular significance in this example, but 4 maps to has a backbone and
is a Vertebrate, and [43] maps to has fur/hair, nurses young, is a
Mammal, and has four legs. As a result, we know that Dog and Cats share
all of these common Attributes. This may not seem like an earth-shattering
result as to Dogs and Cats, but it will prove to be a very useful tool when we
are trying to communicate more complex Ideas that are not so well
understood. Also, it will be a useful tool in artificial intelligence, because
whereas the letters a and b do not communicate any information as to
how the two Objects are related, [432] and [431] immediately convey that the
two Objects share any Attributes that map to 3, 4, or 43. This is something
that a computer can process very easily.
This ability to track Inheritance of Attributes is the principal advantage to
using the Power Set of Coatoms to represent Attributes, rather than just using
Atoms or assigning random numbers. With another numbering system, we
would lose some of the information about Inheritance of Attributes.
To emphasize this point about Inheritance of Attributes, we look more
closely at an example of a Chain.
40
MATH WITHOUT NUMBERS
Chains are particularly useful tools for demonstrating the Inheritance of
Attributes, so we use them frequently when studying almost any subject.
Observation: Attributes form a many-to-1 (or 1-to-1 if we add more Atoms)
mapping to the Power Set of Atoms/Coatoms. A set of Objects completely determines
a corresponding set of Attributes; and vice versa. Mapping Attributes to the Power Set
of Coatoms is useful because we can see how Ideas inherit Attributes from other related
Ideas.
At first it may seem pointless to use two different Power Sets that are
mathematically equivalent. As a practical matter, however, using both a
Power Set of Atoms and its dual Power Set of Coatoms provides a
convenient and mathematically precise way to represent what we think of
intuitively as Objects (examples) and Attributes (properties). The inverse
relationship of the Power Set of Atoms and Power Set of Coatoms matches
nicely with the inverse relationship of Objects and Attributes.
Step 5: We can create an Idea Signature
In this last step, we talk more about the fact that every Idea will have
both an entry in the Power Set of Atoms, and a corresponding entry in the
related Power Set of Coatoms. We can create what we call an Idea
Signature by pairing the Object Set Element with its related Attribute Set
Element. To clarify which are Objects and which are Attributes, we use
parentheses for Objects, and square brackets for Attributes. For a four Atom
Power Set, we have the following:
41
WILLIAM S. VEATCH
Note the following points about the Idea Signature:
We can calculate the Objects from the Attributes; and the Attributes
from the Objects.
The total number of Elements is always equal to n, i.e., the number
of Deemed Atoms in the Universe of Discourse. In the above
example, since there are four Atoms, there are always four Elements,
not counting the Empty Set. For example, (a)[432] has four
Elements, (b)[431] has four Elements, etc.
The Objects and Attributes are inversely related, with fewer Atoms
at the top, and fewer Coatoms at the bottom.
The following are some more examples of Objects and their unique Idea
Signatures given the specified Domains of three, four, or five Atoms. Since
the Power Sets grow very rapidly in size, with 23= 8, 24=16, and 25=32
Elements, we do not list all of the Elements, but just a few to give you a sense
of what the Idea Signatures look like:
42
MATH WITHOUT NUMBERS
Domain X = { a, b, c }:
o {a} = (a)[32],
o {b}= (b)[31],
o {c} = (c)[21],
o {a,b} = (ab)[3],
o {a,b,c} = (abc)[].
Domain X = { a, b, c, d }
o {a} = (a)[432],
o {b} = (b)[431],
o {c} = (c)[421],
o {d} = (d)[321],
o {a,b} = (ab)[43],
o {a,b,c,d} = (abcd)[],
Domain X = { a, b, c, d, e }
o {a} = (a)[ 5432],
o {b} = (b)[ 5431],
o {c} = (c)[5421],
o {d} = (d)[5321],
o {e} = (e)[4321],
o {a,b} = (ab)[543],
o {a,b,c,d,e} = (abcde)[].
As a reminder on how to read an Idea Signature, if we have the Domain
X = { a, b, c }, then when we select the Element a we are simultaneously
selecting not { c, b }. (Typically, we write the not elements in reverse
alphabetical order, but technically these are not ordered sets.) We write ~
instead of not, so we have ~{ c, b }. Substituting numbers for letters,
we have [3,2]. (For small numbers of Coatoms we often omit the commas
and write [32].) We use parentheses for Object Sets, and square brackets for
Attribute Sets to emphasize that we are talking about Attributes not Objects.
In summary, the Object Element a has the Idea Signature (a)[32], where
the Domain has three Atoms. If the Domain has four Atoms, then Object
Element a has the Idea Signature (a)[432].
The significance of the Idea Signature is that we have taken the mystery
out of the notion of Attributes by defining an Attribute as an entry in the
Power Set of Coatoms. The key to understanding Attributes in a
mathematical sense is to determine by experimentation and research, the
correct mapping from Attributes in the World of Ideas to Elements of the
Power Set of Coatoms in the World of Abstract Sets. See Fig. 2-11.
43
WILLIAM S. VEATCH
In summary, we have learned that while in the Physical World and World
of Ideas we view Objects as possessing certain properties that we refer to
as Attributes, in the World of Abstract Sets we can interpret Attributes as
merely corresponding entries in the Power Set of Coatoms. In one sense,
this approach may be unsatisfying because it takes the mystery out of the
concept of properties or attributes. Rather than assume that Objects
somehow possess Attributes, we have shown that in the World of Abstract
Sets, Attributes are merely equivalent Sets of Coatoms. On the other hand,
this makes perfect sense. If we wish to develop a Mathematics of Ideas, then
it needs to be clear and precise, not surrounded by mystery. As discussed
more at the end of this Chapter (see Subchapter 2.8), it is important to
44
MATH WITHOUT NUMBERS
distinguish among the Physical World, the World of Ideas, and the World of
Abstract Sets. The Physical World and the World of Ideas are still cloaked
in mystery; only the World of Abstract Sets is clear and precise. Creating the
map of relationships among the three Worlds, however, remains challenging.
2.5. Cover Plus a Difference: A Third Way of Defining an Idea
There is a third way we can define an Idea which we call Cover Plus a
Difference, that draws upon both Objects and Attributes. We have seen
that we can define an Idea by listing examples of the Ideas which we refer to
as Objects. E.g., Dog Obj = { Lab, Pug, Dalmatian, Other Dog }.
Alternatively, we can define an Idea by listing its characteristics, which we
refer to as Attributes. Attributes can also be thought of as the Elements of
a Test for whether or not an Object can be classified as a particular type of
Object. E.g., Dog Att = { Living Thing, Animal, Vertebrate, Mammal,
Carnivorous, Walks on Four Legs, Barks, Has Dog Shape }. Sometimes,
however, it is helpful to define an Idea with reference to other similar Ideas.
The technique we use is to choose an Attribute of the Idea that we know is
shared by other similar Ideas (i.e., the Cover), and then distinguish the
Idea from the other related Ideas using a specific Attribute that is unique to
the Idea in question (i.e., the Difference). Note that the Cover is a
Superset of the Idea that we are defining, i.e., an Idea that is higher up in a
Chain. In our example of a Dog, there are a number of Attributes that we
could use as a Cover, including Mammal, Vertebrate, Living Thing, etc., but
often it is most helpful to choose a closely related Attribute, such as Mammal
in this case. Suppose we define Mammal as X = { Dog, Cat, Human, Cow,
Pig, Horse }. Is there a particular Attribute that distinguishes a Dog from all
of the other Mammals? One possibility is the Attribute Barks, since of the
listed Mammals, only a Dog barks. Therefore, to define Dog by Cover
Plus a Difference, we define Dog as a Mammal (i.e., the Cover) that Barks
(i.e., the Difference).
In summary, to define an Idea by a Cover Plus a Difference, we need to
follow these steps:
Choose an Idea to define (e.g., Dog).
Select an Attribute of the Idea to use as a Cover Idea, preferably a
Superset higher up in the Chain of Objects, that is closely related to
the Idea that we defining (e.g., Mammal).
Create a Partition Set for the Cover Idea, i.e., an Exhaustive, Mutually
Exclusive Set of Objects that make up the Cover Idea (e.g., Mammal
= { Dog, Cat, Human, Cow, Pig, Horse }).
45
WILLIAM S. VEATCH
Identify an Attribute that distinguishes the Idea that we are defining
from all of the other Ideas in the Cover (e.g., a Dog Barks).
Using the Cover Plus a Difference technique to define an Idea can be
extremely helpful, as it allows us to use the name of the Cover as a shorthand
proxy for all of the inherited Attributes. For example, if we let X = { dog,
cat, snake, crab, salmon } = { a,b,c,d,e }, we can see that the word Mammal
stands for both the Object Set (ab) and the Attribute Set [543]. In terms of
Defining the Idea Dog as a Cover plus a Difference, we can define Dog
in either of two ways:
Attributes: Dog Att = Mammal + Barks = [543] +[2] = [5432]; or
Objects: Dog Obj = Mammal Cat = (a,b) (b) = (a).
This provides a key to simplifying the Idea Signature by defining the Idea
Signature for Dog to be, for example, (a)[Mammal+2]. In an infinite
Universe of Ideas, we may not know exactly what Elements are subsumed by
the Attribute Mammal, but we can discuss any Elements below Mammal
in a Chain or Power Set without any worries.
2.6. The Importance of Specifying the Domain
When we attempt to add/subtract or multiply/divide Ideas, we must be
careful to specify the Domain. The reason for this is that we cannot
add/subtract/multiply/divide items that belong to different Domains. For
example, the Natural Numbers and the Universe of Ideas are different
Domains, so the expression 2 + Dog does not make any sense. We cannot
add elements taken from different Domains. On the other hand, the
expressions 2+2=4 and Dog+Cat=Mammal are fine, because we are
adding elements taken from a single Domain.
As it turns out, there are several Domains that we use regularly when
working with the Mathematics of Ideas, including the following:
Sets of Object Atoms,
Sets of Attribute Coatoms,
Power Sets of Atoms, and
Power Sets of Coatoms.
46
MATH WITHOUT NUMBERS
See Fig. 2-12. As we progress on our journey, we will develop rules for
adding/subtracting/multiplying/dividing Sets of Objects with Sets of
Objects, Sets of Attributes with Sets of Attributes, Power Sets of Objects
with Power Sets of Objects, and Power Sets of Attributes with Power Sets of
Attributes, but we cannot combine elements across different Domains.
2.7. Working with Sets of Atoms (Coatoms) v. Working with
Power Sets of Atoms (Coatoms)
In this section we discuss a means of working with Atoms (and Coatoms),
and converting to or from Powers Sets. Power Sets have the advantage of
illustrating all possible combinations of Atoms (or Coatoms). Admittedly,
they become difficult and cumbersome to use (or to draw!) if there are more
than four or five Atoms. The good news is that clear rules allow us to switch
back and forth from using Atoms (Coatoms) and Power Sets. In fact, a
central theme throughout our study of the Mathematics of Ideas is that we
can easily switch back and forth between working with Atoms and working
with Power Sets of Atoms. This is important, because there are far fewer
Atoms (n) than Elements in a Power Set (2n). As a result, it is often easier to
work with, and manipulate, Atoms. On the other hand, many Ideas are
Compound Ideas, so we need a way of working with the entire Universe of
Ideas, including both Atoms and Compound Ideas. As noted earlier, a Power
Set includes a Set of Atoms, together with all possible Compound Ideas that
we can form from those Atoms.
47
WILLIAM S. VEATCH
Here are some examples of converting from Sets of Atoms to Power
Sets:
In the first three examples in Table 2-8, it may not be clear at first why
we cannot combine elements from the two Domains of Atoms and Power
Sets of Atoms, but in the fourth example, the problem becomes apparent.
Any element that represents a simple Set in the Domain of Atoms, will
represent a Power Set in the Power Set Domain. For example, B
represents a simple Set B = { b, c } in the Domain of Atoms, but represents
a Power Set, BPS= (, b, c, bc ), in the Power Set Domain. The Set B is
entirely different, depending upon the Domain. Note that we use the
subscript PS after the close parentheses as a reminder that any Elements
that are Sets, such as B in the example, are in fact Power Sets.
From these examples, we can derive a number of Rules that will help us
convert back and forth from Atoms(or Coatoms) to Power Sets.
Rule 1: If we take the Power Set of a Set X with 1,2,3, Atoms, we get the
familiar Power Set structure where the number of Atoms in each row (i.e.,
the series 1, 121, 1331, 14641,) matches up with the familiar structure
known as Pascals Triangle. (For those with a deeper interest in mathematics,
a curious fact is that the number of Atoms in each row also corresponds to
the coefficients in equations generated by the binomial theorem.) See Fig. 2-
11. Pascals Triangle is formed by creating a triangle with 1s on the outside,
where each number inside the triangle is equal to the sum of the two numbers
immediately above it.
48
MATH WITHOUT NUMBERS
Rule 2: If an Element in the Set X is itself a Set, then such Element becomes
a Power Set when we take the Power Set of X, i.e., the Power Set operation
distributes to each Element of X. For example, take X = {A, B }, where A
= { a, b } and B = { c, d }. P(X) = { ,APS, BPS, APSBPS }={ ,{a, b, ab },
{c, d, cd },{ a, b, ab} Un { c, d, cd}} = {, a, b, ab,c, d, cd, ac, ad, acd,
bc, bd, bcd, abc, abd, abcd } = P(a, b, c, d) . We use the PS subscript to
emphasize the point that the Element is a Power Set, not a regular Set. (The
symbol Un represents Direct Product followed by Union. See
Subchaper 7.5 for a detailed explanation.)
Rule 3: We must treat Sets of Atoms (or Coatoms) and Power Sets as
49
WILLIAM S. VEATCH
different Domains, meaning that we can perform Operations that combine
Sets with Sets, or Powers Sets with Power Sets, but not Sets with Power Sets.
By Operations, we mean Union, Intersection, Symmetric Difference,
Complement, and Subtraction, all as discussed in Chapters 4 and 5.
When we convert a Set of Atoms to a Power Set, we refer to this as a
Power Set Expansion of a Set of Atoms. Conversely, when we move
from a Power Set to a Set of Atoms, we refer to this as a Power Set
Contraction. It is important to keep track of whether we are working with
Sets of Atoms or Power Sets, however, as we cannot add Sets of Atoms to
Power Sets and vice versa. As we will see in later Chapters, the ability to work
with Atoms (or Coatoms) and switch back and forth to Power Sets when we
need to, will considerably simplify working with the Mathematics of Ideas.
2.8. Three Worlds: Physical v. Ideas v. Abstract Sets
Another central theme in MWN is that there are three different worlds of
interest when we study the Mathematics of Ideas: the Physical World, the
World of Ideas, and the World of Abstract Sets. See Fig. 2-13. In the Physical
World, there are rules of physics and other laws of nature at work that limit
what is possible. We often feel that we are discovering rather than
creating, when we explore the Physical World.
In the World of Ideas, however, there are no such limitations. We can
imagine a purple cow or a flying pig, so it exists in our World of Ideas. At
first, at an intuitive level, it seems that we can create new elements in the
50
MATH WITHOUT NUMBERS
World of Ideas. In reality, however, all possible combinations of Atomic
Ideas are already present in the Universe consisting of the Power Set of
Atomic Ideas. (This brings to mind the Bob Newhart routine: If an infinite
number of monkeys were sitting at an infinite number of typewriters typing
randomly, eventually they would create all of the great works of literature.
Seek it out if you have not heard it before!)
The World of Abstract Sets is once again entirely different from the
Physical World and the World of Ideas. In the World of Abstract Sets, we
have Objects, but the sole properties of an Object are: (i) its existence as an
Atom or Compound Idea, and (ii) its uniqueness, which leads to the ability
to differentiate any Object from any other Object by its unique Idea
Signature. The real power to using Abstract Sets comes from the realization
that we can map Attributes in the World of Ideas to Elements of the Power
Set of Coatoms in the World of Abstract Sets.
2.9. Update to The Definition of Idea
We now expand the definition of Idea as follows:
Objects and Attributes: Each Idea has a dual nature as a Set of Objects
and a Set of Attributes.
We can fully describe an Idea by a list of examples of the Idea, called
Objects.
We can fully describe an Idea by a list of properties or characteristics of
the Idea, called Attributes.
Objects in the Object Set (or List Set) are examples of the Idea (e.g., a Dog
is a Mammal).
Attributes in the Attribute Set (or Test Set) are characteristics or
properties exhibited by each Object in the Object-Set (e.g., an Animal can
move and therefore is Mobile).
We can view an Idea as a single Attribute, which forms a Partition of the
Universe of Ideas into two Equivalence Classes.
Inverse Relationship: Objects and Attributes are related to one another
in an inverse relationship, in that the more Attributes an Idea has, the fewer
the number of Objects in the Universe there are that have all of the required
Attributes. (Note that each Object or Attribute is itself an Idea that has its
51
WILLIAM S. VEATCH
own Object-Set and AttributeSet.)
Atom or Compound: An Idea is either an Atom (or Deemed Atom) or
a Compound made up of a Set of Atoms. Dually, an Idea is either a Coatom
(or Deemed Coatom) or a Compound made up of a Set of Coatoms.
Power Set: The Power Set of Atoms represents all Atoms and all
possible Compound Ideas that we can form from the Set of Atoms. Dually,
the Power Set of Coatoms represents all Coatoms and all possible Compound
Ideas that we can form from the Set of Coatoms.
2.10. References, Historical Notes, and Further Reading
General
[Rodgers 2000]: In Learning to Reason An Introduction to Logic, Sets, and
Relations, Nancy Rodgers provides a very clear foundation in set theory and
logic. This book is highly recommended for anyone wishing to brush up on
the basics of set theory and logic.
Objects v. Attributes; Extension v. Intension/Comprehension
[Aristotle 350 B.C.E.]: Aristotles Categories is highly recommended
reading early on in the study of the Mathematics of Ideas, as Aristotles
writings have been the primary influence on philosophers and
mathematicians in this field for more than 2,000 years.
[Porphyry 260 C.E.]: This ancient work by Porphyry called the
Introduction, discusses the five items: genera, species, differences, properties,
and accidents, and was written as an introduction to the study of Logic and
Philosophy.
[Frisch 1969 ]: Extension and Comprehension in Logic, by Joseph Frisch,
provides a detailed historical perspective, from the time of Aristotle to the
near present, of the terms extension and comprehension, which are
similar to, but not the same as, our MWN terms Object and Attribute.
Reading such a detailed survey of the use of these terms will give the reader
a better understanding of the nuances of meaning and the complexity of these
concepts, as well as a sense of the philosophers desire throughout time to
understand the nature of Ideas.
52
MATH WITHOUT NUMBERS
Atoms and Coatoms
We borrow the Atom and Coatom terminology from the field of
Lattice Theory. To learn more about Atoms and Coatoms in Lattices, see
[Roman 2008] at 7, and [Ganter and Wille 1999] at 7.
53
WILLIAM S. VEATCH
3. SET THEORY AS APPLIED TO SETS OF IDEAS
In this Chapter, we explore what we mean by Set Theory, and in particular
Set Theory as applied to the Mathematics of Ideas. We explore in more detail
what we mean by an Idea, and how Ideas (either Atoms or Compounds)
combine to form new Ideas. While the Physical World is somewhat relevant
as a source of Ideas, our primary focus is on the World of Ideas and the
World of Abstract Sets. In particular, we are interested in how we can map
Ideas (in the World of Ideas) to Abstract Sets (in the World of Abstract Sets).
Most readers will have some understanding of Set Theory already. Our
view of Set Theory for purposes of the Mathematics of Ideas is similar to the
traditional view, except in the following areas:
We must exercise care to distinguish among the three Worlds: the
Physical World, the World of Ideas, and the World of Abstract Sets,
as Sets behave differently in each of these three Worlds. (See Fig. 3-
1 and previous discussion in Section 2.8.)
We assume as a starting point, that a Universe of all Ideas exists.
We must exercise care to distinguish among the different Domains
that are relevant to the Mathematics of Ideas, including:
o Objects v. Attributes, and
o Sets v. Power Sets.
Keeping these points in mind, we will look at the definition of a Set,
including the concepts of a Universal Set and an Empty Set, and then look at
several dual properties of Sets of Ideas, including:
54
MATH WITHOUT NUMBERS
Finite v. infinite,
Discrete v. continuous, and
Subdivision v. consolidation.
In this way, we will develop a solid foundation in Set Theory for our
Mathematics of Ideas.
3.1. The Definition of a Set
In traditional Set Theory, a Set is a collection of objects which we refer to as
Elements of the Set. The Elements in the Set could be anything from
abstract objects in the World of Abstract Sets, ideas in the World of Ideas, to
tangible objects in the Physical World. Traditionally, a capital letter would be
used to indicate an abstract Set, and curly brackets would be used to encase
the Elements of the Set where the order of the Elements is not important:
X = { a, b, c }. In the World of Ideas, we could use words as Elements of
55
WILLIAM S. VEATCH
Sets. For example, we could write: X = { animal, dog, cat }. The principal
difference between the Physical World and the World of Ideas, is that objects
in the Physical World are limited by what we can measure and observe, and
are governed by the laws of nature, whereas in the World of Ideas, the only
limitation is our imagination.
In the World of Abstract Sets we have four principal types of Elements
or Domains that can form Sets:
Objects, which we define as Atoms or Sets of Atoms,
Attributes, which we define as Coatoms or Sets of Coatoms,
Power Sets of Objects/Atoms, or
Power Sets of Attributes/Coatoms.
There are other possible types of elements that could form Sets, such as
Power Sets of Power Sets (see Chapter 7), but the ones listed above are the
four principal types for our present purposes. Note that the four types of
Elements or Domains are related. See Fig. 3-2
56
MATH WITHOUT NUMBERS
Given any one Domain, we can generate the other three. Although in a
sense the four Domains are equivalent, each serves an important purpose:
Atoms: Atoms (or Deemed Atoms) are the smallest Ideas that for
all practical purposes cannot be broken down into smaller
component ideas. Objects in the World of Ideas map to Atoms or
Sets of Atoms in the World of Abstract Sets.
Power Set of Atoms: The Power Set of Atoms includes the Atoms
themselves, as well as all Compound Ideas that can be formed from
the Atoms.
Coatoms: Coatoms (or Deemed Coatoms) represent Sets of all
Atoms except one: 1=~a, 2=~b, etc. Attributes or properties in
the World of Ideas map to Coatoms or Sets of Coatoms in the
World of Abstract Sets.
Power Set of Coatoms: The Power Set of Coatoms includes the
Coatoms themselves, as well as all Compound Ideas that can be
formed from the Coatoms.
While in traditional Set Theory the Elements of Sets could be anything,
in the Mathematics of Ideas we must abide by three limitations or rules:
We cannot create Sets of Elements taken from different Worlds.
Therefore, we must keep track of whether we are working with Sets
in the Physical World, the World of Ideas, or the World of Abstract
Sets, although we may create a mapping from the World of Ideas to
the World of Abstract Sets.
We cannot create Sets of Elements taken from different Domains.
The four principal Domains are Objects, Attributes, Power Sets of
Objects, and Power Sets of Attributes, and Sets of Elements may be
created within each.
Each Element of a Set is unique; the same Atom cannot appear twice
in a Set.
Another way to think of our Universe of Ideas is like a dictionary, where
each Idea appears only one time. In fact, it would be most confusing if the
same Idea could appear in more than one place in the dictionary, with
different definitions. To communicate clearly, we need to assume that Ideas
and their meanings in our Universe of Ideas are unique and remain constant.
The real power of the Mathematics of Ideas is evident when we map
Ideas in the World of Ideas to Atoms, Coatoms, and Power Sets (of Atoms
57
WILLIAM S. VEATCH
and Coatoms) in the World of Abstract Sets. See Fig. 3-3.
Fig. 3-3 illustrates how a simple classification of Animals in the World
of Ideas maps to an Abstract Set in the World of Abstract Sets. The Elements
map as follows:
Cat a,
Dog b,
Mammal ab,
Snake c,
Vertebrate abc,
Crab d, and
Animal abcd.
Within the Universe of Abstract Sets we see that there are several
different equivalent ways to represent the same Sets, including a Hierarchy
Tree, a Nested Partition Equation, and a Nested Lattice. All of these are
helpful techniques for representing Sets of Ideas as Abstract Sets, that we will
discuss in more detail as we proceed through the Book.
58
MATH WITHOUT NUMBERS
3.2. The Universe of Ideas or Universal Set
Imagine a Set that contains every conceivable Idea. This Set is the Universal
Set, as it contains every Idea imaginable in our Universe of Ideas. In terms
of a diagram, we use a square with a single x labelled (a) representing a
single Object in the Universe. This single Object is itself a Set that includes
every possible Idea. See Fig. 3-4. There are no Attributes that are shared by
every Idea in the Universe, so the related Attribute Set is []. (We use
parentheses for Objects; and square brackets for Attributes.)
We can also create a Power Set from the Universal Set. Since the Universe
consists of a single Object or Set, we know that the Power Set will have 2n =
21 = 2 Elements. What is the second Element? The answer is the Empty
Set, which is considered in Set Theory to be an element of every Power Set.
See Fig. 3-5.
59
WILLIAM S. VEATCH
In this case, [1] represents all possible Attributes, including both the
true and false values of an Attribute. Since no Object can exhibit all
Attributes, such as both Red and ~Red, the Attribute [1] corresponds to the
Empty Set of Objects. [] represents the absence of any required
Attributes. Since no Attributes are required, all Objects in the Universe U
meet the test.
Does the Universal Set of Ideas really exist? If we view the Universe of
Ideas as infinite, then we quickly see that we will never achieve our goal of
writing down all of the Elements of the Universal Set. However, for our
present purposes, we choose to work with finite Sets, and the Universe
becomes the Universe of Ideas under discussion, or the Universe of
Discourse. Sometimes we write U for the infinite Universe or Universal
Set of all Ideas, and UD for a finite Universe of Discourse. The subscript
D emphasizes that we are talking about a finite Universe for discussion
purposes, and not really trying to talk about an infinite, innumerable
Universe. Since we are talking about a finite Universe of Discourse, for all
practical purposes we can safely assume that it exists.
In summary, we assume that the Universal Set of Ideas exists in the World
of Ideas, separate and apart from the Physical World and the World of
Abstract Sets.
60
MATH WITHOUT NUMBERS
Key Point: We assume that the Universal Set of all Ideas exists in the World of
Ideas, separate and apart from the Physical World and the World of Abstract Sets.
The Universal Set U encompasses every thought that we are capable of thinking.
3.3. The Empty Set
Picture an empty box. The set of objects in the box is the Empty Set, because
there is nothing in the box. This seems like a simple enough concept when
applied in the Physical World, but what about the world of abstract Ideas?
Does the Empty Set of Ideas really exist? It is difficult to think of
absolutely nothing. (Try it and see!) The Idea of thinking about nothing is
itself an Idea. For our purposes, we do not need to reach a final resolution
of this philosophical question. We assume for now that the Empty Set exists.
Alternatively, even if the Empty Set does not exist, as we remove Ideas from
a set we approach the Empty Set, even if we can never quite reach it in our
imagination.
In the World of Abstract Sets, the Empty Set has some interesting
properties that we will mention briefly now, and discuss later in more detail.
First, the Empty Set is considered to be an element of every Power Set.
Therefore, if we create a set X of Ideas where X = { a, b, c } = { Dog, Cat,
Elephant }, we need to remember that the Empty Set is an element of the
Power Set P(X) = (, a, b, c, ab, ac, bc, abc ).
Note, however, that the Empty Set is not an element of itself. So, ={}
is a true statement, whereas ={ } is not true, because { } has one
element and is therefore not empty.
Before considering particular Ideas as Object Sets and Attribute Sets, note
the important relationship between the Empty Set and the Universal Set. In
particular, we note the following:
= U ; the Complement of the Empty Set is the Universal Set.
See [Solomon 1990] at 21. By Complement we just mean that
and U are Disjoint, and together they constitute the entire
Universe of Ideas.
U = ; the complement of the Universal Set is the Empty Set.
See [Solomon 1990] at 21.
61
WILLIAM S. VEATCH
3.4. Finite v. Infinite
For purposes of our exploration of the Mathematics of Ideas, we take a dual
approach to the concepts of Finite and Infinite. On the one hand, when we
are working with particular Sets and mathematical structures such as Chains,
Lattices, and Power Sets, we assume that we are dealing with Finite Sets. This
simplifying assumption allows us to examine a fixed set of Ideas in order to
explore how these Ideas are related to each other without becoming
overwhelmed.
On the other hand, when choosing or creating the Finite Sets that we
want to explore, we begin with the assumption that Ideas are Infinite in
nature, always capable of further subdivision into smaller and smaller parts.
This Infinite nature of Ideas is something that we will return to from time to
time, as there are different types of Infinite sets that are relevant to Ideas.
We start out by selecting certain Finite Sets of Ideas that we want to
explore in mathematical terms. From time to time we reexamine those initial
Sets in light of the Infinite nature of Ideas. Upon reexamination, we may
need to further subdivide or consolidate Ideas, either at their outer limits or
along the path of a continuum, in order to express the Ideas or Knowledge
Representation structure that we are trying to articulate, discover, or create.
3.5. Discrete v. Continuous
At an intuitive level, we understand the difference between Discrete Objects
a,b,c,, or red, orange, green, blue, as opposed to Continuous spectrums,
such as the gradations of color in a rainbow, or the range of temperature
from 40 degrees below zero to 90 degrees above zero. See Fig. 3-6.
62
MATH WITHOUT NUMBERS
But do discrete Objects really exist? In the Physical World, they do, since
my dog, or the apple I ate for lunch, are clearly discrete Objects. Yet in the
World of Ideas, it becomes a much closer question. As we examine the
Attributes of two related but different Objects, we can imagine a continuum
of Objects between the two original Objects, with infinitesimal variations to
the point that what at first seemed to be discrete Objects now appear to be
points on a line of continuous variation. Such a continuum of Ideas is an
example of the Infinite nature of Ideas.
In any event, we do not need to answer this question now. We simply
assume that discrete Objects do in fact exist. As with any assumption,
however, we do need to reexamine the assumption from time to time to see
whether we still believe it to be valid, or whether we need to make some
adjustments. For example, if we need to further differentiate between the
colors Red and Orange, we could add a number of new Objects between
these original two, in order to articulate the slight gradations from Red-
Orange to Orange-Red.
3.6. Subdivision v. Consolidation
Another example of the infinite nature of Ideas is the Subdivision of Ideas
into smaller parts. The Idea of Animal can be divided into a wealth of
different Animal types, one of which is Mammals. Mammals can be
further subdivided into types of Mammals, one of which is Dogs. Dogs
can be further divided into breeds of Dogs. Ultimately, we could further
subdivide our sets into particular Dogs that were alive in the past, alive today,
or will be alive in the future; and we could differentiate based on how they
exist at any moment in time. At this point, it becomes clear that we can never
enumerate every Dog, let alone every Animal. By virtue of Subdivision,
63
WILLIAM S. VEATCH
we can view every Idea as a Compound Idea.
Yet another example of the infinite nature of Sets is the Consolidation
of Ideas. Consolidation is a principle that we will return to from time to time.
It means that any two Sets can be consolidated to form a new Set consisting
of the Union of the Elements in the original two Sets. If Ideas can be
subdivided in an infinite number of ways, then clearly there is also an infinite
number of ways to combine and consolidate the subdivided Ideas. In fact,
we know from our study of Power Sets with 2n Elements, that each time we
add a new Atom, for example, through Subdivision, we double the size of
the Power Set! Ultimately, by virtue of Consolidation, we can view every
Idea as a Deemed Atom if we consolidate all of its Elements.
See Fig. 3-7. In this Figure, we illustrate a Universe with four Atoms: X
= { a,b,c,d }. We start with a diagram of U without any Subdivisions. Then,
we subdivide U into two Sets: A = { a, b, c } and { d }. Next, we further
subdivide A = { a, b, c } into two Sets: B = { a, b } and { c }. For each
subdivision, we look at four different ways of representing the subdivision:
Partition Equation,
Hierarchy Tree,
Simplified Power Set Lattice, and
Power Set Lattice showing all Atoms.
It is not necessary to understand these diagrams in detail for the moment;
we will review them in detail in subsequent Chapters. For now, it is just
helpful to get a feel for the different ways that we can represent Subdivisions
of Ideas. Consolidations are basically the same as Subdivisions, but in
reverse. Whereas Fig. 3-7 illustrates Subdivisions moving from the top of
the diagram to the bottom, it illustrates Consolidations moving from the
bottom to the top.
64
MATH WITHOUT NUMBERS
As a practical matter, we tend to select Atoms or Deemed Atoms when
studying a subject, but often we use the technique of splitting Atoms and
examining new combinations in the Power Set as a means of correcting or
improving upon our understanding of a subject.
Ultimately, if we consolidate all Atoms into a single Set, we end up with
the Universal Set U. On the other hand, if we subdivide ad infinitum, we
end up with an infinitely large set of Atoms, but where the distinctions
between Atoms are so miniscule as to be worthless in any practical sense.
As a result, our basic methodology is to start with the Universal Set U,
and subdivide it into different Partitions and Sub-Partitions that are
meaningful to us. The complicating factor is that each time we create a new
Atom by subdivision, it doubles the number of Elements in the Power Set.
Still, that is the nature of Ideas and is not something that we can change. We
can, however, develop techniques such as Partition Equations (see Chapter
10), that help us whittle down the Elements to those that are meaningful.
65
WILLIAM S. VEATCH
3.7. Atoms v. Power Sets
We have considered the three worlds relevant to our study of Ideas:
The Physical World,
The World of Ideas, and
The World of Abstract Sets.
The Physical World is characterized by the fact that we describe it in
terms of Sets that we determine by observation, measurement, and scientific
experimentation (e.g., to determine cause and effect). The World of Ideas
differs because we use Sets to describe the Physical World, but then we move
on to describe things that either could exist in the Physical World, but do not
yet, or that are so fantastic that they exist only in our mind. The World of
Ideas is limited only by our imagination. The World of Abstract Sets is
something bare and bland in comparison.
In the World of Abstract Sets, we assume that Atoms labeled a, b, c,
have no characteristics other than existence and uniqueness, making
them identifiable. We can distinguish one Atom from another, but in MWN
we assume that Atoms do not have any other properties, such as color, size,
shape, etc. Although composed of naked Atoms, the World of Abstract Sets
is, arguably, the most powerful of the three Worlds. Sets in the World of
Abstract Sets behave with certainty and in a mathematically precise manner.
By mapping Sets of Objects in the Physical World and the World of Ideas to
Sets in the World of Abstract Sets, we can explain how we think with
mathematical precision. The fact that we can map in a many-to-one, or
perhaps one-to-one, relationship, Attributes in the World of Ideas to
Elements of the Power Set of Coatoms in the World of Abstract Sets,
provides us with a blueprint for how to make our thinking mathematically
precise, with the ultimate goal of improving communication and education.
We could, in fact, map the entire Universe of Ideas in a systematic way.
In the World of Abstract Sets, there is an important distinction between
Atoms and Power Sets of Atoms. The Atoms (or Deemed Atoms) form the
basic building blocks, and the Compound Elements are the more complex
Ideas formed from Sets of Atoms. Given n Atoms, there are always 2n
Elements in the Power Set, but this is not to say that all Elements in the
Power Set are relevant. In fact, most of the combinations will be
meaningless. We avoid becoming overwhelmed by meaningless
combinations of Atoms by working with Partitions of Sets of Atoms, which
66
MATH WITHOUT NUMBERS
we can transform into Power Sets, and then back to Sets of Atoms, at will.
In this way, the meaningless combinations are hidden from view, although
they are still present. This technique allows us to work with Power Sets,
which exhibit amazing properties of Boolean Algebras, Lattices, Rings, and
Topologies (see Chapter 7), without becoming overwhelmed by the number
of Elements. The other possible approach is to try to identify subsets of
Power Sets that describe how Ideas work, but the disadvantage to that
approach is that we inevitably lose the benefit of many of the unique and
sometimes startling properties of Power Sets. Keeping the Power Sets intact,
but finding ways to hide the unwanted Elements from view has its
advantages.
This brings us to the main point of this Section of the Book, namely,
that it is important to understand that within the World of Abstract Sets,
there are in fact three Domains that are relevant to our study of the
Mathematics of Ideas:
Domain 1: Atoms, X = { a, b, c },
Domain 2: Power Sets of Atoms, P(X) = (, a, b, c, ab, ac, bc,
abc ) PS, and
Domain 3: Power Sets of Power Sets of Atoms, P(P(X)) = (, a, b,
c, (a, b, ab), (a, c, ac), (b, c, bc), (a, b, c, ab, ac, bc, abc))DPS.
Note the use of the subscript PS for Power Set, and DPS for Double
Power Set.
Sometimes we speak of the of the Domain of Atoms as being
Dimension 1; the Domain of Power Sets of Atoms as being Dimension
2; and the Domain Power Sets of Power Sets of Atoms as being
Dimension 3.
Joining the three Domains in a one-to-one relationship, are a number
of rules that allow us to transform Sets and Partitions of Atoms into Power
Sets and Nested Power Sets. See Table 3-1.
67
WILLIAM S. VEATCH
(Do not worry if this last example is hard to follow. We will discuss
nested Power Sets in detail in Chapter 7. For now, we just want to introduce
the concept.)
68
MATH WITHOUT NUMBERS
You might ask whether it is necessary to convert Atoms to Power Sets;
could we simply use Sets of Atoms and Partitions? The answer is that
Partitions of Sets of Atoms are not unique; there are many possible Partitions
that are meaningful to us in terms of the mappings from the World of Ideas
to the World of Abstract Sets. Transforming these Partitions of Atoms into
Power Sets shows how the different Partitions are related to one another.
This will become clear in Chapter 7.
Observations:
We may convert any Set of Atoms into a Power Set.
We may convert any Power Set into a Set of Atoms.
We may convert any Partition of Atoms into a nested Power Set Lattice; and vice versa.
3.8. Simplifying Assumption: Finite Universe of Discourse with
Discrete Atoms
We have seen how the World of Ideas can be interpreted as a Set of n
Atoms forming the building blocks of a Universe of Ideas consisting of U
and at opposite extremes, and in between, the remaining 2n Elements of
the Power Set.
We have seen that in the World of Abstract Sets, in theory we can
Subdivide Atoms ad infinitum, which in turn through Consolidation gives rise
to a Power Set that is growing at an exponential rate. We have also seen that
many, if not all, Ideas form a Continuum rather than clearly defined Discrete
Ideas.
In the World of Ideas, however, there comes a point where further
subdivision of Atoms has little practical benefit. At that point we stop, and
treat our Set of Objects as Deemed Atoms. Similarly, while the boundaries
between related Ideas may be fuzzy in some cases forming a continuum
without Discrete boundaries, in the World of Ideas it makes sense to treat
many Ideas as Discrete Objects, rather than a Continuum. As a result, going
forward we will assume that our Universe of Ideas is a Finite Universe of
Discourse (UD) composed of Ideas that are either Discrete Atoms (really
Deemed Atoms) or Compounds made up of Discrete Atoms.
Having said this, we need to remember to revisit our choice of Deemed
Atoms from time to time, and update it as necessary to describe more
accurately what we observe or measure in the Physical World, or imagine in
69
WILLIAM S. VEATCH
the World of Ideas.
3.9. Update to The Definition of Idea
We now expand the definition of Idea as follows:
U and : We assume the existence of both the Universal Set of all Ideas,
and the Empty Set containing no Ideas.
Finite v. Infinite; Discrete v. Continuous: Ideas appear to be Infinite
and Continuous by their nature, but are easier to work with in mathematical
terms if we treat them as Finite and Discrete.
Subdivision and Consolidation: In theory, Ideas are capable of
Subdivision into more and more Atoms ad infinitum, as well as Consolidation
into any Element of the Power Set generated by such Atoms.
Domains of X, P(X), and P(P(X)): Given a Set of Idea Atoms, there
are three Isomorphic Domains: X, the Power Set of X: P(X), and the Power
Set of Power Sets: P(P(X)).
Simplifying Assumption: In practice, we make a simplifying
assumption that the Universe of Ideas has a Finite number of Atoms which
we refer to as Deemed Atoms. If necessary, we can always revisit our choice
of Atoms and create a Finer Universe of Discourse with more Atoms, or a
Coarser Universe of Discourse with fewer Atoms.
3.10. References, Historical Notes, and Further Reading
Finite v. Infinite
[Tiles 1989]: In The Philosophy of Set Theory, Mary Tiles provides an in
depth discussion of many of the foundational issues of Set Theory, including
a discussion of finite v. infinite Sets in Chapter One at page 6 et seq. In
particular, there is a helpful discussion of Zenos Paradoxes at page 21 et seq.,
which highlight the issues arising from notions of infinity and continuums.
70
4. HOW OBJECT SETS OF IDEAS ARE RELATED
TO EACH OTHER
Before continuing to look at particular mathematical structures of Ideas, we
need to understand how Ideas are related to each other. In this Chapter 4,
we look at Object Sets, and then in Chapter 5 we examine Attribute Sets. In
Chapter 6, we step back and compare and contrast Objects and Attributes.
Understanding precisely how two or more Ideas are related to one another
can be extremely difficult to determine in practice. Although we have seen
that Ideas can be viewed as Sets of Objects and Sets of Attributes, this is not
how we ordinarily think of Ideas. We may use a Word in conversation or in
writing to represent a Compound Idea, without fully understanding what the
component Objects of the Set are, or what Attributes define the Idea. If we
make the effort, however, to define Ideas with precision, both in terms of
Object Atoms and Attribute Coatoms, then we can start to compare Ideas to
see if they have common Objects and/or common Attributes. We find that
Ideas can be related to one another in a finite number of ways, and that the
relationships among Ideas form familiar mathematical structures.
4.1. Ideas are Either Atoms or Compounds Made Up of Multiple
Atoms
We assume that discrete Atoms exist, where such Atoms are defined as Ideas
that cannot be broken down further into smaller component parts. As a
practical matter, we can deem certain Ideas to be Atoms, or more precisely
Deemed Atoms, and then treat them as if they were Atoms and not Sets
of component Ideas. If a Deemed Atom later turns out to be a Set, that is
okay, because our Mathematics of Ideas can easily handle this situation as
discussed in Section 3.6, and we will see again in Chapter 12 in the
discussion regarding Subdivision.
71
WILLIAM S. VEATCH
Once we have a Set of Atoms, the Atoms can be combined in various
ways to form Compound Ideas. For example:
{ Dog, Cat, Other Mammals } = Mammals Obj,
{ Red, Orange, Yellow, Green, } = Colors Obj,
{ Period, Comma, Colon, Semicolon, } = Punctuation Obj, and
{ Democracy, Dictatorship, Oligarchy, } = Forms of
Government Obj.
The Power Set formed by a Set of Atoms will contain all possible
Compound Ideas that could be formed from these Atoms, because by
definition the Power Set contains all possible combinations. This is not to
say that all combinations will be relevant or meaningful. Our challenge is to
ascertain which combinations are relevant and which are not. As a general
rule, if a combination is relevant, then it is likely that we have already given it
a name!
In light of the potential for Subdivision and Consolidation (see Section
3.6 and Chapter 12), in theory every Idea can be both a Deemed Atom and
a Compound Idea. See Fig. 4-1.
Once we have specified a Universe of Discourse (UD), we have for the
moment specified which Objects are Atoms and which are Compounds. We
can always change the UD later by Subdivision, Consolidation, or by adding
new Atoms.
72
MATH WITHOUT NUMBERS
4.2. Object Power Set: All Possible Combinations of Atoms - Forms
a Lattice Ordered by Inclusion
In the previous Section, we touched on how every Idea is either an Atom
(including Deemed Atoms) or a Compound Idea made up of Atoms. We
also know that given a Set of Atoms, we can create a Power Set consisting of
all possible subsets. In this Section we touch on the concept of Order,
and how the Elements of a Power Set form a Lattice that is ordered by
inclusion.
To give an example of order by inclusion, consider the following Set:
( a, ab, abc, abcd ).
We say that this Set is ordered by inclusion because each Element is a
subset of the next Element, i.e., a ab, ab abc, and abc abcd, where
a b means ais a subset of ab or each element of a is an element
of ab. This ordered Set is referred to as a Chain. If you look at the four
Atom Power Set (P4) in Fig. 4-2, you will see that it is made up of a series
of overlapping Chains that start at the bottom of the Lattice and work their
way up to the top. (Note the that Empty Set, , is considered to be an
Element of every Power Set, although we do not always write it out.) Other
examples of Chains in P4 include:
( b, bc, bcd )
( a, ab )
( c, bc, abc )
73
WILLIAM S. VEATCH
4.3. The Five Ways in Which Two Object Sets May be Related to
Each Other
Suppose we have two Ideas represented as Sets A and B. As we know, there
are five possible ways in which these two Ideas could be related to one
another as Sets: (1) the Ideas may be exactly the same, i.e., Identical, (2) the
Ideas may be Disjoint, i.e., entirely different from one another, with no
overlapping elements, (3) the Ideas may overlap, containing some common
elements, i.e., Partially Overlapping or POL for short, (4) one Idea, A, may
be a subset of the other, B, i.e., Subset-Superset, or (5) Idea A may be a
Superset of B, i.e., Superset-Subset. See Fig. 4-3 for Euler Diagrams showing
these basic relationships.
74
MATH WITHOUT NUMBERS
We spend some time looking further at these relationships, because
confusion can be avoided if we understand precisely what types of Set
relationships we are talking about in any particular context.
Depending upon the circumstances, we will use Euler Diagrams, words,
set notation, Hierarchy Trees, and Hasse Lattice Diagrams to illustrate the
Set relationships. See, Fig. 4-4.
75
WILLIAM S. VEATCH
Later, in Volume 2, we will see that a Proposition as used in Classical
Logic is, in essence, describing the Set relationship between two Ideas.
We must remember to specify our Domain up-front. Just as we would
not try to add a number like 2 to an Idea like dog, we must be careful
not to try to combine Objects with Attributes. Rather, we combine Objects
with Objects, and Attributes with Attributes.
Similarly, we combine Atoms with Atoms, Coatoms with Coatoms,
Power Sets of Atoms with Power Sets of Atoms, and Power Sets of Coatoms
with Power Sets of Coatoms. We must take care not to mix Sets of Atoms
with Coatoms; Atoms with Power Sets of Atoms; or Coatoms with Power
Sets of Coatoms.
In this Chapter 4, we focus on the Domain consisting of Objects. In
Chapter 5 we turn to Attributes.
76
MATH WITHOUT NUMBERS
4.3.1. Identical Sets: The Same Set of Objects (and Attributes)
Assume that we are looking at Relationships among Sets within the same
Domain. If the Domain is the same, then two Sets with the same Objects
also have the same Attributes. For example, the Word Cat in English
means the same thing as the Word Chat in French. In set notation, if AObj
= {a,b,c} and BObj = {a,b,c}, then A = B, because the two Sets have the same
Objects. In this example the Sets A and B are exactly the same or Identical.
See Fig. 4-3, for an illustration using a Euler Diagram.
If the Domains of two Sets are different, then we cannot use Set
Operations to combine the Elements directly, unless we first convert one of
the Sets to a common Domain. For example, suppose we have the following
Sets:
A Obj = Cat Obj = { Siamese, Persian } = ( a, b ),
B Att = Cat Att = { Meows } = [ 4, 3 ],
C Obj = Dog Obj = { Lab, Dalmatian } = ( c, d ), and
D Att = Dog Att = { Barks } = [ 2, 1 ].
E Obj = Chat Obj = { Siamese, Persian } = ( a, b ), and
F Att = Chat Att = { Meows } = [ 4, 3 ].
See Fig. 4-5. We can apply Set Operations like Union and Intersection
to A, C, and E, or B, D, and F, but not other combinations, because they
have different Domains. Note that in the example, the Sets A and E are
Identical, because they have the same Objects; and Sets B and F are Identical,
because they have the same Attributes. Therefore, we know that the word
Cat in English and Chat in French refer to the same Idea and mean the
same thing.
77
WILLIAM S. VEATCH
4.3.2. Disjoint Sets
If two Object Sets do not have any overlapping elements, then we say that
they are Disjoint. Disjoint Sets typically arise when we create Partitions of
Sets. Note, however, that two Disjoint Object Sets may have Partially
Overlapping (POL) Attribute Sets.
For example, suppose we have the following Sets:
A Obj = Cat Obj = { Siamese, Persian } = ( a, b ),
B Att = Cat Att = { Meows } = [ 4, 3 ],
C Obj = Dog Obj = { Lab, Dalmatian } = ( c, d ), and
D Att = Dog Att = { Barks } = [ 2, 1 ].
We can compare A and C, or B and D, both of which are Disjoint, but
not other combinations, because they have different Domains. See Fig. 4-5.
Note that in the example, the Cat and Dog Object Sets A and C are Disjoint,
as well as the Attribute Sets B and D.
78
MATH WITHOUT NUMBERS
Interestingly, however, the individual Atoms form Disjoint Object Sets,
but have Partially Overlapping (POL) Attribute Sets:
Siamese = (a)[432]
Persian= (b)[431]
Lab = (c)[421]
Dalmatian= (d)[321]
How can this be? The reason is that while at first it may not appear that
Siamese and Persian Cats share much in common with a Lab Dog, in fact
they do have one common Attribute: none of them is a Dalmatian. As a
result, they all share the Attribute ~d=4.
4.3.3. Partially Overlapping Sets
Partially Overlapping Sets often arise when we create two different Partitions
of a Set of Atoms. For example, suppose we take the following Set of
Animals: X Obj = { Cow, Whale, Robin, Salmon } = { a, b, c, d }. We can
create two different Partitions either by sorting the Animals by whether they
live on land or in water, or by whether or not they are Mammals: X =
( Land + Water ) In ( Mammal + ~Mammal ) = ( ac + bd ) In ( ab + cd ),
where + stands for Union and In stands for Direct Product followed
by Intersection. The Set of Mammals (ab) and the Set of Animals that live
on land (ac) are Partially Overlapping Sets, because a cow (a) lives on land
and is a Mammal. Some Mammals live on land, but some do not; some
Animals are Mammals, but some are not. See Fig. 4-6.
79
WILLIAM S. VEATCH
4.3.4. Subset-Superset and Superset-Subset
Subset-Superset and Superset-Subset relationships are typically found in
Chains. As examples of Subsets, we note that Dog is a Subset of the Set of
Mammals. Dog + Other Mammals = the Set of Mammals. Since Mammals
are one type of Vertebrate, we note that Dogs + Other Mammals + Other
Vertebrates = Vertebrates.
In summary:
Mammals = { Dogs, Other Mammals },
Vertebrates = { Mammals, Other Vertebrates } = { Dogs,
Other Mammals, Other Vertebrates },
Animals = { Vertebrates, Other Animals } = { Dogs, Other
Mammals, Other Vertebrates, Other Animals }.
80
MATH WITHOUT NUMBERS
4.4. The Five Principal Set Operations Applied to Object Sets:
Union, Intersection, Symmetric Difference, Set Subtraction,
and Complement
There are five well-established Operations in Set Theory that we use
frequently in our study of the Mathematics of Ideas: Union ( ), Intersection
( ), Symmetric Difference ( ), Set Subtraction (-), and Complement ( x ).
In the following Subsections we examine each of these five Operations as
they relate to Object Sets. (In Chapter 5 we look at how the five Operations
apply to Attribute Sets.)
4.4.1.Union
By the Union of two Sets, we mean a new Set that includes all of the
elements that are contained within either of the original two sets. The Union
of two or more Sets is the Set formed by combining all of the Elements in
such Sets:
A B = {x| x A or x B }; or
x (A B) x A or x B.
We read the first example as: A Union B equals x such that x belongs
to A or x belongs to B, and the second example as: x belongs to A Union
B if and only if x belongs to A or x belongs to B. See [Rodgers 2000] at 233
for more information about the various Set Operations.
The application of the Union Operation to Sets of Ideas expressed as
Objects is fairly straightforward, as illustrated by the following examples:
Types of Government Obj = { Democracy, Dictatorship }
{ Oligarchy, Other Government Types }, and
Religion Obj = { Christian, Muslim, Jewish } { Hindu, Other
Religions }.
4.4.2. Intersection
By the Intersection of two Sets, we mean a new Set that includes only the
elements that are in common with, or contained within, each of the original
two sets. The Intersection of two of more Sets is the Set formed by those
Elements that are in all such Sets:
81
WILLIAM S. VEATCH
A B = {x| x A and x B }; or
x (A B) x A and x B.
What is the Intersection of the Idea Animal and the Idea Dog?
Written as words, the answer is not immediately obvious. If we write the
Idea Animal as a set of its component Objects, however, the answer
becomes clear. Animal becomes Animal List = { Dog, Cat, Cow, Horse,
Other Animals }. Now it is clear that the Intersection of Animal and Dog is
Dog, as Dog is the only element that is common to both Sets (assuming
that Dog includes all subcategories of Dogs).
Although this is a simple example, it illustrates a complex issue that we
must face when working with sets of Ideas. We typically use words to
represent Ideas without articulating what the Idea means in terms of Objects
(or Attributes for that matter). Therefore, we need to get into the habit of
breaking down Ideas into their component List Sets, in order to see how they
are related to each other.
Note that each of Intersection and Union distributes over the other:
A (B C) = (A B) (A C), and
A (B C) = (A B) (A C).
This will turn out to be an important concept later when we look in
more detail at mathematical structures such as Boolean Algebras and Rings.
In particular, we will use this principle of Distribution to take a Partition
Equation in the form:
X = ( A + B ) In ( C + D), and distribute the term ( C + D ) to form:
X = (A In ( C + D )) + (B In ( C + D )) = A:( C + D ) + B:( C + D ),
where + represents Union and In represents Direct Product followed
by Intersection. Also, we sometimes use a colon, :, to represent Direct
Product followed by Intersection, in order to simplify the notation.
To give some examples using Object Sets of Ideas, consider the
following:
If A = Russian Authors = { Dostoyevsky, Tolstoy, Antokolsky },
B = Novelists = { Dostoyevsky, Tolstoy, Cervantes, Shelley,
Dickens }, and C = Poets = { Antokolsky, Dickinson, Longfellow};
then:
82
MATH WITHOUT NUMBERS
A (B C) = { Dostoyevsky, Tolstoy, Antokolsky } = (A B)
(A C), and
A (B C) = { Dostoyevsky, Tolstoy, Antokolsky } = (A B)
(A C).
4.4.3. Symmetric Difference
The Symmetric Difference of two Sets can be defined in a variety of ways,
but the most helpful definition for our purposes is: the Union minus the
Intersection of the two Sets. To give some examples using Object Sets of
Ideas, consider the following:
If A = Russian Authors = { Dostoyevsky, Tolstoy, Antokolsky },
B = Novelists = { Dostoyevsky, Tolstoy, Cervantes, Shelley,
Dickens }, and C = Poets = { Antokolsky, Dickinson, Longfellow};
then:
A B = { Antokolsky, Cervantes, Shelley, Dickens }, and
A C = { Dostoyevsky, Tolstoy, Dickinson, Longfellow }.
Note that if the two Sets in question are Disjoint, then the Symmetric
Difference will be the same as the Union. This fact will be helpful later.
4.4.4. Set Subtraction
Set Subtraction is relatively straightforward, as it is similar to the subtraction
of numbers. We take away from the first Set, the elements that are
contained in the second Set.
A-B = {x| x A and x B }; or
x (A-B) x A and x B.
To give some examples using Object Sets of Ideas, consider the
following:
If A = Russian Authors = { Dostoyevsky, Tolstoy, Antokolsky },
B = Novelists = { Dostoyevsky, Tolstoy, Cervantes, Shelley,
Dickens }, and C = Poets = { Antokolsky, Dickinson, Longfellow};
83
WILLIAM S. VEATCH
then:
Russian Authors who are not Novelists: A B = { Antokolsky },
Novelists who are not Russian Authors: B A = { Cervantes,
Shelley, Dickens}.
4.4.5.Complement
Before discussing Attributes, we want to consider the Object Complement
Set, which is a list of examples of what something is not. When someone
asks us what we mean by the Idea of Dog, we may answer by saying that a
Dog is a Mammal that is not a Cat, Cow, Horse, Pig, or other Mammal. In
set theoretic terms, we are saying that a Dog is the Complement of Other
Mammal relative to the Set of Mammals. We use an _ mark to indicate
a Complement Set; and we use a subscript to indicate the Set relative to which
it is the Complement. In set theoretic notation, we can write a number of
statements:
If {Mammal} = {Dog, Other Mammal},
then:
{ Dog } Mammal = {Other Mammal}: The Complement of Dog
relative to Mammal, is Other Mammal.
{ Other Mammal } Mammal = Dog: The Complement of Other
Mammal relative to Mammal, is Dog.
{Dog} = {Mammal}\ {Other Mammal}, where the symbol \
is like a minus sign for the subtraction of Sets.
{Dog} = {Mammal} - {Other Mammal} = {Mammal}
{Human, Cat, Cow, Pig, Elephant}
Of course, we could also say that a Dog is not a Ham Sandwich, a
Mountain, or a Tree. This would also be a way of saying what we mean by a
Dog, by listing what it is not. However, in most circumstances it is more
meaningful to describe a Complement Set that is closely related to the Idea
that we are defining. Rather than listing everything in the Universe that a
Dog is not, we pick a closely related Set of which Dog is a part, in this case
Mammal, and list the Complement of Dog relative to the Set Mammal.
(Later we will give the Set Mammal a special name, the Cover of the
Idea of Dog.)
The Idea together with the other Elements of the Cover Set (the Object
Complement Set) equal the Cover Set. E.g., a Dog is a Mammal that is not
84
MATH WITHOUT NUMBERS
an Other Mammal = { Cat, Human }. In a sense, we are defining the
Idea by what it is not.
In order to define the Complement operation, we need to compare the
Set in question to a second Set, which is either the Universal Set or some
other Set. If the comparison Set is the Universal Set, then we refer to the
Complement as the Absolute Complement. Otherwise, we refer to the
Complement as the Relative Complement, relative to the Set we have
specified.
4.4.5.1. Absolute Complement
An Absolute Complement Set of a set X, written X , is the set of all objects
in the Universe that are not objects in X. For example, if U = { Dog, Cat,
Red Apple, George Washington, Lasagna, Other Ideas} and Y = { Dog,
Cat}, then Y = { Red Apple, George Washington, Lasagna, Other Ideas }.
Given the overwhelming number of Objects in the entire Universe, the
Absolute Complement is less useful than the Relative Complement.
4.4.5.2. Relative Complement
As mentioned above, sometimes it is useful to define a Complement Set of a
Set X with respect to some other set within the Universe, rather than the
entire Universe. In this case, we refer to the Complement Set as the relative
complement. For example, if U = {Dog, Cat, Red Apple, George
Washington, Lasagna, all other Ideas}, S = { Dog, Cat }, and T = {Cat},
then the complement of T relative to S is Ts = { Dog }. We use the
subscript S to include that the complement is relative to S, not the entire
Universe.
Note that the Union of T = {Cat} and T = {Dog} is S = {Dog, Cat}.
We write T T = S.
4.5. The Distributive Property of Union and Intersection
We mentioned earlier that each of Union and Intersection distributes over
the other. Now, we take a moment to look at a more complicated example
that will prove useful later. In Chapter 10, we will see that the following is a
common way of writing a Partition Equation:
X PE = (A+B+C) In (D+E) In (F+G)
85
WILLIAM S. VEATCH
In a Partition Equation, each term such as (A+B+C) is an Exhaustive,
Pairwise Disjoint Partition of the Base Set X. Technically, + is Union
and In is Direct Product followed by Intersection, so we can rewrite the
Partition Equation as follows:
X PE = (A B C) (D E) (F G)
As noted above, Union is Distributive over Intersection, and vice versa.
We can use this fact to create a Nested Partition Equation by taking the
following steps:
Step 1: Use the Associative Law to group together the last two terms in the
Partition Equation:
X PE = (A B C) ((D E) (F G))
Step 2: In the newly grouped section, use the Distributive Law to distribute
the third term over the second:
(A B C) ((D E) (F G))
= (A B C) ((D (F G)) (E (F G)))
Step 3: Now, use the Distributive Law to distribute the term in bold from
Step 2 over the first term, creating a Nested Partition Equation:
(A B C) ((D (F G)) (E (F G)))
=(A ((D (F G)) ((E (F G))) (B ((D (F G)) ((E
(F G))) (C ((D (F G)) ((E (F G)))
Step 4: Convert to + for Union, and In (or (:()) for Direct
Product followed by Intersection:
(A In ((D In (F+G)) + ((E In (F+G))) + (B In ((D In (F+G)) + ((E In
(F+G))) + (C In ((D In (F+G)) + ((E In (F+G)))
=A:(D:(F+G) + E:(F+G)) + B:(D:(F+G) + E:(F+G)) + C:(D:(F+G) +
E:(F+G)).
In summary, we have demonstrated that using the Associative Law and
Distributive Law for Union and Intersection we can rewrite any Partition
86
MATH WITHOUT NUMBERS
Equation as a Nested Partition Equation:
X PE = (A+B+C) In (D+E) In (F+G)
= A:(D:(F+G) + E:(F+G)) + B:(D:(F+G) + E:(F+G)) + C:(D:(F+G) +
E:(F+G)).
Now that we have established the rules for the Distributive property of
Union and Intersection, we will use + for Union and In or simply
( : ()) for Direct Product followed by Intersection when we write
out a Partition Equation. This will make it much easier to read the Partition
Equation. From the example above, we can see that Union and Intersection
Distribute or multiply out in the same way as Addition and
Multiplication when working with numbers. This will prove to be
extremely valuable in Chapter 11 regarding Partition Equations. (As
illustrated above, sometimes we use a colon : to indicate that there is a
Nested Partition, instead of using the In symbol.)
4.6. Update to the Definition of Idea
Five Set Relationships Among Objects: Any two Sets of Objects
within a specified Domain are related to each other in one of five ways:
Identical,
Disjoint,
Subset-Superset,
Superset-Subset, or
Partially Overlapping.
Within the Object Domain, we can create Partitions of the Atoms.
Five Principal Operations Between Objects: Within any specified
Object Power Set Domain, the five principal Operations of Union,
Intersection, Symmetric Difference, Set Subtraction, and Complement apply.
4.7. References, Historical Notes, and Further Reading
[Rodgers 2000]: The author discusses the distributive properties of
Union and Intersection at page 236 et seq.
See also, [Pinter 2014].
87
5. HOW ATTRIBUTE SETS OF IDEAS ARE
RELATED TO EACH OTHER
In this Chapter, we look in more detail at the nature of Attributes, and how
Sets of Attributes are related to one another. We saw in Chapter 2, how
Ideas have a dual nature as representing both Objects (examples of the
Idea) and Attributes (properties of the Idea), but what exactly is the
difference in mathematical terms between Objects and Attributes? In
Chapter 4 we saw how Objects form Sets that behave in accordance with
the normal rules of Set Theory, such as Union, Intersection, Symmetric
Difference, Set Subtraction, and Complement. Chapter 4 may have seemed
easy, but Chapter 5 will probably seem more difficult. The reason for this is
that we are not familiar with applying Set Operations to Attributes. As we
will see in this chapter, however, Attributes also follow the rules of Set
Theory, but in an inverse fashion as compared to Objects. If we remember
that Attributes are simply the inverse dual of Objects, then the rules that
apply to Attributes will be easy to remember.
5.1. Definition of a Coatom
We have seen that if we start with a Set X of Atoms, we can create a Power
Set, P(X) consisting of all of the subsets of X. If X has n Elements, then
P(X) will have 2n Elements, including X itself and the Empty Set. Whereas
the Atoms are typically placed at the bottom of a Lattice diagram, just above
the Empty Set, Coatoms make up the row at the top of the Lattice, just below
the Set X = (a,b,c,d). See Fig. 5-1.
88
MATH WITHOUT NUMBERS
If we relabel the Coatoms using numbers, then we can create what looks
like an upside Power Set formed by Coatoms. This Power Set of Coatoms
turns out to be extremely useful to us, because, when we map Attributes to
Coatoms, Coatom is synonymous with Attribute. By creating a Power
Set of Attributes, we can apply the same rules for Chains, Partitions, etc. to
Attributes that we apply to a Power Set of Objects. The inverse relationship,
or Duality, between Objects and Attributes becomes readily apparent.
How do we know, or can we prove, that Coatom is synonymous with
Attribute? The argument is as follows: An Attribute is a characteristic or
property exhibited by one or more of the Objects in a Universe of Ideas
UD. If our UD has a Base Set X with n Atoms, then there 2n Objects
in our UD. If we examine each Object in our World of Ideas to see whether
or not it has the property in question, the answer will be either yes or no
for each Object, thereby creating a Partition of UD with two Equivalence
Classes. (An Equivalence Class in this case is simply a Set of Objects that
share the same yes or no response to whether they exhibit a particular
property.) The number of yes answers will lie somewhere between 0 and
n, and the yes Objects will form one of the 2n subsets of the Power Set
of Atoms in X. We know from our definition of Power Set, however, that
every possible subset of Atoms of Set X is represented somewhere in the
Power Set of Atoms. This means that every possible Attribute corresponds
to an Element in the Power Set of Atoms. We also know that every Element
in the Power Set of Atoms maps to a corresponding Element in the Power
Set of Coatoms. In this way, we can map every Attribute to an Element in
89
WILLIAM S. VEATCH
the Power Set of Coatoms. (To the extent that we have Deemed Atoms
in our Power Set of Atoms, we will have corresponding Deemed Coatoms
in our Power Set of Coatoms.)
Also, we know from observation that in a Chain of Objects, as we move
down the Chain to smaller subsets, each subset inherits the Attributes of the
Supersets above it in the Chain. As we can see from Fig. 5-1, our dual Power
Sets of Object Atoms and Attribute Coatoms follow this rule as well. For
example, as we move down the Chain of Objects: (abcd, abc, ab, a), the
related Attributes increase reflecting the inheritance of Attributes higher up
in the Chain: [, 4, 43, 432]. Given the ability to map Attributes to the Power
Set of Coatoms, and the fact that the Power Set of Coatoms preserves and
helps us to visualize the Inheritance of Attributes, the Power Set of Coatoms
works as a representation of Attributes. This is proof enough for our
purposes.
5.2. Tracking Attributes in a Universe of Discourse
In the previous Section 5.1, we considered a four-Atom Universe, and
considered how we would label the Objects and Attributes. What happens
if, as is usually the case, we are working with a limited Universe of Discourse?
We will see no change to how we view Objects. However, in the case of
Attributes, it helps to include an extra Attribute as a placeholder for the
inherited Attributes relating to the rest of the Universe. Fig. 5-2 illustrates
how this works for Objects.
90
MATH WITHOUT NUMBERS
Note that adding an additional Atom, e, representing the rest of the
Universe consolidated into a single Atom, has no impact on the four-Atom
Universe of Discourse with the Base Set of Atoms = { a,b,c,d }, as compared
to the four-Atom Power Set in Fig. 5-1.
The result differs greatly for Attributes. Fig. 5-3 illustrates how the
Attributes in a Universe of Discourse are impacted by the excluded Elements.
Basically, each Attribute is preceded by a 5 reflecting the exclusion of the
Atom e from all Elements in the Universe of Discourse, as compared to
the four-Atom Power Set in Fig. 5-1.
91
WILLIAM S. VEATCH
(One other potential concern is that for small Universes of Discourse
with limited Atoms, the mapping of Attributes is many-to-one. We will
address the concern in Chapter 11, where we discuss various ways to
combine UDs. Generally speaking, a many-to-one mapping is a signal that
we need to expand the base of Atoms in our UD in order to draw finer
distinctions, and separate out the Attributes currently lumped together in the
many-to-one mapping.)
What this means is that when discussing a Universe of Discourse
representing less than the entire Universe, we will have n Atoms, but
n+1 Coatoms, where the extra Coatom represents the excluded portion of
the Universe. This will become important as we map Attributes, because we
typically do not map any Attributes to . For example, it makes a lot more
sense, intuitively, to have the number 5 at the top of the Universe of
Discourse in Fig. 5-3, than to have the Empty Set, . In summary, our
92
MATH WITHOUT NUMBERS
example four Atom Power Set (P4) Universe of Discourse has Atoms X =
( a,b,c,d ), but Coatoms Y = [ 5,4,3,2,1 ].
5.3. Attributes are Either Coatoms or Compounds Made Up of
Multiple Coatoms
Just as an Idea can be Atomic, made up of a single Object Atom or a
Compound made up of multiple Atoms, we can also view Ideas as Coatomic,
represented by a single Attribute Coatom or a Compound made up of
multiple Coatom Attributes. Although we have shown that each Attribute
creates a Partition of the Universe of Objects into two Equivalence Classes,
there is also a relationship among Attributes that follows similar rules to those
governing Objects.
5.4. Attribute Power Set: All Possible Combinations of Coatoms
In the previous Section, we touched on how every Idea is either a Coatom
(including Deemed Coatoms) or a Compound of multiple Coatoms. We also
know that given a Set of Coatoms, we can create a Power Set consisting of
all possible subsets. Just as the Elements of a Power Set of Objects form a
Lattice that is ordered by inclusion, the same is true of a Power Set of
Attributes.
To give an example of order by inclusion, consider the following Set:
[ 5, 54, 543, 5432 ].
We say that this Set is ordered by inclusion because each Element is a
subset of the next Element, i.e., 5 54, 54 543, and 543 5432, where
means is a subset of or is an element of. In fact, this ordered Set
is referred to as a Chain of Attributes. If you look at the P4 Power Set in Fig.
5-3, you will see that it is made up of a series of overlapping Chains that start
at the top of the Lattice and work their way down to the bottom. Other
examples of Chains in P4 include:
[51, 541, 5431],
[543, 5432], and
[54, 541, 5421].
93
WILLIAM S. VEATCH
5.5. The Five Ways in Which Two Attribute Sets May be Related
to Each Other
Suppose we have two Ideas represented as Sets A and B. As we know, there
are five possible ways these two Ideas could related to one another as Sets:
(1) the Ideas may be exactly the same, i.e., Identical, (2) the Ideas may be
Disjoint, i.e., entirely different from one another, with no overlapping
elements, (3) the Ideas may overlap, containing some common elements, i.e.,
Partially Overlapping or POL, (4) one Idea, A, may be a subset of the other,
B, i.e., Subset-Superset, or (5) Idea B may be a subset of A, i.e., Superset-
Subset.
In Chapter 4, our focus was on the Domain consisting of Objects. In
this Chapter 5 we turn to Attributes. We will see that we can combine Sets
of Attributes in the same five ways that we can combine Sets of Objects. The
critical difference between Objects and Attributes, however, is that in the
case of a UD that is less than the whole Universe, we have an extra Coatom.
As a result, we need to specify whether we are working with the entire
Universe (U) or a more limited Universe of Discourse (UD). Most of
our examples use a UD, because more often than not we are studying a subset
of the entire Universe of Ideas.
5.5.1. Identical Sets: The Same Set of Attributes (and Objects)
Assume that we are looking at Relationships among Sets within the same
Domain. Then, as discussed in Section 4.3.1, if the Domain is the same, the
two Sets with the same Attributes also have the same Objects.
5.5.2. Disjoint Sets
As discussed in Section 4.3.2, if two Object Sets do not have any overlapping
Elements, then we say that they are Disjoint. Disjoint Object Sets, however,
do not necessarily result in Disjoint Attribute Sets. For example, Dog and
Cat are disjoint Object Sets, but share many Attributes, such as the
following: they are Mammals (have hair/fur, nurse young, etc.), Vertebrates
(have a backbone), Animals (can move, etc.), and Living Things (consume
food, reproduce, etc.).
94
MATH WITHOUT NUMBERS
Looking at Fig. 5-4, we can see that, if we ignore Attribute 5 that is
common to the entire UD, then some combinations of Attributes such as [4]
and [3] are Disjoint, but the related Object Sets, (abc) and (abd) are Partially
Overlapping and not Disjoint. On the other hand, [43] and [21] are Disjoint,
as are the related Object Sets (ab) and (cd).
The better view is that the 5 should be considered. The 5
represents the Attributes that are shared by all Elements of the UD, i.e., in the
example in Fig. 5-4, all Objects are Mammals.
5.5.3. Partially Overlapping Sets
As we saw in Chapter 4, Partially Overlapping Sets often arise when we
create two different Partitions of a Set of Atoms. As we saw in the previous
Section on Disjoint Sets, however, sometimes two Object Sets will be
Disjoint, but their related Attribute Sets will be Partially Overlapping;
similarly, sometimes two Object Sets will be Partially Overlapping, but their
related Attribute Sets will be Disjoint. See Fig. 5-1, where Object Power Set
Elements (abc) and (abd) are Partially Overlapping, whereas the related
Attribute Power Set Elements [4] and [3] are Disjoint. This is the case,
95
WILLIAM S. VEATCH
because in Fig. 5-1 the Power Sets represent the entire Universe of Ideas. In
Fig. 5-5, however, we have a more limited 4-Atom Universe of Discourse,
so all Elements share the Attribute 5, which in this case refers to the
Attributes common to all Animals.
Tables 5-1, 5-2, 5-3 and 5-4 summarize these Disjoint and POL
relationships for the 15 (16 if we include ) possible Partitions of a four
Atom Set. Table 5-1 shows that Disjoint Object Partitions result in POL
Attribute Partitions, or sometimes Disjoint Attribute Partitions if working
with the entire Universe. Note that when working with the entire Universe,
a two Element Disjoint Object Set Partition always gives rise to a two
Element Disjoint Attribute Partition, whereas a three or four Element
Disjoint Object Partition always gives rise to a POL Attribute Partition.
96
MATH WITHOUT NUMBERS
Table 5-2 shows how Disjoint Attribute Partitions may result in
Disjoint or POL Object Relationships. This could happen if, for example,
Element d represents all other Atoms in U, and U = { a, b, c, d }.
97
WILLIAM S. VEATCH
Note that the above Table 5-2 excludes the Coatom 5, representing
the Attributes that all Elements in the UD have in common. This would be
appropriate, for example, if this were a Partition of the entire Universe, as
opposed to a UD. In a UD, there can never be a Disjoint Partition, because
all Elements have a common Coatom, 5 in this case.
As we would expect, Table 5-1 Disjoint Object Partitions and Table
5-2 Disjoint Attribute Partitions are inverses of one another, provided that
we are looking at the full Universe and not a more limited Universe of
Discourse.
What if two Sets, either Objects are Attributes, are POL? Can we say
with certainty what the other related Sets will be? See Tables 5-3 and 5-4.
98
MATH WITHOUT NUMBERS
5.5.4. Subset-Superset and Superset-Subset
As noted earlier in Chapter 4, Subset-Superset and Superset-Subset
relationships are typically found in Chains. This is true for Chains of Objects
and their related inverse Chains of Attributes. See Fig. 5-4. As discussed in
Chapter 7, we can also have Power Sets that are Subsets of larger Power
Sets.
99
WILLIAM S. VEATCH
As examples of subsets, we noted in Chapter 4 that Dog Obj = (a) is a
subset of the set of Mammals Obj = (ab), because (a) (ab). The result for
the same Ideas expressed as Attributes, however, is different. Dog Att = [432]
is a superset of the set of Mammals Att = [43], because [432] [43]. We still
have a Subset-Superset relationship, but in the Inverse order. Rather than
say that Dog Att is a Superset of Mammals Att, we often reverse the order and
say that Mammal Att is a property of Dog Att . Given the inverse relationship
of Objects and Attributes, this is in fact what we expect.
Table 5-5 provides some more examples based upon Fig. 5-6, that
illustrate the Inverse relationship.
100
MATH WITHOUT NUMBERS
In summary, for any two four-Atom Ideas expressed as Sets of Objects
and Sets of Attributes, we have the relationships set forth in Tables 5-6 and
5-7 below.
5.6. The Five Principal Set Operations Applied to Attribute Sets:
Union, Intersection, Symmetric Difference, Set Subtraction,
and Complement
As with Object Sets, the same five well-established Operations in Set Theory
that apply to Object Sets also apply to Attribute Sets: Union ( ),
Intersection ( ), Symmetric Difference ( ), Set Subtraction (-), and
101
WILLIAM S. VEATCH
Complement ( x ). The Operations, however, work in an Inverse fashion,
which is to be expected given the Duality and Inverse relationship between
Objects and Attributes. For the examples used in this Section 5.6, we refer
to Fig. 5-6.
5.6.1. Union
The application of the Union Operation to Sets of Ideas expressed as Objects
is fairly straightforward, although Attributes behave in an Inverse manner as
illustrated in Table 5-8. The more Attributes we require an Object to have,
the fewer the number of Objects that will meet the test.
The above examples illustrate that because Ideas can be Objects or
Attributes, the Union of two Ideas can yield a different result depending upon
whether we are working with Objects or Attributes. It also illustrates why
we need to specify the Domain, i.e., whether we are working with Objects or
Attributes (or their Power Sets). Sometimes the Domain is clear from the
context. If not, we can use the subscripts Obj and Att to indicate the
Domain: Dog Att Mammal Att = Dog Att or (Dog Mammal = Dog)Att.
5.6.2. Intersection
As with the Union Operation, the application of the Intersection Operation
to Sets of Ideas expressed as Objects is fairly straightforward, although
Attributes behave in an Inverse manner as illustrated in Table 5-9. The more
Attributes we require an Object to have, the fewer Objects will meet the test.
102
MATH WITHOUT NUMBERS
Note the inverse results for Objects and Attributes:
Objects: Dog Mammal = Dog, and
Attributes: Dog Mammal = Mammal.
We noted that Intersection was easy to follow with Objects once we
wrote out the Ideas as Sets. The same is true of Attributes. The result of
(Vertebrate Dog)Att may not be clear at first, but [54] [5432] is
clearly [54] = Vertebrate.
5.6.3. Symmetric Difference
As with Object Sets, we define the Symmetric Difference of two
Attribute Sets as the Union minus the Intersection of the two Sets.
To give some examples using Attribute Sets of Ideas, consider the
examples in Table 5-10.
103
WILLIAM S. VEATCH
As we gain more experience working with Set Operations as applied to
Objects and Attributes, we see that it is much easier to apply the Operations
to the Sets expressed as letters for Objects and numbers for Attributes, than
it is to apply the Operations to words.
5.6.4. Set Subtraction
Set subtraction is relatively straightforward for Attributes as well as Objects.
The results, however, in our Dog-Animal Power Set examples in Fig. 5-6
may not be terribly helpful, because the Difference is typically outside of the
Universe of Discourse. The main point for our present purposes is that the
Operation is closed, meaning that the answer to the Subtraction equation
will be found somewhere in the Power Set, so long as we are looking at the
entire Universe of Ideas.
5.6.5.Complement
As with Object Sets, we can define a Complement Set for Attributes Sets.
For example, referring to Fig. 5-6:
If (Mammal = Vertebrate + ~Snake)Att, then:
Mammal - Vertebrate = [543] [54] = [3]= ~Snake.
At first, it may be difficult to interpret this last equation. What it is
saying, however, is that if we start with the Attributes of a Mammal (i.e.,
hair/fur, nurses young, Vertebrate, Animal, etc. ) and subtract out the
Attributes of a Vertebrate, what we are left with are the Attributes that
distinguish Mammals (i.e., Dog and Cat in Fig. 5-6) from other Vertebrates
104
MATH WITHOUT NUMBERS
(i.e., Snake in Fig. 5-6). Therefore, we are left with [3]= ~Snake. Given our
UD in Fig. 5-6, ~Snake could be a Dog, Cat, Crab, or any Object that is
not an Animal.
As with Object Sets, to define the Complement operation, we need
to compare the Set in question to a second Set, which is either the Universal
Set or some other Set. If the comparison Set is the Universal Set, then we
refer to the Complement as the Absolute Complement. Otherwise, we
refer to the Complement as the Relative Complement, relative to the Set
we have specified.
5.6.5.1. Absolute Complement
An Absolute Complement Set of an Attribute Set Y, written Y , is the set
of all Attributes in the Universe that are not Attributes included in Y. For
example, if U = { Dog, Cat, Red Apple, George Washington, Lasagna, Other
Ideas}Att and Y = { Dog, Cat } Att, then Y = { Red Apple, George
Washington, Lasagna, Other Ideas } Att.
5.6.5.2. Relative Complement
As with Object Sets, sometimes it is useful to define a Complement Set of a
Set Y with respect to some other set within the Universe, rather than the
entire Universe. In this case, we refer to the Complement Set as the relative
complement. For example, if U = {Dog, Cat, Red Apple, George
Washington, Lasagna, all other Ideas} Att, S = { Dog, Cat } Att, and T =
{Cat} Att, then the complement of T relative to S is T s = { Dog}. We use
the subscript S to mean that the complement is relative to S, not the entire
Universe.
Note that the Union of TAtt = {Cat}Att and T Att = {Dog}Att is SAtt =
{Dog, Cat}Att. We write ( T T = S)Att.
5.7. How Many Attributes Can an Idea Have?
We know that there is a many-to-one relationship (or perhaps one-to-one
relationship if we consider the Universe of all Ideas and eliminate
distinctions without a difference) between Attributes of an Idea and the
elements of a Power Set that include that Idea.
If we accept that all Attributes in the Universe map to the Power Set of
Coatoms, then if we have n Coatoms there will be 2n possible distinct
105
WILLIAM S. VEATCH
Attributes in the Universe. If more than one Attribute maps to a particular
Power Set Element, then either (i) there is a distinction without a difference
and the extra Attributes are irrelevant, or (ii) it is a warning sign that we need
to expand our Universe of Atoms. Ultimately, our map of Attributes to the
Power Set of Coatoms should be one-to-one in an ideal World.
We may ask the question: How many Attributes does an Idea have?
In the case of an Idea that is an Atom, this is equivalent to asking: How
many Elements in a Power Set contain a particular Atom? The answer is
2 n-1. The easiest way to visualize this is to add an Atom to an existing Power
Set. The result is to double the size of the Power Set. See Fig. 5-7.
Table 5-11 sets forth the number of Elements in a Power Set that
contain a Particular Atom, for Power Sets with two to six Atoms.
106
MATH WITHOUT NUMBERS
What if the Idea is a Compound? We simply treat it as if it were a
Deemed Atom and apply the same formula, except that we subtract the
number of Atoms in the Compound from the power of two. For example,
in a Power Set with five Atoms, the Element (de) will have 2n-m = 25-2 = 23
= 8 possible distinct Attributes, where n is the number of Atoms in the
Power Set and m is the number of Atoms in the Compound we are
examining. Alternatively, we can think of the Set of Elements that contain
Attributes as the Up-Set or Filter generated by the Compound Element,
(de) in this case. See Fig. 5-8.
107
WILLIAM S. VEATCH
5.8. Update to the Definition of Idea
Five Set Relationships Among Attributes: We can represent any
Attribute Idea as a Set of Coatoms in any one of several different Domains:
The Domain of Attributes/Coatoms; or
The Domain of the Power Set of Attributes/Coatoms.
Any two Sets within a specified Attribute Domain are related to each
other in one of five ways:
Identical,
Disjoint,
Subset-Superset,
Superset-Subset, or
Partially Overlapping.
Within the Attribute Domain, we can create Partitions of the Coatoms.
108
MATH WITHOUT NUMBERS
Five Principal Operations Between Attributes: Within any specified
Attribute Power Set Domain, the five principal Operations of Union,
Intersection, Symmetric Difference, Set Subtraction, and Complement apply.
5.9. References, Historical Notes, and Further Reading
[Rodgers 2000]: The author discusses the distributive properties of Union
and Intersection at page 236 et seq.
See also, [Pinter 2014].
109
WILLIAM S. VEATCH
6. SUMMARY OF THE WAYS IN WHICH OBJECTS
AND ATTRIBUTES ARE SIMILAR TO OR
DIFFERENT FROM EACH OTHER
In many respects, Objects and Attributes are very similar to each other. In
this Chapter we want to highlight both the similarities and the differences.
What we will discover is that Objects and Attributes behave in the same way
as each other in mathematical terms, except that one is the inverse of the
other. This is an example of Duality, which occurs surprisingly often in
Mathematics. In this case, however, it makes perfect sense considering how
we arrived at our definition of Attributes in the World of Abstract Sets, i.e.,
the Power Set of Attributes is the inverse of the Power Set of Objects.
6.1. Similarities Between Objects and Attributes
6.1.1. An Idea Can Be Described Completely By a List of
Objects (or Attributes)
We assume that an Idea can be described completely by a List of Objects that
are examples of the Idea in the World of Ideas. Much of what exists in the
World of Ideas corresponds to what we can observe, experience, or measure
in the Physical World, but the World of Ideas is not limited by the Physical
World. As seen in previous Chapters, we can map Idea Objects in the World
of Ideas to Atoms in the World of Abstract Sets.
While it may be difficult to prove empirically, since we cannot check every
Idea, after much experimentation it becomes apparent that we can also
describe Ideas completely by a list of their Attributes. Through
experimentation we found that the Power Set of Coatoms works in practice,
110
MATH WITHOUT NUMBERS
as a means of representing all possible Attributes. By defining Attributes as
entries in the Power Set of Coatoms, we can completely describe an Idea by
listing its Attributes.
One way to prove this is by looking at the Idea Signature of an Idea, which
consists of the List of the Atoms and Coatoms relating to that Idea. Once
we specify a Set of Object Atoms that represents an Idea, the Set of Attribute
Coatoms is equal to the negative of the Complement Set. For example, in a
UD = { a, b, c, d }, if we create an Object Set A = { a, b } = (ab), then the
Complement Set of A is A = { c, d } and the negative of the Complement
Set is B = { ~d, ~c } = [43]. Therefore, the Idea Signature for our Ideas is
(ab)[43]. This demonstrates that if we know the Universe of Discourse, given
a Set of Atoms we can calculate the related Coatoms, and vice versa, given a
Set of Coatoms we can calculate the related Atoms.
Without a Universe of Ideas, however, we would not have a Complement
Set. By starting with the assumption that the Universe of Ideas exists, we
have implicitly stated that for any Set of Atoms X = { a, b, c,}, there is a
Complement Set X such that U = X + X . Just as we can map Idea
Objects in the World of Ideas to Atoms in the World of Abstract Sets, we
can map Idea Attributes in the World of Ideas to Coatoms in the World of
Abstract Sets.
6.1.2. Each of Objects and Attributes Can Constitute the
Elements of Sets
As we discussed in Chapter 3, under traditional Set Theory the Elements of
Sets can be anything. As a result, both Objects and Attributes (or properties)
in the World of Ideas, as well as their Atom and Coatom counterparts in the
World of Abstract Sets, can be viewed as Elements that form Sets. We saw,
however, in Chapters 4 and 5 that the Set Operations of Union, Intersection,
Symmetric Difference, Set Subtraction, and Complement behave differently
on the same Ideas, depending upon whether the Ideas are expressed as
Objects or Attributes. Therefore, we concluded that Objects and Attributes,
as well as their respective Power Sets, form separate Domains. We can apply
Set Operations to the Sets of Elements, but we must exercise care to form
Sets only from Elements within a single Domain. We cannot mix and match
and combine Elements from different Domains.
111
WILLIAM S. VEATCH
6.1.3. The Five Possible Set Relationships Between any Two
Sets Govern the Relationships Among Objects, As Well
As Attributes
We demonstrated in Chapters 4 and 5, that any two Sets of Objects or any
two Sets of Attributes will necessarily be related to each other in one of five
ways: Identical, Disjoint, Partially Overlapping (POL), Subset-Superset, or
Superset-Subset. This is true whether the Set Elements are Objects or
Attributes. We simply need to remember not to mix Elements from different
Domains.
6.1.4.The Five Principal Set Operations Apply to Both Sets of
Objects and Sets of Attributes
As we saw in Chapter 4 (Objects) and Chapter 5 (Attributes), the five
principal Set Operations of Union, Intersection, Symmetric Difference, Set
Subtraction, and Complement apply to both Object Sets and Attribute Sets.
The Set Operations work exactly the same way for Object Sets and Attribute
Sets. The outcome expressed as an Idea, however, differs as illustrated in the
following example where UD = XObj = { a, b, c, d } = { Dog, Cat, Snake,
Crab }:
Objects: Dog Mammal = Dog; (a) (ab) = (a), and
Attributes: Dog Mammal = Mammal; [432] [43] = [43].
The rules of Set Theory remain constant, but once again we see the
importance of specifying whether we are working with the Domain of Sets
of Objects or Sets of Attributes.
6.1.5. Specifying An Object Set or An Attribute Set Creates a
Partition of the Universe of Ideas
We must begin our analysis with the critical assumption that the Universe
(U) of Ideas exists, as represented by the Object Set XObj = (a), and the
Attribute Set YAtt = []. If we start with this assumption, then when we
specify an Object Set X (or an Attribute Set Y), we have in fact specified
two Sets, i.e., X and its Complement ~X (or Y and its Complement ~Y). In
other words, specifying an Object Set (or Attribute Set) creates a Partition of
the Universe into two parts which we call Equivalence Classes. If we had not
specified the Universe up-front, then a Partition would not have been
created, and we would lose the dual Power Set relationship between Objects
and Attributes.
112
MATH WITHOUT NUMBERS
In the World of Abstract Sets, Partitions play a very special role.
Partitions are equivalent to a special type of Relation called an Equivalence
Relation. This means that there is great mathematical significance to the fact
that an Idea creates a partition. Since Partitions are also Equivalence
Relations, the rules governing Equivalence Relations apply to Ideas.
In mathematics, we often write the Equivalence Relation in the following
format: xRy, meaning that x and y both have the Attribute R, or
alternatively, that x and y are in the same Partition. Here are some of
the well-established rules governing how Equivalence Relations work:
Reflexive Property: aRa (Every element is related to itself.)
Transitive Property: If aRb and bRc, then aRc. (Any middle
term in a relation chain can be eliminated.)
Symmetric Property: If aRb, then bRa. (The order does not
matter.)
See generally, [Rodgers 2000] at 326 et seq.
6.1.6. One Fully Determines the Other: the Idea Signature
As we saw in Section 6.1.2, once we specify a Set of Objects/Atoms (or
Attributes/Coatoms), then we can calculate the corresponding Set of
Attributes/Coatoms (or Objects/Atoms). The only requirement is that we
must assume that we are working with the Universe of Ideas, or a smaller
Universe of Discourse if one has been specified.
Idea Signature: Every Idea has an Idea Signature consisting of a list of Atoms
included in the Idea: X=(a,b,c); and a list of Coatoms representing Attributes
included in the Idea Y = [ 6,5,4], where 4 = ~d, 5=~e etc. In a six Atom UD,
we write the Idea Signature as: (abc)[654], for an Idea with Atoms a, b, and c.
6.1.7. Objects and Attributes in the World of Ideas Map to
Power Sets of Atoms and Coatoms in the World of
Abstract Sets
Every Idea Objectin the World of Ideas maps to an Element of the Power
Set of Atoms in the World of Abstract Sets. We know this is true, because
every Idea is either an Atom or a Compound consisting of a Set of Atoms,
113
WILLIAM S. VEATCH
and the Power Set is the Set of all possible subsets. Therefore, every Idea
expressed as a Set of Objects must correspond to some Element of the Power
Set of Atoms. Similarly, the same logic demonstrates that every Idea
expressed as a Set of Attributes must correspond to some Element of the
Power Set of Coatoms.
6.2. Differences Between Objects and Attributes
6.2.1.Examples v. Properties
In the World of Ideas, Objects are examples of an Idea, and we say that an
Object is a or is an example of a given Idea. In the case of Attributes,
however, we say that the Attribute is a property or characteristic of the
Idea, or that an Attribute is an element of the test for a particular Idea. For
example, we have the following:
Object: A Lab is a Dog.
Attribute: Mammal is an element of the test for a Dog.
Mammal is a property of a Dog.
In an Object Chain or Lattice of Atoms with 1 at the top, and 0 at
the bottom, the Objects for any given Idea come from below in the Power
Set of Atoms. In an Attribute Chain or Lattice of Coatoms with 0 at the
top, and 1 at the bottom, the Attributes for any given Idea come from
above in the Power Set of Coatoms
6.2.2. Equivalence Classes/Sets v. Partitions
Another way to interpret Objects and Attributes is to think of an Attribute
as an Equivalence Relation evidenced by a circle in a Euler Diagram, and an
Object as an Equivalence Class created by the Equivalence Relation. In Fig.
6-1, we have a typical Euler Diagram showing a Universe of Ideas that has
been sorted and divided into various categories. In visual terms, the
Attributes are the circles themselves, that form partitions of the Universe.
Each Attribute divides the Universe into two parts. The Objects that are
trapped within the partition are examples of the Attribute that created the
Partition. In Fig. 6-1, the circle labeled Animal is an Attribute or partition,
that divides that Universe into Objects that are Animals, which are placed
inside the circle and labeled a through h; and Objects that are not
Animals, which are placed outside the circle and labeled i and j.
114
MATH WITHOUT NUMBERS
It is well-established in Mathematics that given an Equivalence Relation
R on a Set S, the set of all Equivalence Classes of R is a Partition of S. See
[Rodgers 2000] at 331 for an example of a proof to this effect.
6.2.3. Inverse Relationship
The most significant difference between Objects and Attributes is that they
are related to one another in an Inverse relationship. We can see this most
clearly in the Chain table that we have examined a number of times in
previous Chapters. See Table 6-1.
115
WILLIAM S. VEATCH
In particular, we note that the more Attributes we require an Object to
have, the fewer Objects will meet the test.
6.3. Duality
We have encountered Duality a number of times in previous Chapters, but
we have not yet given a clear definition. By Duality, in the context of
Ideas, we mean that Ideas exhibit the following properties or Attributes:
Ideas have two parts, like two sides of the same coin, namely Objects
and Attributes,
The Objects and Attributes are related to one another in an Inverse
fashion, and
If we have a formula using Dual Concepts (e.g., Object and Attribute;
Union and Intersection; and U; or and ) that applies to
Ideas, then we can create a dual formula that is also true, by swapping
each dual element, i.e., Object becomes Attribute, Attribute becomes
Object, Union becomes Intersection, etc.
See generally, [Abbott 1969] at 57 for an excellent discussion of Duality.
See also, [Roman 2008] at 12 for a mathematical definition of the Duality
Principle for Posets (i.e., Partially Ordered Sets), and [Ganter and Wille 1999]
at 4. In mathematics, there is also an area of study called Galois
Connections that examines this type of Duality in great detail. See [Ganter
and Wille 1999] at 11.
116
MATH WITHOUT NUMBERS
Table 6-1 illustrates the Chain Duality, and Fig. 6-2 illustrates Power Set
Duality.
In addition to Chains (see Chapter 8) and Power Sets (see Chapter 7),
there are other examples of Duality, including the following:
and in Ordered Sets See [Caspard, Leclerc, and Monjardet
2012] at 10), and [Davey and Priestley 2001] at 15.
Union and Intersection as applied to Elements of a single
Power Set, provided that we also substitute for U, and vice
versa. See [Stoll 1963] at 19.
Chains and Partitions (sometimes called Antichains). See Chapters
8 and 9.
Lattices the dual of a statement about lattices phrased in terms of
and is obtained simply by interchanging and .
See [Davey and Priestley 2001] at 39.
Union and Intersection as applied to Elements of dual
117
WILLIAM S. VEATCH
Power Sets of Atoms and Coatoms, provided that we also substitute
Objects for Attributes, and vice versa.
Boolean Algebras. See [Davey and Priestley 2001] at 111.
Ideals and Filters. See Chapter 7, and see [Davey and Priestley 2001]
at 263 et seq.
What is particularly remarkable about these pairs of Operations is that if
we have a valid Theorem that is true for all subsets of U (and not a mere
Equation that is true for only particular subsets) using these Operations, we
can substitute for and vice versa, and the resulting Theorem will also
be valid. Also, we can substitute for , and for U, within a
single Power Set and vice versa, and the resulting Theorem will also be valid.
For example, we know that the following Theorem is true by virtue of
the Distributive Law:
A (B C) = (A B) (A C).
Applying the principle of Duality, we can replace Union by Intersection, and
vice versa, and obtain the following Theorem that is also true:
A (B C) = (A B) (A C).
Extending this one step further, we can substitute for , and
for U (and vice versa), and substitute Object for Attribute (and vice
versa), within dual Power Sets of Atoms and Coatoms, and the resulting
Equation will be valid. For example, see Table 6-2, assuming UD = { a, b, c,
d } = { Dog, Cat, Snake, Crab }. (Since it is a Universe of Discourse and not
the entire Universe, e represents all other Ideas and [5] represents not-e
or ~e.) Here, the Duality appears to be true for Equations, not just
Theorems, i.e., there is Duality if moving horizontally in Table 6-2.
If we move down a column in Table 6-2, we must change the Sum of
the Equation from Dog to Mammal. This is not a true Duality, because we
had to change the Sum. There are, however, rules governing this situation;
the following are equivalent:
118
MATH WITHOUT NUMBERS
A B (where means is a Subset of or is equal to),
A B= A, and
A B = B.
See [Stoll 1963] at 20. Applying these equivalent statements to the
Object example in Table 6-2, we get the following equivalent statements:
Dog Mammal,
Dog Mammal= Dog, and
Dog Mammal = Mammal.
Note that we have not attempted to list all examples of Duality in the
Mathematics of Ideas. (For example, in Chapter 7 we will discuss the Dual
relationship of Ideals and Filters.) Remarkably, examples of Duality
abound. In fact, if we discover a principle that applies to Objects (or Power
Sets of Objects), there undoubtedly is a corresponding Dual principle that
applies to Attributes (or Power Sets of Attributes), and vice versa. We can use
this as a means of testing a proposed Theorem by asking: Does the Dual
Theorem hold true as well? If not, then we know we have a problem with
our original Theorem.
6.4. Do We Need Both Objects and Attributes?
Since a Set of Objects fully determines its related Attributes, and vice versa, do
we really need both? We have asked this question before, and will continue
to do so, because Objects and Attributes are equivalent, although inverse to
one another. In a strictly mathematical sense, the answer is no, we do not
need both, since we can calculate one from the other. From a practical
standpoint, however, we use both Objects and Attributes in the World of
Ideas. Objects are the denotation whereas Attributes are the
connotation.
Another advantage to using both Objects and Attributes, is that the
number of Objects is smaller in the bottom half of an Object Power Set
Lattice; and the number of Attributes is smaller in the top half of an Attribute
Power Set Lattice. Therefore, to minimize the number of characters used,
such as in a computer program, we can optimize by using Objects and
Attributes accordingly. See Fig. 6-2. For example, in a Universe of Discourse
describing Mammals, a relatively small Set of Attributes at the top of the
Power Set Lattice describes a very large Set of Objects constituting Mammals;
whereas a single Object at the bottom of the Power Set Lattice representing
119
WILLIAM S. VEATCH
my black lab named Lucy takes the place of a much larger Set of Attributes
that would be required to identify Lucy. In practice, without thinking about
it, we tend to choose the smaller Set, whether it be Attributes or Objects.
6.5. Update to the Definition of Idea
Ideas as Sets of Objects/Atoms: In the Domain of Atoms, an Idea is
represented by a list of Atoms, i.e., a single Atom if the Idea is an Atom or a
List Set of Atoms if the Idea is a Compound Idea.
Ideas as Power Sets of Objects/Atoms: In the Power Set Domain of
Objects/Atoms, an Idea is made up of the elements of the Power Set of
Atoms in the List Set referred to above.
Ideas as Sets or Power Sets of Attributes/Coatoms: By the Principle
of Duality, there are dual statements about Coatoms:
In the Domain of Attributes/Coatoms, an Idea is represented by a list
of Coatoms, i.e., a single Coatom or a Test Set of Coatoms if the Idea is a
Compound Idea, as to Coatoms.
In the Power Set Domain of Attributes/Coatoms, an Idea is made up
of the elements of the Power Set of Coatoms in the Test Set referred to
above.
6.6. References, Historical Notes, and Further Reading
Duality
[Hilbert and Ackermann 1938]: Principles of Mathematical Logic by Hilbert
and Ackermann is a classic text on mathematical logic, and includes a
discussion of Duality at page 15 et seq.
[Pinter 2014]: Charles Pinter, in A Book of Set Theory, discusses Duality
at page 95-96, in connection with partially ordered classes.
[Davey and Priestley 2001]: Introduction to Lattices and Order by Davey
and Priestley provides an in depth discussion of Duality principles in
numerous contexts, including ordered sets, lattices, Boolean Algebras, and
Ideals.
120
MATH WITHOUT NUMBERS
Galois Theory
[Stewart 2004] at 85 et seq. and 133 et seq. : In Galois Theory, Ian Stewart
has written a very accessible and well-written book about a very complex
subject.
121
7. A MORE DETAILED LOOK AT POWER SETS:
THE BUILDING BLOCKS OF IDEAS
In this Chapter, we take a closer look at Power Sets. The Power Set is the
basic building block of our Mathematics of Ideas. The Power Set provides a
list of all possible Compound Ideas that we can form from a given Set of
Atoms.
This Chapter involves more mathematics that the other Chapters in the
Book. The reader, therefore, may choose to skip this Chapter if it is hard to
follow, or perhaps read only Sections 7.1 through 7.5. For those readers up
for a challenge, it is well worth the effort to read the whole Chapter.
Given a finite Set of Atoms X, we can form a Power Set, written P(X)
(i.e., the Set of all Subsets of X, including and X itself). If we define
Addition as Union and Multiplication as Intersection, then, along with
the Complement Operation, we can define a Boolean Algebra. In other
words, with the Power Set of X, P(X), as the Domain, we have a closed
Algebra. Later, we will show that we can convert any Boolean Algebra into
a Boolean Ring where Addition is Symmetric Difference and
Multiplication is Intersection.
We will see that we can also form a Domain consisting of a Power Set of
Power Sets, P(P(X)), and that we can perform Operations, including Addition
and Multiplication, on the Power Sets within this Domain.
In fact, we will explore two equivalent ways of performing Operations in
our Domain of a Power Set, P(X), or a Power Set of Power Sets, P(P(X)):
122
MATH WITHOUT NUMBERS
The Atomic Approach, and
The Direct Product of Elements or DPE Approach.
7.1. Why Study Power Sets?
There are a number of reasons why we study Power Sets, including the
following:
The Universe of Ideas can be viewed as a Power Set of Object
Atoms, along with its dual Power Set of Attribute Coatoms,
Objects and Attributes in the World of Ideas map to Dual Power
Sets of Atoms and Coatoms, respectively, in the World of Abstract
Sets, and
A single Idea can be interpreted as a Power Set.
We examine this last example more closely in this Subchapter. To
illustrate the concept, we look at the Idea of a Vertebrate in the Universe
of Discourse = XAtoms = { Dog, Cat, Snake, Lizard, Eagle, Robin, Frog,
Toad}. We can define Vertebrate in any of the following ways:
Vertebrate Obj = { Dog, Cat, Snake, Lizard, Eagle, Robin, Frog,
Toad},
o = { Mammal, Snake, Bird, Lizard, Frog, Toad },
o = { Dog, Cat, Reptile, Bird, Frog, Toad },
o = { Dog, Cat, Snake, Bird, Lizard, Amphibian }
o = { Mammal, Reptile, Bird, Amphibian }
o = etc.
From this example, we see that there are many possible ways to define
Vertebrate in terms of Object Sets. Is one definition better than another?
We may not always want to list every Atom separately, but if we Consolidate
Atoms, how do we decide which ones to Consolidate? We start to see that
Vertebrate is equal to, or can be expressed as, every possible combination
of Subsets formed from Partitions of the component Atoms of Vertebrate.
We know from our study of Power Sets that the Power Set is by definition
the Set of all possible Subsets. Each Partition that defines Vertebrate,
therefore, consists of some combination of Subsets of the Power Set of the
component Atoms of Vertebrate. The implication is that we can view an
Idea as the complete Power Set! There is no need to select any one definition,
because they all appear in the Power Set.
123
WILLIAM S. VEATCH
Table 7-1 lists some of the Partitions of our eight Atom Power Set.
Where combinations of Atoms are meaningful to us, the chances are
that we have given the combination its own name. For example, Dog +
Cat we refer to as Mammals. Not all possible Partitions are meaningful
to us, such as Dog + Snake + Toad. Still, the Power Set is a useful way to
encapsulate all of the relevant possible combinations.
If we view each Idea in the Universe of Ideas as a Power Set, then in
order to combine Ideas in a Mathematical sense, we need to develop a
Mathematics of Power Sets. In Chapters 4 (Objects) and 5 (Attributes), we
saw that the Set Operations of Union and Intersection worked well for
regular Sets, along with Symmetric Difference, Set Subtraction, and
Complement. For reasons discussed in detail in Appendix B, we often view
Union (or sometimes Symmetric Difference) as Addition and Intersection
as Multiplication when combining Elements of a Power Set. In the rest of
this Chapter 7, we will look in more detail at the properties of Power Sets
and then we will develop a Mathematics of Power Sets that allows us to
Add and Multiply Ideas that are expressed as Power Sets. We will see
that Addition of Power Sets is related to Addition (i.e., Union or Symmetric
Difference) of the underlying Atoms in the Power Set; and Multiplication of
Power Sets is related to Multiplication (i.e., Intersection) of the underlying
Atoms in the Power Set. In fact, our techniques of Power Set Expansion and
Power Set Contraction, will simplify working with Power Sets.
7.2. What is a Power Set?
To every Set X of Atoms, there is a corresponding Power Set denoted P(X),
which is the Set whose Elements consist of all possible subsets of X.
124
MATH WITHOUT NUMBERS
Importantly, the number of Elements in the Power Set is equal to 2n, where
n = the number of Elements in X. Here are some examples of Power Sets:
Note that the Empty Set, , is always listed as an Element of the Power
Set.
Interestingly, we can also take the Power Set of a Set of Sets, but we
must exercise care as illustrated in the following examples in Table 7-3. (The
symbol Un refers to Direct Product followed by Union. See page 130
for more information on this Operation.)
Note that the Power Set Operation distributes across the Elements in
X. Also, note that we count the Empty Set only one time. (We can prove
that the Empty Set is unique, and therefore should be counted only one time.)
This concept of treating the Power Set as an Operation is extremely
important in our Mathematics of Ideas, because it allows us to switch back
and forth between Sets of Atoms X, Y, , and Power Sets of Atoms P(X),
P(Y), , as long as we exercise care and remember to distribute the Power
Set Operation to any Elements that are Sets.
In fact, this is the key to simplifying complex operations, such as the
addition of Power Sets. To add two or more Power Sets, we simply: (i)
convert the Power Sets to Sets of Atoms, (ii) add the Sets of Atoms, and (iii)
convert the sum back to a Power Set. What began as a complicated endeavor
becomes manageable.
125
WILLIAM S. VEATCH
7.3. How Power Sets Grow in Size
Remember that the number of Elements in a Power Set = 2n, where n = the
number of Elements in the related Base Set of Atoms. This means that every
time we add a new Atom to our Base Set of Atoms, the size of the Power Set
doubles. See Fig. 7-1.
Note that a Base Set X with a single Atom, { a }, becomes a Set of two
Elements when we take the Power Set P(X) = ( , a ). In Lattice notation,
we write this two Element Set as two circles connected by a line. In a Power
Set of Object Atoms, we place at the bottom of the Lattice; in a Power
Set of Attribute Coatoms, we place at the top of the Lattice.
Fig. 7-2 shows a trick to make it easier to draw expanding Power Set
Lattices. We follow these steps:
Draw the Lattice that we want to double in size.
Draw a copy of the Lattice to the right, but starting one row or Level
up.
Tilt the Lattices to the Left. (If drawing by hand, you really need to
do this in each of the first two steps!)
Connect the corresponding Elements of the two Lattices.
In this way, you can see how the P1 Lattice (i.e., the Power Set Lattice
with one Atom) shows up two times in the P2 Lattice (i.e., the Power Set
126
MATH WITHOUT NUMBERS
Lattice with two Atoms); the P2 Lattice shows up two times in the P3 Lattice;
and so on.
7.4. Combinatorics of Power Sets the Binomial Theorem
In this Subchapter, we look at the Combinatorics of Power Sets.
Combinatorics is the study of counting problems, including permutations
and combinations. We know that there is a total of 2n Elements in a Power
Set, but is there a formula for determining the number of Elements in each
row of the Power Set? It turns out that there is a formula, and it is none
other than a variation of the Binomial Theorem or Pascals Triangle.
Let the exponent of x represent the number of Objects, and the
exponent of y represent the number of Attributes, in any row or Level
of the Power Set Lattice. Then, the coefficients of the terms in the expanded
equation correspond to the number of Elements in each Level of the Power
Set.
(x + y)2 = x2 + 2xy + y2 : 1-2-1,
(x + y)3 = x3 + 3x2y + 3xy2 + y3 : 1-3-3-1, and
(x + y)4 = x4 + 4x3y +6x2y2 + 4xy3 + y4 : 1-4-6-4-1.
See Fig. 7-3.
127
WILLIAM S. VEATCH
The binomial theorem provides a means of calculating the number of
Elements in any row of a Power Set:
!
where denotes the binomial coefficient:
! !
n = the number of Atoms in the Power Set, and
k = the number of Atoms in the particular row. The symbol ! means
factorial, where, for example:
5! = 5 4 3 2 1 = 120,
4! = 4 3 2 1 = 24,
3! = 3 2 1 = 6, and
2! = 2 1 = 2.
As an example of calculating the number of Elements in a row of a
Power Set, in the Power Set with four Atoms (P4), the number of Elements
the third row with two Atoms in each Element is:
128
MATH WITHOUT NUMBERS
!
= 24/(2x2) = 6. Also, note that (n-k) is equal to the number of
! !
Attributes in the corresponding row of the Attribute Lattice of Coatoms.
7.5. A Brief Look at Notation
Before we look in detail at the Operations comprising Boolean Algebras and
Boolean Rings, Table 7-4 summarizes the notation for various Operations
that we of think of intuitively as Addition and Multiplication.
Note that the Direct Product is not a Closed Binary Operation in the
same sense as Union, Intersection, and Symmetric Difference. By Closed
Binary Operation, we mean that Sum or Product obtained from applying the
Operation to any two Elements in the Domain is another Element in the
Domain. Union, Intersection, and Symmetric Difference are all Closed
Binary Operations, because the result is always another Element in the same
Domain. In the case of the Direct Product, however, the result of taking the
Direct Product of two Sets is a Set of Ordered Pairs; Ordered Pairs are not
in the original Domains that we work with in the Mathematics of Ideas.
Rather, in the Mathematics of Ideas we use the Direct Product to create a list
of all possible combinations of Elements in two or more Sets, and then we
apply a Binary Operation, such as Union, Intersection, or Symmetric
Difference, to the Elements of each Ordered Pair (or N-Tuple, if we took
the Direct Product of more than two Sets). Therefore, we need a simple
notation that allows us to show that we are taking the Direct Product and
next applying a Binary Operation to each Ordered Pair (or N-Tuple). Table
7-5 illustrates this concept.
129
WILLIAM S. VEATCH
Note in Table 7-5 that we use a colon, :, before the parentheses to
indicate Intersection rather than Union or Symmetric Difference; and we
use a double colon :: to indicate Symmetric Difference. Although we have
covered Operations in the Power Set Domain, for the moment we have not
specified combined Direct Product-Operation notation for the Power Set
of Power Sets Domain. We can do this later if it becomes necessary. Usually
it is not an issue, because we convert to Sets of Atoms, apply Operations,
and then convert back to Power Sets.
There is one exception to Table 7-4. For clarity and simplicity of
notation in our use of Partition Equations (see Chapter 10), we use +
instead of , along with In to indicate Direct Product followed by
Intersection ( ). We do this because the + symbol is easier to read.
Through careful use of the above notation, we can track whether we are
working with Boolean Algebras or Boolean Rings; and whether we are
working within the Domain P(X) (i.e., the Power Set of a Base Set X), or
within the Domain P(P(X)) (i.e., the Power Set of Power Sets of a Base Set
X).
Also, just to emphasize the point, note that in MWN we use the Direct
Product as a tool for listing all of the possible combinations of Elements in
two or more Sets. The result of the Direct Product of two Sets is an Ordered
Pair; three Sets, an Ordered Triple; and n Sets, an Ordered N-Tuple. The
Direct Product, as we use it in this book, is just the first step. The second
step is to apply the relevant Operation to the Elements of the Ordered N-
Tuple, which Operation could be Union, Intersection, Symmetric Difference,
etc. Therefore, any time you apply a Direct Product Operation, you should
ask what Operation follows, i.e., Union, Intersection, Symmetric Difference,
etc.
130
MATH WITHOUT NUMBERS
7.6. Boolean Algebras
Before we look at how to add and multiply Power Sets, it is important
to understand two key concepts from the field of Abstract Algebra: Boolean
Algebras (discussed in this Section 7.6) and Boolean Rings (discussed in
Section 7.7). The Boolean Algebra of all Subsets of X is defined as
P(X); , , 0, 1, _ where:
P(X) is the Domain consisting of a Power Set with 2n Elements: A,
B, C, , where each Element is either , an Atom from the Base
Set X = { a,b,c,n }, or a Compound derived from these Atoms,
is defined as Set Union,
is defined as Set Intersection,
0 is the Empty Set = ,
1 is the finite Base Set of Atoms X = { a,b,c n }, and
_ is the Complement Unary Operation.
(Union and Intersection are Binary Operations, because they act on two
Elements of the Power Set, whereas Complement is a Unary Operation
acting on a single Element.)
What makes the Boolean Algebra so powerful, is that the three
Operations (Union, Intersection, and Complement) are Closed with
respect to the Elements of P(X), meaning that these Operations are always
defined within the Domain/Range equal to the Subsets of P(X). The
following identities in Table 7-6 are always true for any Subsets A, B, and C
in the Domain.
131
WILLIAM S. VEATCH
In addition, the following are always true:
X =
= X, and
Double Complement Law: ( A ) = A.
There are other Boolean Algebras besides the Power Set Boolean
Algebra. For example, when we look at Logic in Volume 2, we will see that
the Logic Operations of , , and also form a Boolean Algebra. Our
primary interest, however, is in the Power Set Boolean Algebra; as well as the
Boolean Algebra formed by a Power Set of Power Sets, sometimes referred
to as Dual Power Sets of Ideals and Filters. First, we look briefly at Boolean
Rings, then we return to Ideals and Filters.
7.7. Boolean Rings
In this Section 7.7, we look at Boolean Rings. A Boolean Ring is similar to
a Boolean Algebra, except that instead of Union, we use Symmetric
Difference for Addition. Remember, that Symmetric Difference is the
same as Union minus Intersection. The significance of this is that every
Element in the Power Set now becomes its own Additive Inverse. This is
one of the requirements for an algebraic structure to be a Ring. For example:
a a = (a a) (a a)=
ab ab = (ab ab) (ab ab)=
abc abc = (abc abc) (abc abc)=
abcd abcd = (abcd abcd) (abcd abcd)=
A Boolean Ring with Identity is defined as P(X), , , e, 0 , where
for A, B, C, each a Subset of P(X):
P(X) is a Power Set, with Elements A, B, C, (either Atoms or
Compounds),
A B=(A B ) ( A B)=(A B)-(A B)
(Symmetric Difference)
A B = A B (Intersection),
e = 1, where 1 is the finite Base Set of Atoms X = { a,b,c n },
and
132
MATH WITHOUT NUMBERS
0 is the Empty Set = .
See [Abbott 1969] at 178 and 185. Interestingly, Multiplication is
defined as Intersection for both Boolean Algebras and Boolean Rings, but,
as mentioned above, Addition is defined differently. In the case of
Boolean Algebras, Addition is defined as Union, but for Boolean Rings,
Addition is defined as Symmetric Difference, also known in Logic as the
Exclusive Or.
The following identities in Table 7-7 are always true for any Subsets A,
B, and C.
Note the differences between a Boolean Ring and a Boolean Algebra. In
a Boolean Ring:
Symmetric Difference is not an Idempotent Operation, whereas
Union is, and
Symmetric Difference is not distributive over Intersection, whereas
Union is distributive over Intersection.
Of course, if two Sets are Disjoint, then Union and Symmetric
Difference yield the same result.
Boolean Rings are interesting also because the Set of Integers, i.e., the
positive and negative counting numbers that we learn in grade school, form
a Ring. Another interesting fact is that any Boolean Ring can be converted
to a Boolean Algebra; and any Boolean Algebra can be converted to a
Boolean Ring. See [Givant and Halmos 2009] at 14.
7.8. General Methodology for Performing Operations on Power
Sets
As mentioned at the beginning of this Chapter, there are two methodologies
133
WILLIAM S. VEATCH
for performing Operations such as Addition or Multiplication on Power Sets:
The Atomic Approach or [Sum/Product/Complement
etc.] of Atoms Approach, and
The DPE Approach or [Sum/Product/Complement
etc.] of DPE Approach.
Each method yields the same result.
Under the Atomic Approach, we take the following steps:
Convert the Power Sets into Base Sets of Atoms.
Apply the relevant Operation (Sum/Product/Complement etc.) to
the Base Sets.
Convert the resulting Base Set back into a Power Set.
Under the DPE Approach, we take the following steps:
Take the Direct Product of the Elements of the n Power Sets that
we want to Add, Multiply, etc., resulting in a Set of N-Tuples.
Apply the relevant Operation (Sum/Product/Complement etc.) to
the Elements of each N-Tuple.
Simplify by eliminating any redundant Elements.
Generally speaking, the Atomic Approach where we convert to Atoms
before applying the Operation is simpler, because there are fewer calculations
to perform. Calculating the Direct Product in the DPE Approach can
become quite cumbersome given that Power Sets increase in size by a power
of 2 as new Atoms are added.
7.9. Adding Power Sets
Now, we develop a means of adding and multiplying Ideas in the form
of Power Sets. We saw earlier that we could use the Union and Intersection
Operations on a Domain equal to a Power Set to form a Boolean Algebra;
and the Symmetric Difference and Intersection Operations on a Domain
equal to a Power Set to form a Boolean Ring. Now, we will explore how to
add Elements consisting of Power Sets, i.e., our Domain is a Power Set of
Power Sets. Instead of calling the Elements a Power Set of Atoms, we
refer to the Elements as Principal Ideals. In the case of Coatoms, we
refer to Power Sets of Coatoms as Principal Filters. (We use these names
because their use is well-established in other areas of mathematics.)
134
MATH WITHOUT NUMBERS
Our goal in this Section 7.9 is to define the Addition operation in a
new Boolean Algebra of Principal Ideals: P(P(X)); , , 0, 1, _ ,
where:
P(P(X)) is a Power Set, with Elements A, B, C, that are Principal
Ideals,
is defined (using the Atomic Approach) as the Power Set of
the Union of Atoms,
is defined (using the Atomic Approach) as the Power Set of
the Intersection of Atoms,
0 is the Empty Set = ,
1 is the Power Set = P( X ), where X is the Base Set of Atoms =
{ a, b, c, n }, and
_ is the Complement Unary Operation.
If you look at a diagram of two Power Set Lattices, it may not be obvious
how to add them together in a meaningful way. Fortunately, an area of
mathematics called Ideal Theory works nicely, and provides a means of
adding Power Sets in a way that resembles Union; and a means of multiplying
Power Sets in a way that resembles Intersection. See [Northcott 1968] at 5.
As we saw in Section 7.8, there are two ways of performing the Addition
Operation that yield equivalent results: the Atomic or Sum of Atoms
Approach, and the Sum of DPE Approach.
It turns out that the Sum of Atoms Approach is the easiest way to add
two Power Sets: (i) first, convert the Power Sets to Base Sets of Atoms, (ii)
then, take the Union of the two Base Sets of Atoms, and (iii) finally, convert
the new Union of Atoms into a Power Set. For example, if we have P(X) =
( , a, b, ab ) and P(Y) = ( , c, d, cd ), then we take the following steps to
add P(X) and P(Y):
Convert P(X) to a Set of Atoms: X = { a,b },
Convert P(Y) to a Set of Atoms: Y = { c,d },
Take the Union of X and Y: Z = X Y = { a,b,c,d }, and
Take the Power Set of Z: P(Z) = ( , a, b, c, d, ab, ac, ad, bc, bd, cd,
abc, bcd, cda, dab, abcd ).
Therefore, P(X) P(Y) = P(X Y).
135
WILLIAM S. VEATCH
(If we wanted to work with a Boolean Ring instead of a Boolean Algebra,
then in step (ii) we would take the Symmetric Difference of the two Sets,
rather than the Union. (Symmetric Difference = Union Intersection.) If
the two Sets are Disjoint, however, then Symmetric Difference and Union
yield the same result. For our current purposes, however, we focus primarily
on Boolean Algebras, keeping in mind that we can always convert to a
Boolean Ring if we want to.)
The second way to add two Power Sets, is the Sum of the Direct
Product of Elements Approach or Sum of DPE Approach. We take
the sum of each pair of Elements where each such pair has one Element from
the first Set and one Element from the second Set. See [Northcott 1968] at
5. Interestingly, the Sum of P(X) and P(Y) defined in this way is also equal
to the Union of each pair of Elements in the Direct Product of P(X) and
P(Y), if we remove any duplicative Elements. For example, suppose we have
two Power Sets P(A) = { , a, b, ab } and P(B) = { , c, d, cd }. The Direct
Product of P(A) P(B) =
{ (, ), (, c), (, d), (, cd), (a, ), (a, c), (a, d), (a, cd), (b,), (b, c),
(b, d), (b, cd), (ab,), (ab, c), (ab, d), (ab, cd) }.
Note that the Direct Product does not combine the Elements the way
that we do when we add or multiply Elements. Instead, the Direct Product
has created a Set of Ordered Pairs; the first Element in each Ordered Pair is
from P(A) and the second Element is from P(B). To complete our Addition
of Power Sets, we need to take the Union of Elements in each Ordered Pair
from the Direct Product. This leaves us with the following: P(A)+P(B) =
{ , c, d, cd, a, ac, ad, acd, b, bc, bd, bcd, ab, abc, abd, abcd }. If you look
closely, you will see that we got the same answer for the Sum P(A)+P(B)
whether using the Sum of Atoms approach or the Sum of the Direct Product
of Elements approach.
The Direct Product is a useful tool for listing all the possible
combinations of Elements in two or more Power Sets. However, the Direct
Product is not a Closed Binary Operation in the sense that it does not
combine the Elements to create a third Element from within the original
Domain. Rather, the Direct Product results in a list of N-Tuples, to which
we can apply the relevant Operation, whether it is Union, Intersection, etc.
Also, note that a single Idea, represented as a Power Set Object, takes
the form (, a). When we add a Power Set Object to another Power Set
Object, we start by taking the Direct Product of the two Power Sets. (In
mathematical terms, when we take the Direct Product of two or more Power
136
MATH WITHOUT NUMBERS
Sets, the Empty Set plays the role of the Identity Element. This means
that when we take the Direct Product of a Power Set Object and a new Power
Set Object, the Empty Set results in a copy of the original Power Set and the
new Atom results in a second copy of the original Power Set multiplied by
the new Atom. In this way, adding a single Atom to a Power Set will always
double the size of the Power Set. See Fig. 7-1.
As we continue to work with Power Sets and apply other Operations,
such as Ring Addition (i.e., Symmetric Difference) or Multiplication (i.e.,
Intersection), we will see that the same two approaches apply in each case.
We can either (i) apply the Operation to Atoms of the Power Sets being
added or multiplied, and then take the Power Set of the Sum or Product; or
(ii) we can create a Set of Ordered Pairs (or N-Tuples if combining more than
two Power Sets) by taking the Direct Product of the Power Sets being added
or multiplied, and then applying the Operation to the Elements of each
Ordered Pair (or N-Tuple).
Fig. 7-4 illustrates the Sum of two Partially Overlapping (POL) Power
Sets using these two approaches. As you can see, the Sum of Atoms
approach is much easier to apply.
137
WILLIAM S. VEATCH
138
MATH WITHOUT NUMBERS
The Sum of the Direct Product of Elements Approach may be
confusing at first, because we are defining a Sum in terms of a Direct Product.
It makes sense, however, if you remember that adding Power Sets is
accomplished by adding the Atoms in the two Power Sets, and then
calculating the Power Set of the Sum, which is the same as adding each pair
of Elements in the Direct Product. It is helpful to know that we get the same
result either way. See the references to [Givant and Halmos 2009] at the end
of this Chapter for further discussion of the sum and product
terminology as used in various branches of mathematics. The key point to
remember is that as used in MWN, Direct Product is not a Closed Binary
Operation in and of itself; rather it is a tool used to create a list of all possible
combinations of the Elements of two or more Sets, in the form of ordered
N-Tuples. We then apply an Operation (e.g., Union, Intersection,
Complement, etc.) to the Elements of each N-Tuple.
Note that we can apply the same methodology in reverse to split a Power
Set into its component parts. We refer to this as a Sum Decomposition
of a Power Set. For example:
P(Z) = P( X Y ) = P(X) P(Y), where Z = { X Y }.
P( a, b, c )
= (, a, b, c, ab, ac, bc, abc ) PS
= P( a, b ) P( c )
= ( , a, b, ab )PS ( , c )PS.
The following Table 7-8 illustrates some common Operations in a
Boolean Algebra based upon a Domain of a Power Set Atoms, P(X), or a
Domain of a Power Set of Power Sets, P(P(X)):
139
WILLIAM S. VEATCH
7.10. Multiplying Power Sets
As with Addition of Power Sets, there are two equivalent ways of performing
the Multiplication Operation: (i) the Atomic or Product of Atoms
Approach and (ii) the Product of Direct Product of Elements or
Product of DPE Approach.
Using the Product of Atoms Approach, (i) we first convert the Power
Sets to Base Sets of Atoms, (ii) then, we take the Intersection of the two Base
Sets of Atoms, and (iii) finally, convert the new Intersection of Atoms into a
Power Set. For example, if we have P(X) = ( , a, b, c, ab, ab, bc, abc ) and
P(Y) = ( , b, c, d, bc, bd, cd, bcd ), then we take the following steps to
multiply P(X) and P(Y):
Convert P(X) to a Set of Atoms: X = { a,b,c },
Convert P(Y) to a Set of Atoms: Y = { b,c,d },
Take the Intersection of X and Y: Z = X Y = { b,c }, and
Take the Power Set of Z: P(Z) = ( , b, c, bc ).
A second way to multiply two Power Sets, is the Product of DPE
approach, where we take the product (i.e., Intersection) of each pair of
Elements where each such pair has one Element from the first Set and one
Element from the second Set. See [Northcott 1968] at 5. The Product of
140
MATH WITHOUT NUMBERS
P(X) and P(Y) defined in this way also equals the Intersection of each pair of
Elements in the Direct Product of P(X) and P(Y), if we remove any
duplicative Elements. For example, suppose we have two Power Sets P(A)
= { , a, b, c, ab, ac, bc, abc } and P(B) = { , b, c, d, bc, bd, cd, bcd }. The
Direct Product of P(A) P(B) =
{ (,), (,b), (,c), (,d), (,bc), (,bd), (,cd), (,bcd),
(a,), (a,b), (a,c), (a,d), (a,bc), (a,bd), (a,cd), (a,bcd),
(b,), (b,b), (b,c), (b,d), (b,bc), (b,bd), (b,cd), (b,bcd),
(c,), (c,b), (c,c), (c,d), (c,bc), (c,bd), (c,cd), (c,bcd),
(ab,), (ab,b), (ab,c), (ab,d), (ab,bc), (ab,bd), (ab,cd), (ab,bcd),
(ac,), (ac,b), (ac,c), (ac,d), (ac,bc), (ac,bd), (ac,cd), (ac,bcd),
(bc,), (bc,b), (bc,c), (bc,d), (bc,bc), (bc,bd), (bc,cd), (bc,bcd),
(abc,), (abc,b), (abc,c), (abc,d), (abc,bc), (abc,bd), (abc,cd), (abc,bcd)}.
Note that the Direct Product does not combine the Elements the same
way as when we add or multiply Elements, but rather the Direct Product has
created a Set of Ordered Pairs where the first Element in each Ordered Pair
is from P(A) and the second Element is from P(B). To complete our
Multiplication of Power Sets, we take the Intersection of Elements in each
Ordered Pair from the Direct Product. This leaves us with the following:
P(A) P(B) =
{ (), (), (), (), (), (), (), (),
(), (), (), (), (), (), (), (),
(), (b), (), (), (b), (b), (), (b),
(), (), (c), (), (c), (), (c), (c),
(), (b), (), (), (b), (b), (), (b),
(), (), (c), (), (c), (), (c), (c),
(), (b), (c), (), (bc), (b), (c), (bc),
(), (b), (c), (), (bc), (b), (c), (bc)}
= (, b, c, bc)PS.
As you can see, we got the same answer for the Product P(A) P(B)
whether using the Product of Atoms Approach or the Product of DPE
Approach.
As we have seen, there are two different calculations used frequently with
Sets and Power Sets that we think of as multiplication. They are very
different, however, and it is important not to confuse the two. One
calculation is the Direct Product. As discussed above (see discussion of the
Direct Product of Elements approach), however, the Direct Product of
141
WILLIAM S. VEATCH
Power Sets creates Ordered Pairs (or N-Tuples if there are more than two
Sets), and is actually just the first step in what we define as addition or
multiplication of Power Sets. The second step is to apply the Operation
at hand (e.g., Union or Intersection) to each Ordered Set in the Direct
Product.
By defining Addition as the Union of Atoms (or Symmetric Difference
for a Boolean Ring) and Multiplication as the Intersection of Atoms, we
see that Power Sets of Ideals together with the Operations of Addition and
Multiplication will form a Boolean Algebra (or Boolean Ring if we use
Symmetric Difference). See Appendix B for a further discussion of Boolean
Algebras and Boolean Rings.
The following Table 7-9 illustrates some common Operations in a
Boolean Algebra based upon a Domain of a Power Set Atoms, P(X), or a
Domain of a Power Set of Power Sets, P(P(X)):
7.11. Power Set Complements
Now, we explore how to determine the Complement of a Power Set. This
142
MATH WITHOUT NUMBERS
can be very confusing at first, but we can avoid most of the confusion if we
carefully specify the Domain up-front, which is typically either the Power Set
of Atoms P(X), or the Power Set of Power Sets P(P(X)). As we will soon see,
in the Power Set Domain we typically look at the Complement of a single
Element, or the Complement of each individual Element in a Set of
Elements. On the other hand, in the Power Set of Power Sets Domain, we
take our familiar Atomic Approach where we convert to Atoms, determine
the Complement, and then convert back to a Power Set.
7.11.1. Power Set Complement: Domain = Power Set P(X)
In the case of a Base Set X = { a,b,c,d }, we can form a Power Set P(X) = P4
= ( , a, b, c, d, ab, ac, ad, bc, bd, cd, abc, bcd, cda, dab, abcd ). To determine
the Complement of an Element of the Power Set of X relative to X, we
simply determine the result of X { Set Element}. For example:
X a = bcd,
X ab = cd, or
X acd = b.
In the above example, the Domain is the Power Set P4, and the
Complement is determined relative to the X=abcd Element at the top of
the Power Set Lattice. (In Lattice Theory, the top Element in a Lattice is
sometimes referred to as the 1 Element, and the bottom Element as the
0 Element.)
We can also, however, determine the Complement of a Power Set (i.e., a
Subset of our Power Set Domain), sometimes referred to as an Ideal, in a
similar manner, by determining the Complement of each Element of the
Ideal. For example, if we have the Power Set A = { , a, b, ab }, then the
Complement of A in P4 is M = { abcd, bcd, acd, cd }. See Fig. 7-5.
143
WILLIAM S. VEATCH
We commonly refer to the Power Set A as an Ideal or Down Set.
The term Down Set is descriptive of the fact that we can form an Ideal in
a Power Set (sometimes referred to as a Principal Ideal) by starting with the
top Element, ab in this case, and forming a Set that includes all of the
Elements in a downward path to 0=.
Note that the Complement of A, referred to here as the Set M, is a
Filter or Up Set, formed by starting with the bottom Element, cd in this
case, and forming a Set that includes all of the Elements in a upward path to
X= abcd.
It is a well-established fact that Ideals and Filters in a Boolean Algebra
come in pairs that are Complements of one another. See [Burris and
Sankappanavar 1981] at 128.
Finally, note that if we relabel our Filter using Coatoms rather than
Atoms, we can more easily see that we can also view the dual Ideal and Filter,
as a dual Power Set of Atoms and related Power Set of Coatoms. See Fig. 7-
6.
144
MATH WITHOUT NUMBERS
The key points to remember are that if we are working in the Power Set
Domain P(X), then the Complement will be another Element in the Power
Set Domain P(X); and if we are determining the Complement of a group of
Elements, such as an Ideal or a Filter, then we determine the Complement
of the group by determining the Complement of each Element in the group,
one-by-one.
7.11.2. Power Set Complement: Domain = Power Set of
Power Sets P( P(X))
Now, we look at how to determine the Complement of a Power Set,
sometimes referred to as an Ideal, when we are in the Domain of the Power
Set of Power Sets. The Complement of an Ideal in P(P(X)) is very different
from the Complement of an Ideal in P(X). The methodology for determining
the Ideal is the same, but the result is different because we are working in a
different Domain.
We have seen that we can Add (i.e., take the Union (for Boolean
Algebras) or Symmetric Difference (for Boolean Rings)) and Multiply (i.e.,
take the Intersection) of Power Sets by Adding or Multiplying the related Sets
of Atoms, and then creating the Power Set of the Sum or Product, as the case
may be. We can take a similar approach to determine the Complement of a
145
WILLIAM S. VEATCH
Power Set within a broader Universe of Discourse. For any given Power Set
within a broader Universe of Discourse, we start by converting the Power
Set to Atoms. We then take the Complements of the Set of Atoms, relative
to the Base Set of Atoms for the Universe of Discourse. Finally, we convert
the Complement Set of Atoms to a Power Set. We call this the
Complement of Base Set Atoms Approach. For example, what is the
Complement of the Power Set P( a,b ) in the UD = P( a,b,c,d )?:
Step 1: P( a,b ) = ( , a, b, ab ). The Atoms are A = { a, b }.
Step 2: The Complement of Atoms relative to the Base Set X =
{ a,b,c,d } for the UD = P( a, b, c, d ) is A = { c, d }.
Step 3: Converting A = { c, d } to a Power Set, we get P( A )=
P( c, d) = (, c, d, cd ).
Therefore, using the Complement of Base Set Atoms Approach, the
Complement of P( a,b ) = ( , a, b, ab ) in UD = P( a, b, c, d ) is P( c, d ) =
(, c, d, cd ). We can check our answer using the generic test for a
Complement in a Lattice:
a a = 1
a a = 0.
Using either Union (for Boolean Algebras) or Symmetric Difference (for
Boolean Rings) for , and Intersection for , we can confirm that the
results are valid:
P( a,b ) P( c,d ) = P( a,b,c,d ) = 1
P( a,b ) P( c,d ) = = 0.
Here is another example of how we determine the Complement of a
Power Set in the context of the Subdivision of Atoms. We start with a Power
Set X = ( , A, B, AB )PS, and then see what happens to the Complements if
we Subdivide A so that A = (, a,b,ab )PS, and subdivide B so that B = ( ,
c,d,cd )PS. (Remember the key point that when a Set appears inside a Power
Set, it is itself a Power Set.) In our initial Power Set X, we can form an Ideal
(or Down Set) equal to (, A); and the corresponding Filter (or Up Set)
would be ( B, AB ). See Fig. 7-7.
146
MATH WITHOUT NUMBERS
In this example, A = B, and B = A. If we subdivide A, so that A =
(, a, b, ab}, and B, so that B = {, c, d, cd}, then we arrive at the same P4
Power Set that we looked at earlier. Intuitively, however it seems like the
Complement of the Power Set A = (, a, b, ab) should be B = (, c, d, cd),
not the Filter as determined in Section 7.11.1, which would be (cd, acd, bcd,
abcd)! How can this be? The answer is that we are working in a Domain
consisting of a Power Set of Power Sets, not Atoms in a simple Power Set
Domain. The first clue that this was the case was when we subdivided A into
(, a, b, ab), not simply {a,b}, and B into (, c, d, cd), not simply {c,d}.
A more helpful and less confusing way to draw the lattice, would be as
a lattice of Power Sets, where the Domain is a Power Set of Power Sets, not
a Power Set of Atoms. See Fig. 7-8.
147
WILLIAM S. VEATCH
Now, we still have a four Atom Power Set, but each Element in the
Power Set is itself a Power Set. (Note that the Empty Set is unique, so it
appears only one time in the Power Set Lattice.) Now, using this Domain of
UD = P(P(X)), we can easily determine that the Complement of the single
Element (,a,b,ab)PS is in fact (,c,d,cd)PS.
Again, we can check our answer using the generic test for a Complement
in a Lattice:
a a = 1,
a a = 0.
Using the Power Set Addition (i.e., based upon Union for Boolean
Algebras and Symmetric Difference for Boolean Rings) and Multiplication
(i.e., based upon Intersection) that we worked out earlier in this Chapter 7,
we see that the test for a valid Complement is satisfied.
In summary, at first it appears that there are two different approaches or
methodologies to determining the Complement of a Power Set within a larger
Universe of Discourse. In fact, however, the methodologies are the same;
what is different is the Domain. The Complement of an Ideal in the Domain
148
MATH WITHOUT NUMBERS
P(X) is different from the Complement of the same Ideal in the Domain
P(P(X)). In particular, when we subdivide an Idea (as we did when we had A
= (, a, b, ab )PS; and B = (, c, d, cd )PS, we should use a Power Set of
Power Sets Lattice as in Fig. 7-8, in order to preserve the Complement
Operation.
7.12. Power Sets Viewed as a Sum of Chains; or a Sum of
Antichains/Partitions
As illustrated in Fig. 7-9, we can view a Power Set as either a sum of Chains,
or as a sum of Partitions. This is important, because Chains and Partitions
are critically important in our study of Ideas. Chains and Partitions, along
with Partition Equations (which are a way of combining Chains and
Partitions), are our primary tools for organizing and sorting Ideas.
A Chain runs vertically in a Power Set Lattice, and is made up of Sets
that are ordered by inclusion. Chains have the following characteristics:
In the traditional way that we draw a Power Set Lattice, a Chain
will appear vertically.
The Elements of the Chain form an ordered Set under inclusion,
e.g., Chain X = ( a, ab, abc, abcd ), where a ab abc abcd.
There are many possible Chains in a Power Set, and many
possible Sets of Chains the sum of which constitute a Partition
of the entire Power Set.
149
WILLIAM S. VEATCH
Partitions have the following characteristics:
In the traditional way that we draw a Power Set Lattice, a
Partition will appear horizontally, as a Set of Disjoint Elements.
The Elements of each Partition form an Exhaustive, Mutually
Exclusive (i.e., Pairwise Disjoint) Partition of the applicable Set
of Atoms into Equivalence Classes.
A Partition also constitutes an Equivalence Relation, such that
the following Relations are true for Elements within the same
150
MATH WITHOUT NUMBERS
Equivalence Class:
o Symmetric: aRa
o Reflexive: aRb bRa
o Transitive: aRb and bRc aRc
There are many possible Partitions in a Power Set.
All of the possible Partitions taken together form a Lattice as
illustrated in Fig. 7-9. See, [Parrochia and Neuville 2013] at 60.
Before we leave Power Sets, we make one quick observation that will
prove to be extremely useful later:
Fact: We can represent any Power Set both as a sum of Disjoint (or overlapping)
Chains, or as a sum of Disjoint (or overlapping) Equivalence Classes.
7.13. Techniques to Simplify Working with Power Sets
Ideally, we would like to develop some techniques that make it easier to work
with Power Sets of Ideas. The issue, as we have seen, is that Power Sets grow
very quickly given that the number of Elements in the Power Set is 2n, where
n the number of Atoms. As we can see from the following sequence,
Power Sets quickly become unmanageable: 2, 4, 8, 16, 32, 64, 128, 256, 512,
1024, 2048, Also, as a practical matter many of the mappings of
Compounds from the Power Set in the World of Abstract Sets to Ideas in
the World of Ideas, are not useful to us. For example, the following
Compound Ideas are not particularly useful: Dog-Rock, Rock-Plane, Sweet-
Mud, etc. Fortunately, there are a number of tools that we can use to limit
the number of Elements that we are working with at any given time, without
sacrificing mathematical rigor. In the following sections, we will develop the
following tools to simplify working with Power Sets:
Power Set Expansion and Contraction of a Set of Atoms,
Partition Equations, and
Nested Partition Equations.
7.14. Power Set Expansion and Contraction of a Set of Atoms
First, we develop a set of rules that allow us to move back and forth between
Sets of Atoms on the one hand and Power Sets of Atoms on the other. In
this way, we can perform operations on the smaller sets of Atoms, and then
151
WILLIAM S. VEATCH
convert to Power Sets when we need to consider the full Set of Compound
Elements.
In the following Table 7.10, + represents Union when working with
a Boolean Algebra (or Symmetric Difference = Union minus Intersection,
when working with a Boolean Ring), and Un represents Direct Product
followed by Union:
Note that the Power Set Operation distributes to any Sets. To avoid
confusion, we use {} for regular Sets and () or ()PS for Power Sets.
One issue that arises frequently is the following: what happens if an
152
MATH WITHOUT NUMBERS
Element that we thought was an Atom turns out in fact to be a Set? In the
following example, we illustrate this by an Element b which started out as
an Atom, but later became a Set B={b,c}. The good news is that the
mathematics of Power Sets can easily handle this sort of change. The key is
to note that the Power Set Operation distributes to any Sets:
P(X) = ( , a, P(B), a Un P(B) ) PS
= ( , a, (b, c, bc), a(b, c, bc) ) PS
= ( , a, b, c, bc, ab, ac, abc )PS
= P ( a, b, c ).
7.15. Partition Equations
Another simplifying technique that we can use is to create a Partition
Equation. A Partition Equation has the following form:
X = Label 1: (Partition 1) In Label 2: (Partition 2) In Label 3:
(Partition 3), where each term in parentheses is an Exhaustive,
Disjoint Partition of the Elements (usually Atoms or Deemed
Atoms) in the Set X. The symbol In in this context represents
Direct Product followed by Intersection. The Label designation
is a label for the type of Partition, e.g., in the case of Planets we use
Size, Distance from the Sun, and Has Moons.
An obvious, but important, point is that the Intersection of any number
of Partitions will always be equal to the original Set X, i.e., since each Partition
is Exhaustive, it contains all Elements of U, and U U U = U. As a
result, we can add any number of Partitions to the Partition Equation.
Example: X = { a, b, c, d, e, f, g } = (abc + de + fg) In (abc + def
+ g) x (ab + cd + efg), where + represents Union and In
represents Direct Product followed by Intersection.
In Chapter 10, we will see how Partition Equations like this become a
very useful and practical tool for sorting Atoms based upon certain related
Attributes. What is most extraordinary is that we can easily convert these
Partition Equations of Atoms to Power Set Partitions as follows for the
above example:
153
WILLIAM S. VEATCH
P(X) = P(abc + de + fg) In P(abc + def + g) In P(ab + cd +
efg) =
P(X) = P(abc + de + fg) P(abc + def + g) P(ab + cd + efg).
7.16. Nested Partition Equations
The next related tool that we develop is a Nested Partition Equation. In a
Nested Partition Equation, some or all of the terms of a Partition may contain
a further Partition separated by a colon : as follows:
X = { Trout, Salmon, Shark, Water Lily, Cattail, Frog, Toad,
Bryophyte, Spike-Weed, Corn, Dog, Garter Snake, Eagle }
= ( a,b,c,d,e,f,g,h,i,j,k,l,m )
= abcde + fghi + jklm
= abcde:(abc+de) + fghi:(fg+hi) + jklm:(j+klm:(k+l+m))
The key to creating Nested Partition Equations is to recognize that
Union is Distributive over Intersection, and vice versa. Therefore, given a
Partition Equation, we can multiply out the Elements as follows:
X=
Dwelling:(Water + Both + Land ) In Living Thing Type:( Animal +
Plant) In Vertebrate Type:( Fish + Amphibian + Mammal + Reptile +
Bird + ~Vertebrate ) =
Dwelling:(abcde + fghi + jklm) In Living Thing Type:(abcfgklm +
dehij ) In Vertebrate Type:( abc+ fg +k+l+m+dehij ).
Step 1: Multiply the first two Partition Sets by distributing one term over the
other:
(abcde + fghi + jklm) In (abcfgklm + dehij ) =
abcde:( abcfgklm + dehij) + fghi:( abcfgklm + dehij) + jklm:( abcfgklm +
dehij).
Step 2: Recognizing that Intersection will eliminate and simplify the
Elements in the inner parentheses, we get the following simplified equation:
abcde:(abc+de) + fghi:(fg + hi) + jklm:(j+klm).
154
MATH WITHOUT NUMBERS
Step 3: Distribute the third Partition Set:
abcde:(abc+de) + fghi:(fg + hi) + jklm:(j+klm) In (abc+ fg +k+l+m+dehij)
=
abcde:(abc: (abc+ fg +k+l+m+dehij)) + de(de:( abc+ fg +k+l+m+dehij)) +
fghi:(fg: (abc+ fg +k+l+m+dehij) + hi: (abc+ fg +k+l+m+dehij)) +
jklm:(j: (abc+ fg +k+l+m+dehij)+klm: (abc+ fg +k+l+m+dehij)).
Step 4: Next simplify by taking the Intersection within the innermost
parentheses:
abcde:(abc: (abc)) + de(de:(de)) +
fghi:(fg: (fg) + hi:(hi)) +
jklm:(j: (abc+ fg +k+l+m+dehij)+klm: (abc+ fg +k+l+m+dehij))
Step 5: Eliminate duplicates or redundancies:
abcde:(abc + de) + fghi:(fg+hi) + jklm:(j+klm:(k+l+m)).
The Lattice diagram in Fig. 7-10 shows how this Nested Partition
Equation looks when we draw the related Nested Power Sets.
155
WILLIAM S. VEATCH
There are a number of interesting points that this diagram illustrates:
There are eight Deemed Atoms, i.e., Elements that cover, or
are just above, the Empty Set (): abc, de, fg, hi, j, k, l, and m.
If we multiply out the number of Elements in the Diagram, we
see that we have 28 = 256 Elements, although many of the
Elements are rolled up or hidden from view.
We can also see a P3 structure with three Deemed Atoms: abcde
+ fghi + jklm.
156
MATH WITHOUT NUMBERS
By using this technique of a Nested Partition and the related
Nested Power Set, we have simplified a complex Set of variables
to a reasonably easy to understand diagram.
Fig. 7-11 illustrates the related Hierarchy Trees.
7.17. The Substitution Principle
The Substitution Principle is one of the most useful techniques
available to us. It is what allows us to study a specialized topic in detail and
then later substitute our work and analysis into the overall Knowledge
Representation structure that we have developed. For example, in Fig. 7-10
we have 13 Atoms and 8 Deemed Atoms, which would give rise to 213 =
8,192 or 28=256 Elements respectively in the related Power Sets. Yet by
consolidating the three abc Atoms into a single Deemed Atom, we reduced
the size of the Power Set from 213 = 8,192 to 211 = 2,048.
If, however, we decide to separately analyze abc as three separate
Atoms, we could do so, and then substitute the result into our previous
Power Set as illustrated in Fig. 7-12.
157
WILLIAM S. VEATCH
While there are still some limitations in terms of the complexity of
diagrams that we are able to draw, we are now representing complex
structures with over 1,000 Elements! Note that the Substitution Principle
allows us to move in both directions, i.e., Coarser (fewer Atoms) to Finer
(more Atoms) diagrams, or from Finer to Coarser.
Fig. 7-13 illustrates the related Hierarchy Trees.
158
MATH WITHOUT NUMBERS
How is this Mathematics of Power Set Lattices helpful to us? One
example, is that we can program a computer to take a Power Set Equation
and generate all possible Hierarchy Trees for us to review, and we can select
the ones that are most helpful in learning, teaching, or communicating Ideas
relating to a given subject.
7.18. Selecting the Order of the Nested Partition Sets
Suppose we have a Partition Equation in the standard format: X = Label 1:
(Partition 1) In Label 2: (Partition 2) In Label 3: (Partition 3) In Label
n: (Partition n). Given n Partition Sets in a Partition Equation, there are n!
(i.e., n factorial) possible combinations or different orders in which we can
apply the Partition Sets to form a Nested Partition Equation. For example,
using our Planet example with three Partition Sets, there are 3! = 3x2x1 =
6 possible Hierarchy Trees:
Size In Distance from Sun In Has Moons,
Size In Has Moons In Distance from Sun,
159
WILLIAM S. VEATCH
Distance from Sun In Size In Has Moons,
Distance from Sun In Has Moons In Size,
Has Moons In Distance from Sun In Size, and
Has Moons Size In Size In Distance from Sun.
For purposes of the Planets example, we will use the following Sets
and Partition Equation:
Planet Atoms = { Mercury, Venus, Earth, Mars, Jupiter, Saturn,
Uranus, Neptune, Pluto } (For present purposes, we treat Pluto as a
Planet, although it is often considered a dwarf planet.).
UD = Planet Deemed Atoms = { a, b, c, d, e }, where:
o a = { Mercury, Venus },
o b = { Earth, Mars },
o c = { Jupiter, Saturn },
o d = { Uranus, Neptune },
o e = { Pluto }.
Planets PE = Size:( Small + Medium + Large ) In Distance from
Sun:( Near + Far ) In Has Moon:( Moon + ~Moon )
= Size:( S + M + L ) In Distance:( N + F ) In Moon:( Mo +
~Mo ).
Given the ability to select the order in which we apply Partition Sets, we
update our methodology as follows:
Step 1 Select Relevant Partition Sets: From the Partition Equation,
select the Partition Sets that will be used to create a Nested Lattice Diagram
or Hierarchy Tree. Selecting n Partition Sets will result in a Hierarchy Tree
with n+1 Levels. Therefore, for most routine purposes, it is convenient to
select up to four Partition Sets to generate a Hierarchy Tree with up to five
Levels. In theory, however, there is no limit to the number of possible
Partition Sets or Levels.
Step 2 - Select the Order in which to apply the Partition Sets: From the
chosen Partition Sets, select the order in which they will be applied.
Mathematically speaking, there is no preference of one order over another,
although some may be more useful than others from a practical standpoint.
Choosing the order in which to apply the Partition Sets is often more of an
art than a science, but there are some guidelines:
160
MATH WITHOUT NUMBERS
Try to proceed from more general at the beginning to more
specific at the end.
Stated another way, look for is a relationship Chains moving
from general to specific.
Some Partition Sets will not, in the World of Ideas, be part of
the is a relationship Chain. In those cases, there is a judgment
call to be made as to the order in which to apply the Partition
Set.
Step 3 Multiply/Distribute the first two Partition Sets:
Multiply/distribute the first two Partition Sets, but do not multiply them all
the way out:
Size:( S + M + L ) In Distance:( N + F )
= ( S:( N + F ) + M:( N + F ) + L:( N + F ) )
(Remember, the colon : indicates that we are taking the Direct Product
followed by Intersection.)
Step 4 Multiply/Distribute by the next Partition Set:
( S:( N + F ) + M:( N + F ) + L:( N + F ) ) In Moon:( Mo + ~Mo )
= S:( N:( Mo + ~Mo) + F:( Mo + ~Mo) ) +
M:( N:( Mo + ~Mo) + F:( Mo + ~Mo) ) +
L:( N:( Mo + ~Mo) + F:( Mo + ~Mo) )
Step 5 Repeat one-by-one with each remaining Partition Set: N/A in
the example.
Step 6 Substitute in Atoms (or Deemed Atoms, if applicable): In the
Planet example, we have the following relationships:
S = {a, b, e}
M={d}
L={c}
N = { a, b }
F = { c, d, e }
Mo = { b, c, d, e }
~Mo = { a }
161
WILLIAM S. VEATCH
S:( N:( Mo + ~Mo) + F:( Mo + ~Mo) ) + M:( N:( Mo + ~Mo) + F:( Mo
+ ~Mo) ) + L:( N:( Mo + ~Mo) + F:( Mo + ~Mo) )
= abe:( ab:( bcde + a) + cde:( bcde + a) ) + d:( ab:( bcde + a) + cde:( bcde
+ a) ) + c:( ab:( bcde + a) + cde:( bcde + a) )
Step 7 Simplify: Recognizing that Intersection will eliminate and simplify
the Elements in the inner brackets, we get the following simplified equation
by simplifying the equation starting with the innermost parentheses and
working our way out:
abe:( ab:( bcde + a) + cde:( bcde + a) ) + d:( ab:( bcde + a) + cde:( bcde +
a) ) + c:( ab:( bcde + a) + cde:( bcde + a) )
= abe:( ab:( b + a) + cde:( cde) )
+ d:( ab:( b + a) + cde:( cde) )
+ c:( ab:( b + a) + cde:( cde) )
= abe:( ab:( b + a) + e:(e) ) + d:(d) + c:(c)
Step 8: Eliminate duplicates or redundancies:
abe:( ab:( b + a) + e:(e) ) + d:(d) + c:(c)
= abe:( ab:( b + a) + e ) + d + c.
Now, can use our sorted Nested Partition Equation to create a Nested
Partition Lattice Diagram. See Fig. 7-14.
162
MATH WITHOUT NUMBERS
We can also use Partition Equations to create Hierarchy Trees. See Fig.
7-15.
163
WILLIAM S. VEATCH
At first this may seem like a complicated process, but with practice it
becomes easier. With each step clearly specified, we can program a computer
to automate the process of generating Nested Partition Equations, along with
Hierarchy Trees and Lattices.
7.19. Multi-Dimensional Power Sets
We end this Chapter with an update to our earlier discussion of Domains, by
updating our list of Domains to include a Power Set of Power Sets Domain
= P(P(X)).
164
MATH WITHOUT NUMBERS
We take a moment now to look more closely at the definition of
P(P(X)), and in particular at the fact that the number of Elements in
P(P(X)) is the same as the number of Elements in P(X). How can this be? If
P(X), where X = { a, b, c }, has 2n = 8 Elements, then should P(P(X)) have
2n = 28 = 256 Elements? The answer is that we need to be very careful to
keep track of the Domain in which we are operating. In the Domain of the
Power Set of Power Sets as used in the Mathematics of Ideas in this Book,
P(P(X)) will have 2n Elements, where n is the number of Atoms in the Base
Set of Atoms. (As discussed below, some authors define P(P(X)) in a
different way, so this is something to watch out for when reading books on
Set Theory.)
165
WILLIAM S. VEATCH
In the Domain of Atoms, we define P(X) as a Unary Operation on a
single Set of Atoms labeled X. P(X) = the Set of all Subsets of X, including
and X, so if X = {a, b, c} then P({a, b, c}) = (, a, b, c, ab, ac, bc, abc)PS.
In MWN, we interpret the result obtained by applying the Power Set
Operation as being in a different Domain, as indicated by the PS subscript.
In the Domain of a Power Set of Atoms, we define P(P(X)) as a Unary
Operation on each Set of Atoms within the Power Set labeled P(X). P(X)
= the Set of all Subsets of X, including and X, so if X = { a, b, c } then
P(X) = (, a, b, c, ab, ac, bc, abc )PS = ( , {a}, {b}, {c}, {a,b}, {a,c}, {b,c},
{a,b,c} )PS. Therefore:
P(P(X)) =
P(, {a}, {b}, {c}, {a,b}, {a,c}, {b,c}, {a,b,c}) =
(P(), P(a), P(b), P(c), P(a,b), P(a,c), P(b,c), P(a,b,c))PS =
( , (, a), (, b), (, c), (, a, b, ab), (, a, c, ac), (, b, c, bc), (, a,
b, c, ab, ac, bc, abc) )DPS.
Note that the definition of the Power Set Unary Operation is the same
in each Domain. What is different is the Domain itself: one is a single set
and one is a Power Set of Sets, where the Power Set Operation distributes to
each Set. As a result, the number of Elements in P(P(X)) is still 2n, where
n is the number of Atoms in the Base Set X.
Some authors define the Power Set Operation in a different way, so the
reader must exercise care when comparing our MWN Power Set Operation
with the discussion in other books on Set Theory. For example, in [Suppes
1972] at 47-48, Suppes defines P(0) = {0} and P(P(0)) = {0, {0}}. In MWN,
if 0 is meant to be , then the Power Set of is always . Cf.,
[Stewart and Tall 1977] at 59, where the authors state that P(0,1) = { ,
{0},{1}, {0,1}}. Here, 0 is treated as a number that is distinct from ,
so we would get the same result in MWN.
7.20. Update to The Definition of Idea
Power Set of Power Sets of Atoms: There is a third Domain called the
Power Set of Power Sets Domain, in addition to the Domain of Atoms
and the Power Set of Atoms Domain. In the Power Set of Power Sets
Domain, an Idea is made up of the elements of the Power Set of Power Sets
formed from the Base Set of Atoms, i.e., each Element of the Domain is a
Power Set.
We can use the Power Set Operation to transform a Set of Idea Atoms to
166
MATH WITHOUT NUMBERS
a Power Set, or a Power Set of Power Sets.
Therefore, we can represent an Idea by a Set, a Power Set, or a Power Set
of Power Sets of Atoms (or Coatoms).
Addition and Multiplication of Ideas: We can Add and Multiply Ideas
in any of the three Domains, where Addition is based on Union (or
Symmetric Difference in a Boolean Ring) and Multiplication is based on
Intersection. We can use either the Atomic Approach or the DPE Approach
to add or multiply Power Sets, but we get the same result either way.
7.21. References, Historical Notes, and Further Reading
[Givant and Halmos 2009]: Introduction to Boolean Algebras, by Halmos
and Givant, is a comprehensive review of Boolean Algebra and Boolean
Rings. The authors also provide a helpful summary of the terminology
relating to sums and products used in various branches of mathematics
that touch on Boolean Algebras. See [Givant and Halmos 2009] at 221-222,
223, 225, 427, and 432. (The confusion comes about because to calculate a
sum of Power Sets, e.g., when viewed as Ideals or Filters in a Boolean
Algebra, we start by taking the Direct Product of the Elements in the two
Power Sets and then take the Union (or Symmetric Difference for a Boolean
Ring) of each pair of Elements. A product of Power Sets is calculated in a
similar way, but using Intersection instead of Union. In summary, for
purposes of MWN we view the Direct Product as a useful technique that we
use on a regular basis (e.g., when adding or multiplying Power Sets), but the
Direct Product is not viewed as the multiplication operation for purposes of
defining Algebras or Rings. Rather, the multiplication operation for purposes
of defining Algebras or Rings is Intersection.)
[Davey and Priestley 2001]: In Introduction to Lattices and Order, Davey and
Priestley provide a very helpful overview of Formal Concept Analysis,
including a detailed discussion of the planet Formal Context at page 65 et
seq.
167
WILLIAM S. VEATCH
[Caspard, Leclerc, and Monjardet 2012]: In Finite Ordered Sets Concepts,
Results, and Uses, the authors provide a detailed analysis of Chains and
Antichains, as well as a discussion of Boolean Lattices at 53 et seq.
168
8. CHAINS: IDEAS THAT FORM ORDERED SETS
OF ATTRIBUTES
In the last Chapter, we saw the definition of a Chain in mathematical terms
relating to abstract Sets. Now we explore Chains of Ideas as they arise in the
World of Ideas. We will see that we can use Chains of Ideas to organize Ideas
in a number of different contexts, including:
Classification Chain: Classification of Objects in the Physical
World, which includes tangible objects, but also intangible objects
such as Historical Facts relating to events that occurred in the
Physical World,
Value System Chain: Classification of abstract Values that we use
to guide our actions in daily life, and
Action Chain: We can use an Action Chain to present the Elements
of an argument in a logical order, and to compare competing
arguments in a debate. Typically, we use an Action Chain to evaluate
proposed actions that we may take in the future, as in a political
debate.
We want to emphasize that we can use a Chain for absolutely any type
of Idea, not just Classification schemes. By discussing Value System Chains
and Action Chains, we see that the Mathematics of Ideas relates to how we
think, not just the Classification of Objects in the Physical World.
8.1. Creating a Valid Chain; Examples of Chains
We start by giving some examples of Chains of Ideas. It may not be obvious
at first that they are Chains, but we will demonstrate that they are in fact
Chains. Then we will look at how to identify Chains when you are studying
169
WILLIAM S. VEATCH
a new subject area.
Here are some examples of different types of Chains:
( Living Thing, Animal, Vertebrate, Mammal, Dog ),
( Land Territory, Country, Country in Europe, Switzerland ),
( Human Rights, Individual Rights, Right to Free Speech ),
( Actions that the Federal Government could take, Actions that
are within Federal jurisdiction, Actions that are cost effective,
Actions that are consistent with our Values, Actions the Federal
Government should take).
Definition: Chain: A nonempty Subset S of U is a Chain in U if S is totally
ordered by . A finite Chain with n elements can be written in the form:
c1 c2 cn. Example: X = (a, ab, abc, abcd, abcde, abcdef,
abcdefg), where a ab abc abcd abcde abcdef abcdefg.
Cf. [Roman 2008] at 5.
8.2. Each Attribute of An Idea Can Be Interpreted as a Level in a
Chain
Interestingly, we can show that each Attribute of an Idea can be viewed as
creating a Level in a Chain. We may not want every Attribute to become a
new Level, but we will discuss that in the next Section. For now, we just
want to demonstrate that any Attribute can become a Level.
The concept is quite simple. We know that the typical dictionary
definition of an Idea is comprised of a list of Attributes. We also know that
each Attribute creates a Partition of the Universe into two Equivalence
Classes: for any Object in the Universe of Ideas, either the Attribute is
True, or the Attribute is False. On this basis, we can create a Chain,
ordering the Attributes from general (at the top) to specific (at the bottom).
See Fig. 8-1. In this case, the dictionary definition of Switzerland includes
the following Attributes, from general to specific: territory of land, that
constitutes a Country, and that is found in Europe.
Key Point: Dictionary definitions are typically lists of Attributes.
170
MATH WITHOUT NUMBERS
Starting at the top of the Chain and working our way down, if the
Attribute is True, we move on to the next Level down; if the Attribute is
False, we move to the right and stop there.
For ease of display, we sometimes limit ourselves to using up to five
vertical Levels (as well as up to five horizontal Partition Classes), along with
a listing of up to 10 Elements for Level (or Partition Class), giving us a total
of 50 Attributes and 50 Objects. In theory, however, there is no limit on the
number of Chain Levels (or Partition Classes) that are possible. For in depth
study of a topic, we could create a Chain with dozens of Levels. It becomes
more of an art than a science, to select Chains (and Partitions) that are useful
to us in our studies.
171
WILLIAM S. VEATCH
Key Point: We can use Attributes to form a Chain of Ideas, where each Attribute is
a Level in the Chain.
8.3. Level of Detail in a Chain
We can create Chains with varying Levels of Detail, so long as the Chain is
an Exhaustive Partition of the Coatoms or Attributes in our Universe of
Discourse. We refer to a Chain with fewer Equivalence Classes (or Levels)
as a Coarse Chain; and a Chain with a greater number of Equivalence
Classes (or Levels) as a Fine Chain. See Fig. 8-2.
172
MATH WITHOUT NUMBERS
8.4. Creating a Chain of Objects
Often when we study a particular Idea, we find that it has numerous
Attributes. We also find that there are many other Ideas that share various
subsets of those Attributes. There is a natural tendency to sort and classify
Objects by the Attributes that they have in common. A Chain is a useful tool
for sorting Ideas that share overlapping Sets of Attributes.
Suppose we have an Idea that we are examining, we have found its
dictionary definition, and we have a list of its Attributes. Each Attribute
represents a quality that every Idea in the Universe either has or does not
have. If an Idea has all of the qualities represented by our list of Attributes,
then it meets the test for that Idea. When we said that Each Attribute
represents a quality that every Idea in the Universe either has or does not
have, this is what allows us to create a Chain. For example, a Dog has the
following Attributes:
Living thing, vertebrate, backbone, fur, barks, mammal, female
produces milk and nurses young, carnivorous, walks on four legs.
We can select some or all of these Attributes to use as Levels of a Chain,
and order them from general to specific. We group together Attributes that
are shared by a common Set of Objects to form Equivalence Classes of
Attributes. We then select an Object Label for each Equivalence Class of
Attributes. In this case, we select Living Thing, Vertebrate, Mammal, and
Dog, as the Object Labels to use as Levels of a Chain. Each Object Label
represents an Equivalence Class of Attributes, and together we have an
Exhaustive Partition of the Attributes of the Idea Dog:
Dog Att = ( Living Thing: Thing, Alive ) + ( Vertebrate: Has a
Backbone ) + ( Mammal: Female Produces Milk and Nurses
Young, Has Hair/Fur ) + (Dog: Carnivorous, Walks on Four Legs,
Barks).
Now, we are ready to create the Chain set forth in Fig. 8-3.
173
WILLIAM S. VEATCH
In the Chain, each branch of the Chain to the right with a ~ represents
the Set of Ideas that do not have the Attribute immediately above. The
Elements in the Chain satisfy the is a relationship test. For example, in
Fig. 8-3, Dog is a Mammal, Mammal is a Vertebrate, and Vertebrate is
a Living Thing. Each Element in the Chain inherits all of the Attributes
of all the Elements above it in the Chain.
To better see how the Chain is made of Sets of Objects, and why we
refer to labels as Object Labels, we can add letters to represent each Set as
shown in Fig. 8-4.
174
MATH WITHOUT NUMBERS
One possible Set of Objects that fits this Chain is the Set X = { a,b,c,d,e},
where:
a = dog,
b = cat,
c = snake,
d = snail, and
e = rock.
By looking at the Sets on the left side of the diagram, we see that we do
in fact have an Ordered Set of Ideas that meets our definition of a Chain:
( a, ab, abc, abcd, abcde ).
We have chosen a familiar example to use as a starting point, but we can
use the same methodology to create a Chain for any Idea. As we create more
examples of Chains, we will see that it is both an art and a science. Initially,
there is a lot of trial and error, and we find that we will update and refine our
175
WILLIAM S. VEATCH
Chains and Level of Detail on a regular basis. For general everyday purposes,
a Chain of three to five Levels usually is sufficient.
8.5. Creating a Dual Chain of Attributes; Inheritance of Attributes
We know from our study of Power Sets that for every Chain of Objects, there
is a dual Chain of Attributes. See Fig. 8-5.
The numbering for the Elements on the right side of the Chain Hierarchy
Tree may be confusing at first, but remember that each number corresponds
to a not Element, so if our Universe of Attributes is [54321], then:
{a} = [5432], i.e., since {a} is present, we exclude 1=~a.
{b} = [5431], i.e., since {b} is present, we exclude 2=~b.
{c} = [5421], i.e., since {c} is present, we exclude 3=~c.
Etc.
176
MATH WITHOUT NUMBERS
We refer to the number in square brackets as the Attribute Signature
of an Idea. We can easily see how Attributes are inherited by looking at the
numerical Attribute Signature. For example, Mammal is represented by
[543], so we know that every Idea that has [543] in its Attribute Signature has
inherited all of the Attributes of Mammals.
8.6. Using a Chain to Sort Attributes from the World of Ideas
Now, to illustrate how we use the Chain based on Sets of Atoms and
Coatoms to sort Attributes from the World of Ideas, we place each Attribute
from the list we created earlier at the highest Level in the Chain for which
the Attribute is True for all Objects at that Level. See Fig. 8-6.
177
WILLIAM S. VEATCH
8.7. Impact of Changes in Level of Detail on the Sorting of
Attributes
Note that if we were to select a Chain with a smaller number of Levels, the
total number of Attributes, inherited and new, remains the same, although
the allocation between inherited and new may change. See Fig. 8-7.
178
MATH WITHOUT NUMBERS
8.8. Partition Sets in a Partition Equation Create Levels in a
Chain; Datasets
Suppose we create a Partition Equation for Individual Data as follows:
Individual PE = (Last Name) In (First Name) In (Country) In (State)
In (City), where:
Last Name = ( Jones + Smith + Chan ),
First Name = ( Sam + Robin + Melissa + Peter ),
City = ( Los Angeles + New York + San Francisco ),
State = ( CA + NY ), and
Country = ( USA ).
Recall that each term in the Partition Equation represents an
Exhaustive, Disjoint Partition of all Objects in the Universe of Discourse.
Also, because Intersection is Commutative, we can place the terms of the
Partition Equation in any order. For example, we could sort our UD first by
Last Name, and then by Country; or by Country first, and then by Last Name.
179
WILLIAM S. VEATCH
We want to demonstrate in this Section 8.8 that each term of the
Partition Equation can create a new Level in a Chain. We illustrate this
concept in Fig. 8-8.
We typically use two special types of Tables to organize Data: Design
View (see Table 8-1) and DataSet (see Table 8-2). In the Design View Table,
we list the Fields used in our Database. Each Field represents a Label for a
Partition of the Universe of Discourse. To be a proper Field, the Values for
the Field must constitute an Exhaustive partition of the Universe of
Discourse into Equivalence Classes that are Pairwise Disjoint. In our
Individual data example, this means that each individual must have a single
Last Name, First Name, and Address.
180
MATH WITHOUT NUMBERS
8.9. Comparing Chains in the Physical World, the World of Ideas,
and the World of Abstract Sets
There are some important differences among Chains that relate to the
Physical World, those that relate to the World of Ideas, and those that relate
to the World of Abstract Sets. In a nutshell, Chains that relate to the Physical
World can be verified by observation or scientific experimentation designed
to determine cause and effect. This is not to say that it is easy to make
scientific discoveries. The reality is quite the opposite. In theory, at least, it
is possible to determine the truth of a Proposition (two Ideas), Syllogism
(three Ideas), or Lattice of Ideas (any number of Ideas) concerning the
Physical World.
In the World of Ideas, however, we are limited only by our imagination.
We can live on Mars, or travel through time, or live in a land inhabited by
unicorns and purple cows. The rules governing the construction of
Knowledge Representation structures and the Logic of Ideas, however,
181
WILLIAM S. VEATCH
remain the same as in the Physical World.
In the World of Abstract Sets, we attempt to build our world upon a
foundation with as few assumptions as possible. Chapter 12 on Set Theory
discusses the mathematical foundations in more detail, but the key elements
in the World of Abstract Sets are the following:
Objects exist and are unique.
o Given two Objects, we can distinguish one from the
other. (This allows us to label them a,b,c, , although
they are exactly the same as each other.)
Other than their separate existence, Objects have no
distinguishing characteristics.
An Object can be subdivided ad infinitum.
Two or more objects can be consolidated into a single Object.
o Ultimately, all Objects can be consolidated into the
Universal Object U.
The Empty Set exists, which contains no Objects.
Given any n Objects, we can create a Power Set consisting of
all possible subsets, including the Empty Set, that we can create
from these Objects.
With these basic assumptions, we can take any Set of Objects, for
example, X = { a,b,c,d }, and do a number of things, including:
Create a Power Set,
Create Chains,
Create Partitions, and
Create Partition Equations and Nested Power Sets.
A Chain is a series of Sets that are Ordered by inclusion. We can create
a number of Chains from our set X = { a,b,c,d } using sets from the Power
Set, including:
( a, ab, abc, abcd )
( b, bd, bdc, bdca )
( c, cb, cba, cbad )
Etc.
In the World of Abstract Sets, all of these chains are equally valid. Look
what happens, though, when we attach labels from the Physical World. For
182
MATH WITHOUT NUMBERS
example, let X = { a,b,c,d } = { dog, cat, snake, tree }. Now our first Chain
reads as follows:
( a, ab, abc, abcd ) = ( {dog}, {dog,cat}, {dog, cat, snake}, { dog,
cat, snake, tree }) = ( dog, mammal, vertebrate, living thing ).
In the last term of the equation, we substituted mammal for {dog, cat};
vertebrate for {dog, cat, snake }; and living thing for {dog, cat, snake,
tree }. We were able to do this because the Compound Ideas mammal,
vertebrate, and living thing have special significance to us, to the point
that we find it useful to give these Compound Ideas their own names.
Our second Chain, however, is not as meaningful in the Physical World:
( b, bd, bdc, bdca ) = ( {cat}, { cat,tree}, { cat, tree, snake}, { cat,
tree, snake, dog }).
We have not given these particular combinations of Ideas their own
special name, because they do not have common Attributes that distinguish
them in a meaningful way, although the structure of the Chain is exactly the
same. We mention this to highlight a key difference between the World of
Ideas and the World of Abstract Sets. In the World of Abstract Sets, both
of the above Chains are equally valid, whereas in the World of Ideas, one has
significance and the other does not. The art is in selecting Chains that have
significance when we map them to the Physical World or the World of Ideas.
We will see that the mapping is critical, not just the concept of a Chain.
Once again, mathematics can tell us how to think, but not what to think.
8.10. Three Principal Types of Chains: Classification, Value
System, and Action
As we saw in the previous section, any Ideas that we are capable of thinking
can be organized into Chains. This is true because in mathematical terms,
Objects have no characteristics other than uniqueness; it is the combination
of Ideas into Sets that has significance. If we create Sets that are ordered by
inclusion, then we have a Chain. In the World of Ideas and in the Physical
world, however, some Chains have significance, whereas others do not. In
this Subchapter we consider three types of Chains:
Classification: By Classification Chains, we mean classification of
Ideas that relate to Objects, tangible or intangible, that we can
observe or measure in the Physical World, or which are related by
183
WILLIAM S. VEATCH
cause and effect as verified by scientific experiment. Classification
Chains could be facts or explanations if they are not
controversial. If controversial, then the Chain becomes an
Argument Chain that must be backed up with Evidence.
Value System: Values are different from Classifications, because
they are intangible and subjective by their nature. Values may be
codified into law or a code of ethics, or they may be unwritten
general principles that guide us in our daily lives. For some, religion
is a source of values, but for others the source of values may be
simply a sense of what is right and what is wrong. In any event,
Values are a type of Idea that we can organize and sort with Chains
as we would any other Ideas.
Action: An Action Chain may be as simple as a Syllogism in
Classical Logic, but more often it is a complex Chain of elements
and reasoning in the form of an Argument. In a Debate, we typically
compare and contrast two competing Arguments to determine
which is the better Argument. We use Facts to support each
Element in the Argument, and we use Values to try to persuade the
audience that one Argument is better than another.
We saw earlier in Chapter 1, an example of an Argument Chain. Now,
we look at the same Argument Chain, but with three different Levels of
Detail: Coarse, Standard, and Fine. See Fig. 8-9.
184
MATH WITHOUT NUMBERS
In the Coarse Argument Chain, we have two Test Elements. In the
Standard Argument Chain, there are six Test Elements. In the Fine
Argument Chain, we take each of the six Elements in the Standard Chain and
break it down into three parts:
Is there Relevant Evidence to support the argument that the
Element is satisfied or True?
Is the Evidence Reliable?
185
WILLIAM S. VEATCH
Is the Evidence sufficiently Probative, i.e., does the evidence
carry sufficient weight to meet the applicable standard of proof?
For now, we just want to give an example of an Argument Chain and how
there can be varying Levels of Detail. In Book One, Volumes 2 and 3, we
will explore Argument Chains in more detail, and explain what we mean by
Relevant, Reliable, and Probative Evidence.
8.11. Simplifying Assumption: Using a Five Level Chain
In theory, there is no limit to the number of Levels that we can have in a
Chain. As we have seen, each Attribute of an Object can be viewed as a
Level in a Chain. The maximum number of Attributes that an Object can
have is 2n-m, where n is the number of Atoms in the Universe and m is
the number of Atoms in the Idea itself, which is obviously a staggeringly large
number. As a practical matter, for everyday purposes we can define most
Ideas with a Chain that has five Levels. This is a convenient size of Chain
for display on-line on a computer screen, or in a print dictionary, along with
a list of 10 Attributes per Level in the Chain, for a Total of 50 Attributes
sorted into five Levels.
As we will see shortly, we can take a similar approach to sorting Objects,
and use a five Class Partition together with a List of ten Objects per Class,
for a total of 50 Objects. This gives us a MWN definition of any Idea, with
up to 50 Attributes and 50 Objects sorted into five Levels and five Partition
Classes.
Of course, when studying a subject in more detail than is necessary for a
general purpose Dictionary, there is no limit to the number of Levels,
Partition Classes, Attributes, or Objects.
8.12. Displaying a Chain as a Nested Lattice
Fig. 8-10 uses a Nested Lattice Diagram to illustrate how we can create a
Chain by adding one Atom at a time. In addition to adding a new Atom at
the top of the Chain, we need to multiply everything below by the new Atom.
186
MATH WITHOUT NUMBERS
Fig. 8-11 uses a Nested Lattice Diagram to illustrate creating a Chain
from a P8 Power Set Lattice, by Subdivision.
187
WILLIAM S. VEATCH
188
MATH WITHOUT NUMBERS
8.13. Update to the Definition of Idea
Chains: We can use each Attribute in the Test Set of an Idea to create a
Level in a Chain.
Level of Detail: We can create Chains of varying Levels of Detail from
Coarse to Fine, by selecting fewer or more Attributes to use as Levels in the
Chain. If an Idea has m Attributes, then the applicable Chain can have
from 1 to m Levels.
The Levels of a Chain, expressed as Object Labels, form an
Exhaustive, Pairwise Disjoint Partition of the Attributes of an Idea.
8.14. References, Historical Notes, and Further Reading
[Roman 2008]: In Lattices and Ordered Sets, Steven Roman provides a
detailed discussion of Chains, Lattices, and Boolean Lattices.
[Schrder 2003]: In Ordered Sets, Bernd Schrder provides another
detailed view of Chains, Antichains, and Lattices.
189
9. PARTITION SETS / ANTICHAINS: IDEAS THAT
FORM DISJOINT SETS OF OBJECTS
In this Chapter, we study Partitions in more detail. Together with Chains,
Partitions are one of the primary tools that we use to define Ideas. The term
Partition Set may be new to the reader, but the concept of a partition goes
back to the time of Aristotle, when Aristotle described the Ten Categories
into which all Ideas can be classified. See [Kreeft 2010] at 55. The two
principal rules for creating valid Partitions are the following:
To constitute a valid Partition, a Partition must constitute an
Exhaustive partition of a Set, meaning that all of the Elements of the
partitioned Set must belong to one of the Equivalence Classes in the
Partition; and
The Elements of a Partition Set must be Mutually Exclusive,
meaning that each pair of Equivalence Classes must be Disjoint.
Stated another way, an Element may belong to one and only one
Equivalence Class.
If we follow these two rules, then we can create valid Partitions of any
Idea. See Fig. 9-1, which illustrates how a Partition divides a Set into a
number of Equivalence Classes. The Elements in the same Equivalence
Class are equivalent to one another in respect of the property or Attribute
that we are measuring. We use the term Partition Set to refer to the Set of
Elements that serve as labels for the Equivalence Classes of the Partition.
For example, we can write Vertebrate Part = ( Mammal + Reptile +
Amphibian + Fish + Bird ). Each Element of the Partition Set is an Attribute
Label used to identify Equivalence Classes of Objects that share the specified
Attributes. This Set is similar to a List Set, except that the subscript Part
190
MATH WITHOUT NUMBERS
indicates that the Set is an Exhaustive Partition, and that the Elements are
Pairwise Disjoint. See Fig. 9-1.
Each section in the Partition is an Equivalence Class, and we refer to
the Set of Equivalence Classes as a Partition Set.
Given any Set with a finite number of Elements, the rules of
mathematics, and combinatorics in particular, tell us how many partitions are
possible. See [Parrochia and Neuville 2013] at 59. Our goal here, however,
is to find particular Partitions that mirror the way that Ideas are structured in
the real world. As you read this Chapter 9, note the Inverse Duality of
Partitions as compared to Chains in Chapter 8.
9.1. Creating a Valid Partition Set; Examples of Partitions
Just as listing Attributes is the first step towards creating a Chain, listing
Objects is a first step towards creating a Partition Set. Then, we group
together Objects that share common Attributes to form Equivalence Classes,
and select an Attribute Label for each Equivalence Class of Objects. To have
a valid Partition Set, however, we must check that it is Exhaustive and
Mutually Exclusive (Pairwise Disjoint). For example, if we take the Idea
Animal, it is easy to create an Object Set as follows:
191
WILLIAM S. VEATCH
Animal List = { Vertebrate, Mammal, Dog, Cat, Crab, Eagle, Dolphin,
Other Animal }.
The Set is Exhaustive, due to the use of the Element Other Animal,
but we can quickly see that the Elements of the Set are not Mutually
Exclusive. To be Mutually Exclusive, the Elements must be Pairwise
Disjoint, meaning that the Intersection of any two Elements must be the
Empty Set. Here, to give one of many possible examples, Vertebrate Dog
= Dog. This also highlights one of the difficulties to using an Element like
Other Animal, because there is no way to be certain that the Element is
Pairwise Disjoint; rather, we must assume that it is Disjoint since we cannot
test it directly. That is, by definition Other Animal cannot be Vertebrate,
Mammal, Dog, Cat, or the other Elements in the List.
Note that Partition Sets are not unique. To illustrate this point, we can
show that there are a number of different ways to fix the prior Set to make it
a valid Partition Set. Consider the following:
Animal List = { Vertebrate, Crab, Other Animal } ME ,
Animal List = { Dog, Cat, Crab, Eagle, Dolphin, Other Animal } ME ,
and
Animal List = { Mammal, Crab, Eagle, Other Animal }ME.
(The subscript ME in the above examples indicates that the Elements in
each Set are Mutually Exclusive, i.e., Pairwise Disjoint.)
Using our Partition Set notation, we can re-write the above Sets as
follows:
Animal Part = ( Vertebrate, Crab, Other Animal ) = ( Vertebrate +
Crab + Other Animal ),
Animal Part = ( Dog, Cat, Crab, Eagle, Dolphin, Other Animal ) =
( Dog + Cat + Crab + Eagle + Dolphin + Other Animal ), and
Animal Part = ( Mammal, Crab, Eagle, Other Animal ) = ( Mammal
+ Crab + Eagle + Other Animal ).
These two forms of notation are equivalent, but the Partition Set
notation will be helpful when we look at Partition Equations in Chapter 10.
192
MATH WITHOUT NUMBERS
9.2. Each Object in An Idea List Set Can Be Interpreted as an
Equivalence Class in a Partition or Antichain
In Section 8.2, we saw that each Attribute of an Idea can become a Label
for a Level in a Chain. Similarly, there is an Inverse, Dual concept for Objects
and Partitions: each Object can become a Label for an Equivalence Class in
a Partition (sometimes referred to as an Antichain). We saw this in the
examples given above in Section 9.1. The only limitations are that the
Elements of the Partition Set taken as a whole must be Exhaustive, i.e., they
must account for all Objects in our Universe of Discourse, and the Elements
of the Partition Set must be Mutually Exclusive, i.e., Pairwise Disjoint.
9.3. Level of Detail in a Partition
As with Chains, we can create Partitions with varying Levels of Detail, so
long as the Partition is an Exhaustive partition of the Atoms in our Universe
of Discourse. We refer to a Partition with fewer Partition Classes as a
Coarse Partition; and a Partition with a greater number of Partition
Classes as a Fine Partition. See Fig. 9-2.
193
WILLIAM S. VEATCH
While each Object has the potential to become its own Equivalence
Class in a Partition, it is not necessary. Rather, there is an art to selecting
Partition Sets with an appropriate or useful Level of Detail in the particular
circumstances.
194
MATH WITHOUT NUMBERS
9.4. Creating a Partition (or Antichain) of Objects
In Section 8.5, we looked at how we could create a Chain of Attributes (as
well as a dual Chain of Objects in Section 8.4) by selecting certain Attributes
to use as Levels in a Chain, and assigning an Object Label to each Level.
Similarly, there is an Inverse, Dual concept for an Antichain/Partition of
Objects. We can select certain Objects to use as Equivalence Classes in a
Partition, and assign an Attribute Label to each Equivalence Class in the
Partition. Fig. 9-3 illustrates this dual relationship between Chains and
Partitions. We use a Chain to create a Partition of Attributes; we use an
Antichain to create a Partition of Objects.
195
WILLIAM S. VEATCH
Why do we call the labels of the Levels in a Chain Object Labels and
the labels of the Equivalence Classes in an Antichain Attribute Labels?
Since every Idea has both an Object Set and an Attribute Set associated with
it, either label would work. However, we chose Object Label for Levels in
a Chain to reflect the sequence of events when creating a Chain:
Select an Idea to define by a Chain (e.g., Vertebrate),
Select a Cover Object Superset, or several Supersets and create a
Chain (e.g., (Vertebrate: Animal: Living Thing: Tangible Thing)), and
Use the Chain to sort Attributes. See Fig. 9-3.
196
MATH WITHOUT NUMBERS
Similarly, we chose Attribute Label for Equivalence Classes in an
Antichain to reflect the sequence of events when creating an
Antichain/Partition:
Select an Idea to define by an Antichain/Partition (e.g., Vertebrate),
Select Attribute Sets that create an Exhaustive Partition (e.g.,
Vertebrate = (Fish + Reptile + Bird + Amphibian)), and
Use the Antichain/Partition to sort Objects. See Fig. 9-3.
9.5. Using Partition Sets to Sort Objects from the World of Ideas
Just as we can use Chains to sort Attributes, we can use Antichains/Partition
Sets to sort Objects. (This is another example of the Dualities that abound
in mathematics.) Note that the Objects in the Partition Set are Disjoint and
Exhaustive, whereas the Attrributes are usually Partially Overlapping (POL).
(The Attributes of Atoms are Disjoint in the two-Element Power Set P2, but
are POL in larger Power Sets.) See Fig. 9-4 for an example of the P4 Power
Set.
197
WILLIAM S. VEATCH
Consider the following Universe of Discourse, UD = { Dog, Snake,
Frog, Eagle, Trout }. We can create the following Partition Set:
UD= { Mammal, Reptile, Amphibian, Bird, Fish }.
We can then use that Partition Set to sort the Objects in our Universe of
Discourse. See Fig. 9-5.
198
MATH WITHOUT NUMBERS
199
WILLIAM S. VEATCH
9.6. Impact of Changes in Level of Detail on the Sorting of Objects
If we reduce the Level of Detail in our Partition, the result is to redistribute
the Objects in our UD, so that there are more Objects in a smaller number of
Equivalence Classes. See Fig. 9-5, where the Coarse Partitions are at the top,
and the Finer Partitions are at the bottom.
9.7. Partition Sets in a Partition Equation Create Levels in a
Chain; Datasets
We touched on the concept in Section 8.8 and we will explore it in more
detail in Chapter 10, but each time we add a Partition Set to a Partition
Equation, we create a new Level in a Chain. See Fig. 8-8.
9.8. Displaying a Partition as a Nested Lattice
Fig. 9-6 uses a Nested Lattice Diagram to illustrate how we can create
various Partitions of a P4 Lattice.
9.9. Update to the Definition of Idea
Partitions: We can use each Object in the List Set of an Idea to create an
Equivalence Class in an Antichain / Partition.
200
MATH WITHOUT NUMBERS
Level of Detail: We can create Antichains of varying Levels of Detail
from Coarse to Fine, by selecting fewer or more Objects to use as
Equivalence Classes in the Antichain. If an Idea has m Objects, then the
applicable Antichain can have from 1 to m Equivalence Classes.
The Equivalence Classes of an Antichain, expressed as Attribute
Labels, form an Exhaustive, Pairwise Disjoint Partition of the Objects of an
Idea.
9.10. References, Historical Notes, and Further Reading
[Roman 2008]: In Lattices and Ordered Sets, Steven Roman provides a
detailed discussion of Chains, Antichains, Lattices, and Boolean Lattices.
[Schrder 2003]: In Ordered Sets, Bernd Schrder provides another
detailed view of Chains, Antichains, and Lattices.
201
10. PARTITION EQUATIONS: A PRACTICAL TOOL
FOR ORGANIZING IDEAS
A Partition Equation is a practical tool that we can use to organize Ideas. It
also demonstrates the power and usefulness of our new Mathematics of
Ideas, combining Chains and Partitions to Create Hierarchy Trees and
Outlines of Subjects. We start by creating a list of relevant Partitions of a
given Set of Atoms. We connect the Partitions by In representing Direct
Product followed by the Intersection Operation. By multiplying out the
Partition Equation in part, we can create Hierarchy Trees and Outlines in a
systematic way. Only the relevant Power Set Elements appear in the
Hierarchy Tree or Outline; the other Elements are still present, but are
hidden from view.
10.1. Definition of a Partition Equation
A Partition Equation is a list of Partition Sets with respect to an Idea. A
Partition Equation has the following form:
X = Label 1: (Partition 1) In Label 2: (Partition 2) In Label 3:
(Partition 3), where each term in parentheses is an Exhaustive,
Disjoint Partition of the Elements (usually Atoms or Deemed
Atoms) in the Set X. The symbol In in this context represents
Direct Product followed by Intersection.
The Elements inside each Partition are joined by Set Union ( ), and
the Partitions themselves are linked together by the symbol In which
stands for Direct Product followed by Intersection ( ) in this case, so
we could write a Partition Equation as follows:
202
MATH WITHOUT NUMBERS
U PE = Label 1: (A B) Label 2: (C D) Label 3: (E F)
Label 4: (G H I).
Sometimes for ease of reading, however, we write the Partition Equation
using + instead of ( ), and ( In ) instead of ( ), but the reader
should keep in mind that the operations used are Union and Intersection in
a Boolean Algebra:
U PE = Label 1: (A+B) In Label 2: (C+D) In Label 3: (E+F) In
Label 4: (G+H+I).
Normally, an sign, as we use it in MWN, would stand for Direct
Product. In this case, however, we use In to mean Direct Product
followed by Intersection. We take the Direct Product first and then apply
the Intersection Operation to each Ordered Pair (or N-Tuple) in the Direct
Product. (Recall from Chapter 7, that the Direct Product, as we use it in
MWN, is a tool for creating a list of N-Tuples of all combinations of
Elements in two more Sets, but is not itself a Closed Operation like Addition
or Multiplication. Usually, we use the Direct Product to create a list of
Ordered N-Tuples, and then we apply the relevant Operation to the
Elements of each N-Tuple.)
We sometimes use the subscript PE to show that we have a Partition
Equation. Remember that a Partition Set, by definition, is an Exhaustive
partition of the Idea into Mutually Exclusive subset Ideas.
Here are some examples of Partition Equations:
Vertebrate List = { Human, Dog, Cat, Snake, Frog, Eagle, Trout,
Salmon },
o Vertebrate PE
= Type In Dwelling In Mobility
= Type: (A+B+C+D+E) In
Dwelling: (F+G+H) In
Mobility: (I+J+K+L)
= Type: ( Mammal + Reptile + Amphibian + Bird + Fish )
In
203
WILLIAM S. VEATCH
Dwelling: ( Land + Water + Land/Water ) In
Mobility: ( Walks on Two Legs + Walks on Four Legs +
Slithers + Swims ).
Planets List = { Mercury, Mars, Earth, Venus, Saturn, Jupiter,
Neptune, Uranus, Pluto }
o Planets PE = Size In Distance from Sun In Has Moons
= Size:( Small + Medium + Large ) In
Distance from Sun:( Near + Far ) In
Moons: (True + False )
Note that sometimes we include a label before the Partition Set, such as
Type before the following Partition Set: Type:( Mammal + Reptile +
Amphibian + Bird + Fish ).
For any Partition Equation, we can substitute letters in order to make it
easier to see patterns. For example, we can re-write our Vertebrate example
as follows:
Vertebrate List = { Human, Dog, Cat, Snake, Frog, Eagle, Trout,
Salmon }
Vertebrate PE =
Type:( Mammal + Reptile + Amphibian + Bird + Fish ) In
Dwelling:( Land Dwelling + Water Dwelling + Land/Water
Dwelling ) In
Mobility:( Walks on Two Legs + Walks on Four Legs + Slithers +
Swims )
= Type:(A+B+C+D+E) In Dwelling:(F+G+H) In
Mobility:(I+J+K+L).
10.2. Creating Nested Partition Equations with Two Partition Sets
We start with a Base Set of Atoms X = { a, b, c, }. Then, we identify and
select two Partition Sets, keeping in mind that to be a valid Partition Set, the
Partition must be Exhaustive and the Equivalence Classes in the Partition
must be Mutually Exclusive.
For example, we can use the Set of Planets X = { Mercury, Venus,
204
MATH WITHOUT NUMBERS
Earth, Mars, Jupiter, Saturn, Uranus, Neptune, Pluto }.
Next, we create a Cross Table to highlight those Attributes that we
are most interested in examining. See Table 10-1.
The Attributes listed in the Cross Table in Table 10-1 are as follows:
1 Small (S)
2 Medium (M)
3 Large (L)
4 Near the Sun (N)
5 Far from the Sun (F)
6 Has a Moon (Mo)
7 No Moon (~Mo)
A Cross Table is a tool that we have borrowed from the field of Formal
Concept Analysis (FCA). See Appendix B for more information about FCA.
See [Davey and Priestley 2001] at 65 et seq., and [Carpineto and Romano 2004]
at 17 et seq., for discussion of the Planet example in FCA. (We understand
that Pluto is no longer considered a planet, but we include it anyway in the
example. Technically, Pluto is now considered a dwarf planet.)
Note that although we have nine planets, there are four pairs where there
is no distinction given the Attributes that we have chosen to examine. We
can, therefore, reduce our Set of Deemed Atoms to five Sets, a through
e. We list the relevant Attributes across the top of the Cross Table, and
the Objects down the left side. Then, we place an x in each row where an
Object has the corresponding Attribute in a particular column. The Cross
Table is very helpful in identifying patterns and determining whether or not
Partitions are Exhaustive and Mutually Exclusive.
205
WILLIAM S. VEATCH
The next step is to look for Partition Sets so that we can create an
Attribute Partition Equation. We can identify Partition Sets by looking for
two or more columns of Disjoint Elements that create an Exhaustive
Partition of the Base Set. In this example, we have columns 1, 2, and 3: Size
= ( Small + Medium + Large ); columns 4 and 5: Distance from Sun = ( Near
+ Far ); and columns 6 and 7: Moon = ( Moon + ~Moon ).
Given the three Partition Sets, we can create a number of different
Partition Equations that each have two Partition Sets:
X = ( S + M + L ) In ( N + F ).
X = ( S + M + L ) In ( M + ~M ).
X = ( N + F ) In ( M + ~M ).
X = ( N + F ) In ( S + M + L ).
X = ( M + ~M ) In ( S + M + L ).
X = ( M + ~M ) In ( N + F ).
In fact, if we have n Partition Sets and we select any two of the
Partitions Sets to make a Partition Equation, then there are n (n-1)
possible Partition Equations where the order of the Partition Sets makes a
difference. In this case, if we select two of the three Partition Sets, then there
are 3 2 = 6 possible Partition Equation with two Partition Sets, where order
makes a difference.
To create a Nested Partition Equation, we use the Distributive Law to
multiply each term of first Partition Set by the entire second Partition Set:
X = ( S + M + L ) In ( N + F ) = S:( N + F ) + M:( N + F ) +
L:( N + F ).
We could stop here and create Hierarchy Tree or Lattice, but if you do
so you will see that some of the branches are empty. Therefore, we add one
more step first; we substitute in the Atoms for the Equivalence Class names,
and then simplify the equations:
X = S:( N + F ) + M:( N + F ) + L:( N + F )
= abe:(ab+cde)+c:(ab+cde)+d:(ab+cde)
= abe:(ab+e)+c+d.
206
MATH WITHOUT NUMBERS
To simplify the Partition Equation, we look at the second factor in
each term and eliminate any Atoms that do not appear in the first factor. For
example, the term abe:(ab + cde) becomes abe:(ab + e). We deleted
cd because when we take the Intersection of abe and cde, cd
disappears. (We explained the mathematics behind this in Chapter 7, but
here we just want to apply the end result.)
Now, we can create a Hierarchy Tree. See Fig. 10-1.
Lets look at two other variations of Nested Partition Equations with
two Partition Sets each.
Using Size and Whether the planet has a Moon, see Fig. 10-2:
X = ( S + M + L ) In ( Mo + ~Mo )
= ( S:( Mo + ~Mo ) + M:( Mo + ~Mo ) + L:( Mo + ~Mo ) )
= abe:(bcde+a) + c:(bcde+a) + d:(bcde+a)
= abe:(be+a) + c + d.
207
WILLIAM S. VEATCH
Using Distance from the Sun and Whether the planet has a Moon,
see Fig. 10-3:
X = ( N + F ) In ( Mo + ~Mo )
= N:( Mo + ~Mo ) + F:( Mo + ~Mo )
= ab:(bcde+a) + cde:(bcde+a)
= ab:(b+a) + cde.
208
MATH WITHOUT NUMBERS
10.3. Creating Nested Partition Equations with Three Partition
Sets
Given the three Partition Sets in our planet example, we have the following
Partition Equation with three Partition Sets:
X = ( S + M + L ) In ( N + F ) In ( M + ~M ).
Given n Partition Sets in a Partition Equation with three Partition Sets,
we have n! = n(n-1)(n-2) = 3(2)(1) = 6 possible Nested Partition Equations
where order makes a difference:
X = ( S + M + L ) In ( N + F ) In ( M + ~M )
X = ( S + M + L ) In ( M + ~M ) In ( N + F )
X = ( N + F ) In ( M + ~M ) In ( S + M + L )
X = ( N + F ) In ( S + M + L ) In ( M + ~M )
X = ( M + ~M ) In ( S + M + L ) In ( N + F )
X = ( M + ~M ) In ( N + F ) In ( S + M + L ).
We will not take the time to simplify all six possible Nested Partition
209
WILLIAM S. VEATCH
Equations, but we will give three examples that build upon the examples in
the previous Section. Starting with the two Partition Set equation, we
multiply each term of second Partition Set (in each place that it shows up) by
the entire third Partition Set:
X = (S:( N + F ) + M:( N + F ) + L:( N + F )) In ( M + ~M )
= S:( N:( M + ~M ) + F:( M + ~M ) ) + M:( N:( M + ~M ) +
F:( M + ~M ) ) + L:( N:( M + ~M ) + F:( M + ~M ) )
Then, starting with the simplified two Partition Set Equation, we add
the Atoms for the third Partition Set:
S:( N:( M + ~M ) + F:( M + ~M ) ) + M:( N:( M + ~M ) + F:( M
+ ~M ) ) + L:( N:( M + ~M ) + F:( M + ~M ) )
= abe:(ab:(bcde+a)+e:(bcde+a))+c:(bcde+a)+d:(bcde+a)
= abe:(ab:(b+a)+e)+c+d.
To simplify the equation, we look at the third nested factor in each
term and eliminate any Atoms that do not appear in the second factor. (We
explained the mathematics behind this in Chapter 7, but here we just want
to apply the end result.)
Now, we can create a Hierarchy Tree. See Fig. 10-4.
210
MATH WITHOUT NUMBERS
Lets look at two other variations of Nested Partition Equations with
three Partition Sets each.
Using Size, Whether the planet has a Moon, and then Distance
from the Sun, see Fig. 10-5:
X = ( S:( Mo:( N + F ) + ~Mo:( N + F ) ) + M:( Mo:( N + F ) +
~Mo:( N + F ) ) + L:( Mo:( N + F ) + ~Mo:( N + F ) ) )
= abe:(be:(ab+cde)+a:(ab+cde)) + c:(ab+cde) + d:(ab+cde)
= abe:(be:(b+e)+a ) + c + d.
Using Distance from the Sun, Whether the planet has a Moon, and
then Size, see Fig. 10-6:
X = N:( Mo:( S + M + L ) + ~Mo:( S + M + L ) ) + F:( Mo:( S +
M + L ) + ~Mo:( S + M + L ) )
= ab:(b:(abe+d+c)+a:(abe+d+c)) + cde:(abe+d+c)
= ab:(b+a) + cde:(e+d+c).
211
WILLIAM S. VEATCH
10.4. Adding a Partition Set Creates a New Level in a Chain
When we create a Hierarchy Tree or an Outline from a Partition Equation,
the number of Levels is equal to n+1, where n is the number of Partition
Sets in the Partition Equation. We can add any number of Partition Sets to
a Partition Equation joined by In , i.e., Direct Product followed by
Intersection, because each Partition Set is equal to U, and the Intersection
of U with U is equal to U.
For ease of reading, we often try to choose our UD so that we can use
up to four Partition Sets, creating a Chain with up to five Levels. In this
book, we also try to limit our Partitions to five Element Partition Sets for
ease of display in the Book or on the computer screen. There is in theory,
however, no limit on the number of possible Levels in a Chain or Elements
in a Partition Set. Note, however, that we allow for a listing of 10 Attributes
for each Chain Level, and 10 Objects for each Partition Set, so even with
these practical limitations we can display up to 50 Attributes, and 50 Objects
when defining a single Idea. For most everyday purposes, this is adequate.
When programming computers for Artificial Intelligence, however, there is
no need to impose any such limitations.
As we have discussed earlier, the Direct Product symbol In refers to
Direct Product followed by Intersection. Returning the the Vertebrate
example that we looked at in Subchapter 10.1, since the Intersection
Operation is Commutative, we can multiply out the Partition Sets in any
order that we like in order to sort and classify our Universe of Vertebrates.
212
MATH WITHOUT NUMBERS
See the Cross Table in Table 10-2.
The Attributes listed in the Cross Table in Table 10-2 are as follows:
Type:
1 Mammal
2 Reptile
3 Amphibian
4 Bird
5 Fish
Dwelling:
6 Land
7 Water
8 Both (Land and Water)
Mobility:
9 Walks on Two Legs
10 Walks on Four Legs
11 Slithers
12 Swims
Being Commutative means that the multiplication of the terms
Type, Dwelling, and Mobility can take place in any order. We could, in fact,
multiply out the terms completely, but then we would end up with 5x3x4=60
different combinations of Type, Dwelling, and Mobility. Since we have only
eight Objects in our Universe of Discourse, however, most of the entries
would be empty. A better approach, then, is to sort the Elements by one
category, such as Type, and then sort by another category, such as
Dwelling.
213
WILLIAM S. VEATCH
To give a couple of examples, we will start with Type In Dwelling
In Mobility, and then show what happens if we change the order to
Dwelling In Mobility In Type.
For Type In Dwelling In Mobility, we have the following Partition
Equation:
Vertebrate PE = Type:(abc+d+e+f+gh) In Dwelling:(abcdf+e+gh)
In Mobility:(af+bce+d+gh)
=
(abc:(abcdf:(af+bce+d+gh)+e:(af+bce+d+gh)+gh:(af+bce+d+gh))+
d:(abcdf:(af+bce+d+gh)+e:(af+bce+d+gh)+gh:(af+bce+d+gh))+
e:(abcdf:(af+bce+d+gh)+e:(af+bce+d+gh)+gh:(af+bce+d+gh))+
f:(abcdf:(af+bce+d+gh)+e:(af+bce+d+gh)+gh: (af+bce+d+gh))+
gh:(abcdf:(af+bce+d+gh)+e:(af+bce+d+gh)+gh:(af+bce+d+gh)))
= (abc:(abc:(a+bc)+e:(e)+gh:(gh))+
d:(d:(d)+e:(e)+gh:(gh))+
e:(e:(e))+
f:(f:(f))+
gh:(gh:(gh)),
= (abc:(abc:(a+bc))+
d:(d:(d))+
e:(e:(e))+
f:(f:(f))+
gh:(gh:(gh))),
= (abc:(a+bc)+d+e+f+gh).
See Fig. 10-7.
214
MATH WITHOUT NUMBERS
For Dwelling In Mobility In Type, we have the following Partition
Equation:
Vertebrate PE = Dwelling:(abcdf+e+gh) In
Mobility:(af+bce+d+gh) In Type:(abc+d+e+f+gh)
=
(abcdf:(af:(abc+d+e+f+gh)+bce:(abc+d+e+f+gh)+d:(abc+d+e+f+g
h)+gh:(abc+d+e+f+gh)) +
e:(af:(abc+d+e+f+gh)+bce:(abc+d+e+f+gh)+d:(abc+d+e+f+gh)+
gh:(abc+d+e+f+gh)) +
gh:(af:(abc+d+e+f+gh)+bce:(abc+d+e+f+gh)+d:(abc+d+e+f+gh)+g
h:(abc+d+e+f+gh)))
= (abcdf:(af:(a+f)+bc:(bc)+d:(d)) +
e:(e:(e)) +
gh:(gh:(gh)))
= (abcdf:(af:(a+f)+bc+d) + e + gh).
215
WILLIAM S. VEATCH
With practice, we can create Hierarchy Trees very quickly without going
through all of the detailed steps listed above. Our main purpose, however,
in demonstrating the steps in detail is to show that we can automate the
process, so that, for example, we could program a computer to generate
Hierarchy Trees by simply selecting the order of Partition Sets and letting the
computer do the rest of the work.
10.5. Lattice Representations of a Nested Partition Equation
In Fig. 10-9, we have a Lattice Representation of the Hierarchy Tree that we
created in Fig. 10-8.
216
MATH WITHOUT NUMBERS
217
WILLIAM S. VEATCH
Below is the same Hierarchy Tree expressed as an Outline.
Vertebrate
I. Land
a. Two Legs
i. Mammal
1. Human
ii. Bird
1. Eagle
b. Four Legs
i. Dog
ii. Cat
c. Slithers
i. Land Snake
II. Water
a. Trout
b. Salmon
III. Both Land and Water
a. Frog
Note that each Level in the Outline constitutes a Partition of relevant
Atoms.
10.6. The Sorting Room: a Laboratory for Experiments with
Ideas
As in any scientific endeavor, it is critical that we test our hypotheses and
theories in a laboratory to see if they are valid. In our case, we need a place
to test our theories about Ideas, in order to confirm whether or not our
theories correctly explain how Ideas are structured and behave in the
Universe of Ideas. Therefore, we create the Sorting Room.
The easiest way to think of the Sorting Room is as a laboratory where we
conduct experiments on Ideas. As illustrated in Fig. 10-10, we can use a
Partition to Sort Object-Ideas; and we can use a Chain to sort Attribute-Ideas.
In this way, we can take a set of Idea-Objects or Idea-Attributes, chosen using
a particular methodology or at random, and test whether a given Partition or
Chain is an Exhaustive, Mutually Exclusive partition of the Objects or
Attributes, respectively. For example, if we have Ideas left over, then we
know we do not have an Exhaustive Partition. Similarly, if an Idea fits into
more than one Equivalence Class, then we know that our Partition did not
create Mutually Exclusive Equivalence Classes.
218
MATH WITHOUT NUMBERS
In addition to using a Partition or Chain to sort Objects or Attributes,
respectively, we can also use an Attribute Lattice of Coatoms to sort Objects,
or an Object Lattice of Atoms to sort Attributes, all as illustrated in Fig. 10-
11. Sorting Objects is like pouring marbles into a funnel opening over an
219
WILLIAM S. VEATCH
Attribute Lattice. Each Object-Idea marble will fall into the Equivalence
Class having the Attributes exhibited by that Object-Idea. It is also like
playing the game of 20 Questions, where we start with questions like Is it
an animal [vegetable] [mineral]?, followed by Is it bigger than a breadbox?,
to determine the size. The concept, in mathematical terms, is to use
Partitions that become increasingly specific. We start with a general Partition,
as we expect to see at the top of an Idea Lattice, and move to more and more
specific Partitions, with more and more Attributes, as we move down the
Lattice.
Similarly, sorting Attributes is like releasing helium balloons into a
funnel opening under a Chain or Lattice of Objects. Each Attribute-Idea
balloon will float up to the highest Level in the Chain or Lattice where all
Objects exhibit such Attribute-Idea.
220
MATH WITHOUT NUMBERS
Sorting Ideas in this way, whether expressed as Objects or Attributes, is
how we create a Knowledge Representation map of the Universe of Ideas.
221
WILLIAM S. VEATCH
10.7. Update to the Definition of Idea
Partition Equations: Having defined Closed Binary Operations of
Addition (based upon Union) and Multiplication (based upon Intersection),
we can create Partition Equations, as well as Nested Partition Equations that
combine Partitions and Chains in a way that reflects the relationships among
Ideas in the World of Ideas.
An Idea (Atomic or Compound) is defined by a Universe of Discourse
together with a Partition Equation that identifies how the Idea is related to
each other element of the Universe of Discourse.
Outlines, Hierarchy Trees, and Nested Lattice Diagrams: We can
represent Nested Partition Equations either as Outlines or Hierarchy Trees,
as wells as Nested Lattice Diagrams.
If there are n Partition Sets of a UD, then there are n! (read n
factorial) possible Partition Equations where order of the Partition Sets
makes a difference. Therefore, there are n! possible Hierarchy Trees or
Outlines.
10.8. References, Historical Notes, and Further Reading
The concept of a Partition Equation is a new Idea that we explore in
MWN, but there are many excellent resources regarding the component
parts, including Chains, Partitions, and Latticss.
[Abbott 1969]: This is one of the best all-around resources on Sets,
Lattices, and Boolean Algebras.
[Caspard, Leclerc, and Monjardet 2012]: Along with the next three
books listed below, this is one of the go-to sources of information on Finite
Ordered Sets and Lattices.
[Roman 2008]
[Schrder 2003]
[Grtzer 2003]
222
MATH WITHOUT NUMBERS
11. UNIVERSES OF DISCOURSE
Creating a Universe of Discourse is another tool that we can use to simplify
working with complex sets of Ideas. The concept is simple: we select a Subset
of the Universe of all Ideas, and analyze that Subset using our familiar tools
of Chains, Partitions, Lattices, etc. We have seen, however, that every Idea in
the Universe of Ideas is potentially related to every other Idea in the Universe,
so we need to have a means of expanding a Universe of Discourse, or
combining Universes of Discourse. In fact, if we could combine Universes
of Discourse, this would allow us to study and map out specific areas of study
in detail, and then combine the results to further our lofty goal of creating a
Knowledge Representation map of all Ideas. In this Chapter, we will begin
to study how to combine Universes of Discourse. We will see that there are
rules and guidelines for how to do this, but that it is not a trivial task.
We have seen that Ideas can be classified as either Atoms or
Compounds made up of Atoms. The Atoms and Compound Ideas bear a
one-to-one relationship to the Power Set created by the Base Set of Atoms.
The Power Set quickly becomes unwieldy, however, because it doubles in
size each time we add a new Atom. As a result, we found that we could
develop a number of techniques to simplify working with large sets, including
working with Chains, which we typically draw vertically, and Partitions, which
we typically draw horizontally. In fact, we can interpret a Power Set as a sum
of Chains, or alternatively as a sum of Partitions. Next, we looked at the
concept of a Partition Equation, and saw that we could create nested
Partitions that then created Chains. Now, in this Chapter we look at another
technique for simplifying working with large Sets by looking at a subset of
the entire Universe of Ideas, which we call a Universe of Discourse or UD for
short. By studying a smaller Universe of Discourse, we can focus in on a
particular subject area without being distracted by other unrelated topics.
The real power of Universes of Discourse will become apparent when we see
223
WILLIAM S. VEATCH
that, with care, we can combine Universes of Discourse.
We will see that Universes of Discourse are really just Power Sets, or, in
mathematical terms, Ideals consisting of Atoms, and Filters consisting of
Coatoms. As such, we can add and multiply Universes of Discourse just
as we would add or multiply any other Power Sets. The challenge is to
determine how the Sets of Atoms in the two Universes of Discourse compare
to one another. We know, however, that there are only five possibilities:
Identical, Disjoint, POL, SuperSub, or SubSuper. What makes the analysis
potentially difficult, as we will soon see, is that sometimes an Atom in one
UD is a Set in another UD.
11.1. What is a Universe of Discourse?
In Subchapters 2.6 and 7.11 we looked at the importance of specifying the
Domain when working with the Mathematics of Ideas, and not mixing
elements from different Domains. (We will revisit this topic again in
Subchapter 12.3.) We identified a number of different Domains including
the following:
Object Atoms: X,
Attribute Coatoms: Y,
Power Set of Atoms: P(X),
Power Set of Coatoms: P(Y),
Power Set of Power Sets of Atoms: P(P(X)), and
Power Set of Power Sets of Coatoms: P(P(Y)).
In a Universe of Discourse, we still have all of these same different
possible Domains, but the difference is that we start with a subset of the
entire Universe of Atoms or Coatoms, as the case may be.
11.2. How Do We Combine Two or More Universes of Discourse?
When we try to combine two different Universes of Discourse, the starting
point is to compare the Sets of Atoms in the Universes of Discourses. (In
theory, we could use Coatoms as well, but the analysis is the same.) As with
any two Sets, there are only five possible ways in which the two Sets of Atoms
can be related to one another:
Identical / The Same: The two UDs could have the same Atoms,
in which case the two UDs are in fact the same.
Disjoint: The Intersection of Atoms in the two UDs is the Empty
224
MATH WITHOUT NUMBERS
Set.
Partially Overlapping: The two UDs share some Atoms, but the
Sets of Atoms are not the same, and are not in a Subset-Superset
relationship.
Subset-Superset: The first UD is a Subset of the second.
Superset-Subset: The first UD is a Superset of the second.
What complicates the analysis is that what is considered an Atom in one
UD, may need to be subdivided into two or more Atoms before combining
with the second UD.
11.3. Combining Universes of Discourse with One and Three
Atoms, Respectively
To begin our analysis of how to combine two UDs, we start with a simple
example of adding a one-Atom UD to a three-Atom UD. See Fig. 11-1. On
the top left side, we have a three-Atom UD1 with Atoms X = { a, b, c }. Note
that when comparing UDs, we always start with the full Universe U. As a
result we have R1 which represents the remainder or residual, i.e., all of the
Ideas in the Universe other than a, b, c. On the right side, we have a one-
Atom UD2 with Atoms X = { N }. We have R2 which represents all of the
Ideas in the Universe other than N. We use a capital letter N for our single
Atom UD2, because we do not know yet how this single Atom will compare
to the Atoms in UD1; N could turn out to be a Set rather than an Atom. We
know from our study of Set Theory, however, that there are only five possible
ways in which the Atoms in UD1 and UD2 may be related: Identical, Disjoint,
POL, Subset-Superset, or SuperSet-Subset. We will look at each in turn.
225
WILLIAM S. VEATCH
11.3.1. The Same - Identical Sets of Atoms
At the bottom of Fig. 11-1, we see what the combined UDs look like where
the Atoms are the same, i.e., although stated as a single Element N, upon
further investigation it turned out that N = { a, b, c }, i.e., UD1 is simply a
Subdivision of UD2. We sometimes refer to UD1 as being Finer, with more
Atoms, and UD2 as being Coarser, with fewer Atoms (or Deemed Atoms).
11.3.2. Disjoint Sets of Atoms
In Fig. 11-2, we look at what happens when we combine two Universes of
Discourse that have Disjoint Sets of Atoms. Here, the task of combining the
226
MATH WITHOUT NUMBERS
two Universes of Discourse is relatively easy, as we simply add a new Atom.
Our combined Universe of Discourse is now the four-Atom Power Set (P4)
rather than the three-Atom Power Set (P3). Note, however, that we will still
need to look at the impact on Chains, Partitions, and Partition Equations
when we add a new Atom. We will touch on this in Subchapter 11.4.
227
WILLIAM S. VEATCH
11.3.3. Subset-Superset of Atoms
In Fig. 11-3, we look at a more complicated scenario where the Atom in our
second Universe of Discourse is a Subset of the Set of Atoms in our first
Universe of Discourse. What makes this scenario more complicated is the
possibility that N in UD2 could be a Subset of a single Atom in UD1. To take
this possibility into account in Fig. 11-3, we labeled the third Atom in UD1 as
M, where M = { c, d }. (There are other possible scenarios as well. For
example, UD2 could be a subset of a portion of two or even all three Deemed
Atoms in UD1. In that case, the new combined UD could have up to six
Deemed Atoms. In any event, we focus for now on the example where the
third Atom in UD1 is M, where M = { c, d }, resulting in a new UD with
four Deemed Atoms.) Therefore, N could be equal to any one of the 14
Elements of the P4 Power Set; we exclude and { a,b,c,d }, because we are
assuming that N is a non-empty, proper Subset.
228
MATH WITHOUT NUMBERS
229
WILLIAM S. VEATCH
11.3.4. Superset-Subset of Atoms
If N is a Superset of UD1 = { a, b, c }, then N must have at least one additional
Atom, which we represent by d, which could be a single Atom or a Set of
Atoms. In any event, we can consolidate the additional Atoms into a single
Deemed Atom d. As you can see from Fig. 11-4, UDs that are in a
Superset-Set relationship are relatively easy to combine.
230
MATH WITHOUT NUMBERS
11.3.5. Partially Overlapping (POL) Sets of Atoms
If UD1 = { abc } and UD2 = { N } are POL, i.e., N {abc} 0, then we
know that UD2 must contain at least one additional Atom as compared to
UD1. We represent this additional Atom by d. See Fig. 11-5.
231
WILLIAM S. VEATCH
There is another POL scenario that we need to consider. It is possible that
N in our one Element UD2 is partially overlapping with a Subset of an
Atom in UD1. Remember, the whole concept of an Atom is really a fiction,
232
MATH WITHOUT NUMBERS
because any Atom can be further subdivided in theory. As a result, what may
be an Atom (or Deemed Atom, really) in UD1 may need to be broken down
into its sub-Elements before we can combine it with the Deemed Atoms in
UD2. In the case of our three Atom UD1, if we split one of the Deemed Atoms
into two parts, then we now have a total of five Atoms. See Fig. 11-6. If we
split each of the three Deemed Atoms in UD1 into two parts, then, together
with the new Atom in UD2, we would have a total of seven Atoms.
233
WILLIAM S. VEATCH
234
MATH WITHOUT NUMBERS
11.4. How Do We Combine More Complicated Universes of
Discourse?
Now, we look at a more complicated example. In Section 11.3, we examined
what happens to the underlying Base Set of Atoms when we combine a
Universe of Discourse (UD1) with three Atoms, with a Universe of Discourse
(UD2) with one Atom. We saw that the first step was to compare the Base
Sets of Atoms to see which of the five Set Relationships governed (i.e.,
Identical, Disjoint, Subset-Superset, Superset-Subset, or Partially
Overlapping). We saw that the principal potential complication was that a
Deemed Atom in one Universe of Discourse could be a Set of Atoms in
another Universe of Discourse. Once we have determined the new Base Set
of Atoms in our new combined Universe of Discourse, however, the
question is: what is the impact on any Chains, Partitions, and Partition
Equations that were in our original Universes of Discourse when we combine
them together? To begin to answer this question, we look at our Living
Thing: Animal-Plant example. We will see that it is possible to develop rules
and algorithms for combining Universes of Discourse, including updating
any Chains, Partitions, and Partition Equations.
11.4.1. The Same, Identical Universes of Discourse
If the two UDs have the same Atoms, then there is no impact on the Power
Set. See Fig. 11-7.
235
WILLIAM S. VEATCH
The obvious question is: why it was not apparent from the beginning
that the UDs were the same? It may be, for example, that the UDs were in
different languages, so it was not apparent that they were the same until they
were translated.
For example, if we have UD1 = { red, green, yellow }Atoms and UD2 =
{ rouge, vert, jaune }Atoms, then we can show that the two UDs are the same.
236
MATH WITHOUT NUMBERS
Another possibility is that two experts in a particular field of study may
have been using different terminology to describe events. When examined
more closely, however, it turns out that the analysis is the same.
In terms of the impact on Chains, Partitions, and Partition Equations,
there is no change, given that the Atoms are the Identical.
11.4.2. Disjoint Universes of Discourse
If the two UDs have Disjoint Sets of Atoms, then it is easy to form the
combined Power Set, as all we have to do is form a Power Set from the Union
of Atoms. As illustrated in Fig. 11-8, each of the original UDs is preserved
within the combined Power Set.
Chains and Partitions within the original Power Sets are preserved,
although they now relate to only a portion of the whole Power Set, and would
need to be extended to be complete and exhaustive. To extend Chains and
Partitions to cover the combined Universe of Discourse, we look at the
Partition Equation, which in effect combines Chains and Partitions into a
single equation.
The impact on Partition Equations is more complicated. For each
Partition from one Power Set, we need to ask whether the Atoms from the
other Power Set would fall within an existing Equivalence Class, or whether
we need to create a new Equivalence Class.
For example, suppose that for UD1 we have the following Partition
Equation for P3 with X = { a, b, c } = { Black Dog, White Cat, Brown Crab}:
U = X PE = Name In Type of Animal1 In Type of Animal2 In
Color In Dwelling
=
Name: ( a + b + c ) In
Type of Animal1: ( Vertebrate ab + Invertebrate c ) In
Type of Animal2 :( Mammal ab + Crustacean c ) In
Color: ( Black a + White b + Brown c ) In
Dwelling: ( Land ab + Water c )
Now, lets look at the impact of adding a fourth Disjoint Atom d, for
237
WILLIAM S. VEATCH
example:
Brown Snake, or
Silver Rock.
If we add d = Brown Snake, then for the new combined UD we have
the following Partition Equation for P4 with X = { a, b, c, d } = { Black Dog,
White Cat, Brown Crab, Brown Land Snake }:
U = X PE = Name In Type of Animal1 In Type of Animal2 In
Color In Dwelling
=
Name: ( a + b + c + d ) In
Type of Animal1: ( Vertebrate abd + Invertebrate c ) In
Type of Animal2: ( Mammal ab + Crustacean c + Reptile d ) In
Color: ( Black a + White b + Brown cd ) In
Dwelling: ( Land abd+ Water c ).
For each of the five Partitions, we had to ask whether our new Atom fit
within an existing Equivalence Class, or whether we need to add a new
Equivalence Class. In the cases of Name and Type of Animal2, we had
to add a new Equivalence Class.
If we add d = Silver Rock, then for the new combined UD we have the
following Partition Equation for P4 with X = { a, b, c, d } = { Black Dog,
White Cat, Brown Crab, Silver Rock }:
U = X PE = Name In Type of Object: ( Living Thing: ( Type
of Animal1 In Type of Animal2 In Dwelling) + ~Living ) In
Color =
Name: ( a + b + c + d ) In
Type of Object:
(Living Thing abc:
(
Type of Animal1: ( Vertebrate ab + Invertebrate c ) In
Type of Animal2: ( Mammal ab + Crustacean c ) In
Dwelling: ( Land ab+ Water c )
)
238
MATH WITHOUT NUMBERS
+ ~Living d ) In
Color: ( Black a + White b + Brown c + Silver d ).
Once again, for each of the five Partitions, we had to ask whether our
new Atom fit within an existing Equivalence Class, or whether we need to
add a new Equivalence Class. Since a Rock is not alive and is therefore not
an Animal, we had to add a new Equivalence Class to distinguish between
Living and Not-Living things. See Fig. 11-8 for an illustration of how adding
a new Disjoint Atom changes the Lattice interpretation of our Partition
Equation.
239
WILLIAM S. VEATCH
Note that when we added d = Brown Snake, it shifted the location of
Vertebrate and Animal in the Lattice; whereas, adding d = Silver Rock did
not change the Lattice in this way.
This highlights one of the difficulties of adding UDs together: the
location of defined terms within the Lattice may change. As we have seen,
there may be no easy way to check the validity of Partition Equations when
combining Disjoint Universes of Discourse, other than to check each Atom
one-by-one. On the other hand, we need to map them only one time, and
then the Mapping remains valid for all time.
11.4.3. Subset-Superset Universes of Discourse
In many cases, if the Atoms of the second UD are a Subset of the first, then
combining the two Universes of Discourse is easy, because the Subset is
already subsumed within the Superset. The one exception is if the Atoms of
the second UD are a Subset of one or more Deemed Atoms in the first UD,
in which case we need to split one of more Atoms in the first UD. Fig. 11-9
illustrates both a simple Subset (N = Vertebrate) and a more complex
splitting of an Atom (N = Black Lab).
240
MATH WITHOUT NUMBERS
If we add N = Black Lab, then for the new combined UD we have the
following Partition Equation for P4 with X = { a, b, c, d } = { Black Lab,
Black Pug, White Cat, Brown Crab }:
U = X PE = Name Type of Animal1 In Type of Animal2 In
Color In Dwelling
241
WILLIAM S. VEATCH
Name: ( a + b + c + d ) In
Type of Animal1: ( Vertebrate abc + Invertebrate d ) In
Type of Animal2: ( Mammal abc + Crustacean d ) In
Color: ( Black ab + White c + Brown d ) In
Dwelling: ( Land abc+ Water d ).
For each of the five Partitions, we had to ask whether our new Atom fit
within an existing Equivalence Class, or whether we need to add a new
Equivalence Class. Since all we were doing was splitting an Atom from a
to ab, we did not have to change any of our Partitions; we just had to add
an additional Atom. Also, note that the length of the Chain increased from
a three Level Chain (Black Dog, Vertebrate, Animal ) to a four Level Chain
(Black Lab, Black Dog, Vertebrate, Animal).
11.4.4. Superset-Subset Universes of Discourse
In a sense, the results are the same as for Subset-Superset discussed above,
but in reverse, i.e., UD2 is the Superset and UD1 the Subset. We give a few
more examples, however, where a single Set is a SuperSet of our standard P3
example. Fig. 11-10 illustrates three different scenarios with a new UD2 :
N = { Black Dog, Brown Cow, White Cat, Brown Crab },
N = { Black Dog, White Cat, Brown Crab, Worm }, or
N = { Black Dog, White Cat, Brown Crab, Corn }.
242
MATH WITHOUT NUMBERS
Note that in each case, the original P3 Subset Lattice remains within the
P4 Superset Lattice. The Atoms remain unchanged, but notice how the
Labels in some cases have moved:
By adding Brown Cow, Vertebrate moved from ab to abd,
By adding Worm, Animal moved from abc to abcd, and
By adding Corn, Living Thing moved from abc to abcd.
243
WILLIAM S. VEATCH
If we add N = { Black Dog, White Cat, Brown Crab, Brown Cow },
then for the new combined UD we have the following Partition Equation for
P4 with X = { a, b, c, d } = { Black Dog, White Cat, Brown Crab, Brown
Cow }:
U = X PE = Name In Type of Animal1 In Type of Animal2 In
Color In Dwelling
=
Name: ( a + b + c + d ) In
Type of Animal1: ( Vertebrate abd + Invertebrate c ) In
Type of Animal2: ( Mammal abd + Crustacean c ) In
Color: ( Black a + White b + Brown cd ) In
Dwelling: ( Land abd + Water c ).
If we add N = { Black Dog, White Cat, Brown Crab, Brown Worm },
then for the new combined UD we have the following Partition Equation for
P4 with X = { a, b, c, d } = { Black Dog, White Cat, Brown Crab, Worm }:
U = X PE = Name In Type of Animal1 In Type of Animal2 In
Color In Dwelling
=
Name: ( a + b + c + d ) In
Type of Animal1: ( Vertebrate ab + Invertebrate cd ) In
Type of Animal2: ( Mammal ab + Crustacean c + Worm d ) In
Color: ( Black a + White b + Brown cd ) In
Dwelling: ( Land abd + Water c ).
If we add N = { Black Dog, White Cat, Brown Crab, Yellow Corn },
then for the new combined UD we have the following Partition Equation for
P4 with X = { a, b, c, d } = { Black Dog, White Cat, Brown Crab, Corn }:
U = X PE = Name In Type of Living Thing In Color In
Dwelling
=
Name: ( a + b + c + d ) In
Type of Living Thing: ( Animal abc + Plant d )
244
MATH WITHOUT NUMBERS
Animal:
(
Type of Animal1: ( Vertebrate ab + Invertebrate c ) In
Type of Animal2: ( Mammal ab + Crustacean c ) In
)
Color: ( Black a + White b + Brown c + Yellow d ) In
Dwelling: ( Land abd + Water c ).
11.4.5. Partially Overlapping Universes of Discourse
Similar to combining Subset-Superset or Disjoint Sets of Atoms, when we
have Partially Overlapping Sets we will need to check to see whether Partition
Equations remain valid. In a sense, POL is like a combination of Subset
and Disjoint if we break down the pieces. For the Subset portion, the key
question is whether or not we need to split any of the Atoms. For the
Disjoint portion, we simply add the additional Atoms to form the combined
Universe of Discourse.
Since we can view POL as a combination of Subset-Superset and
Disjoint, we do not give any further illustrations of combining POL
Universes of Discourse.
11.5. Summary of the Rules for Combining Universes of Discourse
While not a trivial task, we have shown that it is possible to combine
Universes of Discourse, and that we can develop rules for doing so. There
is an overall two-step process that we can follow:
First, compare the Base Sets of Atoms (or Deemed Atoms) for the
Universes of Discourse to be combined, and determine the Base Set
of Atoms for the combined Universe of Discourse. In some cases,
this may result in a larger Set of Atoms than if we simply added the
Atoms in the two Universes of Discourse, particularly if a Deemed
Atom in one UD is a Subset of, or Partially Overlapping with, a
Deemed Atom in another UD.
Second, we look at the impact of adding any new Atoms on any
Chains, Partitions, or Partition Equations. In fact, since Partition
Equations are really a combination of Chains and Partitions, we
really just need to look at the impact on Partition Equations.
While it may not be a trivial task to combine two or more UDs, it can be
245
WILLIAM S. VEATCH
done. There are a number of observations that we can make based upon the
examples given previously:
Identical UDs are important in areas such as translation or complex
areas where different experts may be using different terminology for
the same sets.
We can interpret POL UDs as a combination of Subset-Superset and
Disjoint UDs; and Superset-Subset is really the same as Subset-
Superset if we reverse the order (which we can do, because Addition
of Atoms or Power Sets of Atoms is Commutative). Therefore, of
the five possible ways in which two UDs can be related to one
another, we can focus our attention on two: Subset-Superset and
Disjoint.
Within Subset-Superset, there are two scenarios:
o No Atom Splitting: One UD may be a Subset of the other
without splitting any Atoms. If so, combining the UDs is
easy, as the combined UD will be the same as the component
UD that is the Superset.
o Atom Splitting: If we need to split Deemed Atoms, then
the number of Atoms in the combined UD will be greater
than the sum of the Atoms in the component UDs. The
impact on the Partition Equation is easy to handle, because
we simply take an existing Deemed Atom in the Partition
Equation and split it into two or more parts, but there no
other change to Equivalence Classes of the Partition
Equation.
If Disjoint, calculating the number of Atoms in the combined UD is
easy, because all we have to do is add the number of Atoms in the
UDs that we are combining.
We can simply the above observations into two principal rules to
remember when combining UDs:
If there is a Subset-Superset relationship (or POL = Subset +
Disjoint) between UDs, check whether there are any Deemed Atoms
that we need to split into component Atoms.
If there is a Disjoint relationship (or POL = Subset + Disjoint)
between UDs, then we need to check the placement of each new
Atom in any relevant Partition Equation. Sometimes this will be
easy, but sometimes it will require that we add new Equivalence
Classes to a Partition. If an Atom is added to an Equivalence Class,
then the Label for that Equivalence Class will move up in the Power
246
MATH WITHOUT NUMBERS
Set Lattice.
Once the Partition Equations have been checked and updated, they
remain valid. As a result, as Knowledge Representation structures are
mapped out and shared, over time the entire Universe of Ideas could, in
theory be mapped out for all time. The only variations would occur as new
Ideas are articulated. Communication could be improved, and fields of study
such as translation and artificial intelligence could benefit enormously.
This allows us to study areas of knowledge independently and map out
smaller component pieces of our Knowledge Base, and then assemble the
components to create a larger map of Knowledge, like combining pieces of
a puzzle to create the overall picture.
11.6. Update to the Definition of Idea
Unverse of Discourse: A Universe of Discourse is equivalent to a Set of
Atoms and Coatoms, and their related Power Sets.
We can represent an Idea by a Universe of Discourse or an Element of
a Universe of Discourse.
Relationships Between Universes of Discourse: Any two Universes
of Discourse are related in the same ways as any two Power Sets, i.e., the
related Base Sets of Atoms are related in one of the five ways that any two
Sets may be related: Identical, Disjoint, Subset-Superset, Superset-Subset,
and Partially-Overlapping.
11.7. References, Historical Notes, and Further Reading
See generally, the following books regarding Knowledge Representation and
conceptual structures:
[Sowa 2000]
[Brachman and Levesque 2004]
[Hitzler and Schrfe 2009]
[Kaburlasos 2006]
247
12. AXIOMS OF SET THEORY
One of our primary goals in the study of the Mathematics of Ideas has been
to see how far we can extend the concepts of Set Theory and related areas of
Mathematics to explain how we think, learn, and communicate. We have
seen that Set Theory explains a lot, although in the course of our studies we
have made some adjustments to traditional Set Theory, primarily to explain
Attributes in terms of Coatoms. In this Chapter, we provide a brief
introduction to traditional Set Theory, and examine how Set Theory as
applied to Ideas differs from traditional Set Theory, which is applied
primarily to numbers.
We will examine the classic Axioms of Set Theory that form the
foundation for the Mathematics of Numbers as well as our Mathematics of
Ideas. Our goal in identifying Axioms is to come up with a list of
fundamental principles from which we can derive the rest of our Mathematics
of Ideas. To qualify as a proper Axiom, ideally the Axiom would be
something that we view as self-evident, and beyond attack. Of course, we
must review our list of Axioms on a regular basis in light of new discoveries,
and update it as necessary. The Axioms form the foundation for the rest of
our Mathematics of Ideas, so if the foundation is flawed, the entire structure
could collapse.
As a simplifying assumption, we will assume that Sets are Finite, rather
than Infinite. This may seem like cheating, but it is necessary, at least in
the beginning, in order to make the mathematics manageable. We will
explore Infinite Sets in later volumes, but for now we can remove the most
difficult issues by working only with Finite Sets. There is one caveat,
however, i.e., that we allow ourselves to constantly revisit our choice of Finite
Atoms, and to expand the Set if we so desire. In this way, we retain the
248
MATH WITHOUT NUMBERS
benefits of Infinite Sets, without having to address the mind-boggling issues
of describing Infinite structures.
12.1. What are the Differences Between Traditional Set Theory
and MWN?
The main differences between traditional Set Theory and the Mathematics of
Ideas are that in MWN, we assume the following:
(i) We assume that the Universal Set (U) of all Ideas exists, and
that U is made up of Atoms.
(ii) We assume that Atoms exist, i.e., certain Ideas that cannot be
broken down further into component parts; other Ideas are
Compounds made up of Atoms. In this view, an Atom is not a
Set, but an Atom can be an Element of a Set. Cf. [Devlin 1993]
at 1 and 7, expressing the common view that everything is a Set,
and Sets are viewed as Sets of Sets. Later we will see that by
assuming the existence of Atoms, there is no need for the Axiom
of Foundation or, stated another way, assuming that Atoms exist
arguably gets us to the same result as the Axiom of Foundation.
See infra Subchapter 12.4.
(iii) Belonging ( ) applies only to Atoms as Elements of Sets,
i.e., the left side of must be an Atom and the right side must
be a Set; we can use Subset ( ) for Sets that are Subsets of
other Sets. Therefore, the statement a a is false, because a
cannot be both an Atom and a Set; the following statement,
however, is true so long as a is an Atom: a {a}.
(iv) We treat a Set of Atoms X, the Power Set of X P(X), and the
Power Set of Power Sets P(P(X)), as separate, but related,
Domains.
(v) We treat the Physical World, the World of Ideas, and the World
of Abstract Sets as three different Worlds that are related through
Mapping relationships.
(vi) We assume that in the World of Abstract Sets, Atoms and
Compounds do not have properties other than existence and
uniqueness. Rather, properties are treated as Mappings from
the Physical World or the World of Ideas, to Elements of the
Power Set of Atoms (or Coatoms) in the World of Abstract Sets.
(The only uncertain factor is the number of Atoms required to
accommodate all Ideas in U either as Atoms or as Compounds.)
(vii) We make a simplifying assumption that the Universe of Ideas is
Finite, but we reserve the right to revisit our choice of Atoms at
any time, and increase or decrease the number of Atoms if
249
WILLIAM S. VEATCH
necessary.
Note that perhaps the most significant difference between traditional Set
Theory and MWN, is Item (vi) above regarding how we treat properties
in MWN. By treating properties as Mappings and eliminating the concept
of Elements of abstract Sets having properties, we eliminate the need for
the Axiom of Class Construction and its various related forms: Axiom of
Selection, Axiom of Subset Selection, Property Axiom, Axiom of
Abstraction, and Axiom Schema of Separation. As noted in [Potter 2009] at
291, the Axiom of Separation has long been criticized for its vagueness. By
removing the concept of properties from the realm of Abstract Sets, we
have removed the vagueness and created mathematical certainty and
precision. The challenge becomes one of determining the proper Mapping
relationships among the Physical World, the World of Ideas, and the World
of Abstract Sets. This is something that will always be subject to debate and
possibly controversy; Mathematics can help us with how to think and
communicate with precision and clarity, but Mathematics cannot tell us
what to think. Determining the proper Mapping relationships will always
be a challenge.
250
MATH WITHOUT NUMBERS
Fig. 12-1 illustrates the three Object Domains, i.e., Atoms, the Power
Set of Atoms, and the Power Set of Power Sets. The three Domains are
related in an interesting and useful way. Although the Domain of Atoms
contains only Atoms, and not Sets, we are able to create Sets using Partitions.
To establish Binary Operations among Sets, however, we must move up to
the next Dimension, i.e., the Power Set Domain. Similarly, we can create a
Partition in the Power Set Domain that creates an Ideal. To establish Binary
Operations among Ideals, however, we must move up to the next
Dimension, i.e., the Power Set of Power Sets Domain. We saw an example
of this earlier in Subchapter 7.11.2.
251
WILLIAM S. VEATCH
12.2. The Is a Relation: Belongs to ( ) or Is a Subset of
( )?
Before we start looking at the Axioms of Set Theory, we take a slight
diversion to look at the meaning of is a as used in connection with Ideas
when expressed as Objects. Specifically, does is a mean belongs to the
Set ( ) or is a subset of ( )? Note that is usually assumed as a
valid Operation, and is defined in terms of as follows: Let A and
X be Sets. A X for every a, a A a X, and A X. See [Rodgers
2000] at 221 and 225.
First, we look at some of the Propositions that we believe, intuitively, to
be true in the Chain:
( a: ab: abc: abcd: abcde ) = ( Dog: Mammal: Vertebrate: Animal: Universe ),
where X = { a, b, c, d, e } = { Dog, Cat, Snake, Crab, Rock }:
A Dog is a Mammal.
A Dog is a Vertebrate.
A Dog is an Animal.
A Cat is an Animal.
A Mammal is an Animal.
One of the key questions becomes: is Mammal = { a, b } an Element of
Vertebrate, or are the Elements a and b Elements of Vertebrate? The
distinction is important because in Set Theory, Vertebrate = { {a, b}, c } is
not the same as Vertebrate = { a, b, c }.
The difficulty we have is that even what we call Atoms are really
Deemed Atoms in the World of Abstract Sets, because in theory we could
continue to subdivide Elements ad infinitum. At first, it would appear, then,
that Dog should be viewed as a Set. We could, however, view Dog as
being a proxy for a list of Elements, but not a Set in itself. For example, per
[Quine 1981] at 185 writing about Mathematical Logic, in Quines view each
cat is an animal, but the class of cats is not. This suggests that we should use
Subset rather than Belonging as the interpretation for the is a
relationship; but we have a dilemma because in theory every Set Element is
capable of subdivision. As Pinter points out (see [Pinter 2014] at 9) almost
every set is a set of sets.
252
MATH WITHOUT NUMBERS
One way to resolve this difficulty is to assume that we have a fixed Set
of Atoms (Deemed Atoms really), and that all Ideas are either Atoms or
Compounds made up of Atoms. That way, depending upon the Domain, all
Ideas are in the Universe of Ideas that is represented by the Power Set of the
selected Atoms. If we have an infinitely subdividable Set of Atoms, then
will not work, because there are no Atoms/Elements that we can
specify. The issue that we are struggling with is really a result of the Infinite
nature of Ideas. If we assume that Ideas are Finite, i.e., that Deemed Atoms
exist and that there is a Finite number of them, then we can interpret is a
as . Therefore, a Mammal is a Vertebrate if all of the Deemed Atoms
in the Set Mammal List are also Deemed Atoms of Set Vertebrate List:
Mammal Vertebrate, but
Mammal Vertebrate, because the Atoms that are Elements of
Mammal are also Elements of Vertebrate.
We see, then, that the question of whether is a means belongs to
( ) or is a subset of ( ) depends upon whether or not Atoms exist,
and whether or not we can identify them. Fig. 12-2 illustrates this point with
a familiar syllogism, the question being whether Socrates is an Atom. Cf.
[Kneebone 1963] at page 170-171, discussing the unit class concept used
by Whitehead and Russell to describe a class with a single object.
Do Atoms exist? For purposes of our Mathematics of Ideas, we take the
position that the answer depends upon whether we are working in the
Physical World, the World of Ideas, or the World of Abstract Sets. In the
Physical World, we think of Atoms existing in the case of tangible matter
253
WILLIAM S. VEATCH
made up of Atoms of Carbon, Oxygen, Hydrogen, Iron, etc. In the World of
Ideas, it is a more difficult question to answer. For example, the Idea of a
Dog can be broken down into breeds of Dogs. Labs can be broken down
further in Black, Chocolate, and Yellow Labs. What about my Black Lab
named Lucy? In theory, we could distinguish Lucy as a puppy from Lucy in
her middle age; in fact we could distinguish a different Lucy at every moment
in time. Since time is, arguably, infinitely divisible, we end up with an infinite
number of Lucys. We have a similar problem with Abstract Sets; there is no
reason to believe that an abstract element of a Set cannot be subdivided ad
infinitum. There is also, however, no reason that we cannot, for the sake of
argument, assume that Atoms do exist. That is the approach that we choose
to take. We assume that Atoms exist in the World of Abstract Sets, but we
constantly revisit our choice of Atoms and how they Map to Objects in the
Physical World and in the World of Ideas, and if necessary we revise and
update our choice of Atoms, both the number of Atoms and how they map
to Ideas.
In summary, going forward, for purposes of the Mathematics of Ideas,
when we use a statement in the form a X, we take the position that a
must be an Atom, and X must be a Set. If a is a Set, then we would write
a X or preferably A X using a capital letter to indicate that A is a
Set. Later, in Subchapter 12.4, we will see that this interpretation of
helps us explain the Axiom of Foundation; and we will see that it helps us to
avoid Russells Paradox for the time being.
12.3. The Importance of Specifying the Domain
One of the recurring themes in our study of the Mathematics of Ideas is that
we must always specify the Domain in which we are working. We have rules
for converting Sets and Partitions Equation back and forth from one Domain
to another, but we must exercise care not mix calculations across Domains.
This is also true in our study of Set Theory; in the Mathematics of Ideas, we
formulate rules and axioms of Set Theory that are specific to a particular
Domain. As illustrated in Fig. 12-3, in our Mathematics of Ideas we find it
helpful to assume that there are three Worlds: the Physical World, the World
of Ideas, and the World of Abstract Sets. Within the World of Abstract Sets,
there are six related Domains: Atoms (X), Coatoms (Y), Power Set of Atoms
(P(X)), Power Set of Coatoms (P(Y)), Power Set of Power Sets of Atoms
(P(P(X))), and Power Set of Power Sets of Coatoms (P(P(Y))).
We sometimes refer to Dimensions of Sets, where the Base Set X
of Atoms (or Coatoms) is Dimension 1; the Power Set P(X) is Dimension
2; and the Power Set of Power Sets P(P(X)) is Dimension 3. The
254
MATH WITHOUT NUMBERS
Dimensions that we refer to are similar to the hierarchies of Sets described
in certain books on Set Theory. See e.g., [Devlin 1993] at 36 et seq.; and at 47,
discussing iteratively constructing new sets from old ones.
If we are working in the Domain of the Set of Atoms, then the
Elements in the Universe are Atoms, not Sets of Atoms. We can define
255
WILLIAM S. VEATCH
an Equivalence Relation on a Set of Atoms, because an Equivalence Relation
is defined in terms of Atoms, rather than Sets of Atoms. For example, on
the Set X = { a, b, c }, we can define an Equivalence Relation that creates
the Partition [a]=[b] = { a, b}, and [c]={c}.
A Binary Operation such as Union, however, is not well-defined on a
Domain of Atoms. For example, a b = ab is not defined within our
Domain, because the Element ab does not appear within our Domain X =
{ a, b, c }. In effect, we can create Partitions of our Domain of Atoms, but
we cannot apply Binary Operations or Laws of Composition.
If the Domain is the Power Set of Atoms, then the Elements of our
Domain are all Subsets of the Base Set of Atoms, including the Atoms
themselves: P(X) = ( , a, b, c, ab, ac, bc, abc ). Now, we can define a Binary
Operation or Law of Composition that is Closed on the Domain; a b = ab
is well-defined within our Power Set Domain, because ab is an Element of
the Domain.
Alternatively, we can form a Domain that is a Power Set of Power Sets
of Atoms, in which case the Elements of our Domain are all Power Sets,
which we sometimes refer to as Ideals in the case of Atoms (or Filters in the
case of Coatoms). As with the Power Set of Atoms, the Operations of
Addition and Multiplication on P(P(X)) are well-defined and Closed within
our Domain.
12.4. Axioms of Set Theory
When we speak of Axioms of Set Theory, of which world are speaking: the
Physical World, the World of Ideas, or the World of Abstract Sets? We start
by looking at Axioms that apply in the World of Abstract Sets, and then
examine to what extent we can apply those Axioms to either the Physical
World or the World of Ideas. In this book we provide a brief overview; later
volumes in the MWN series will examine the Axioms of Set Theory in more
detail.
There are many different statements of the Axioms of Set Theory, but we
choose to start with the Axioms as articulated by [Pinter 2014], which are a
slightly modified form of von Neumanns system of Axioms (see [Pinter 2014]
at 13), with further refinement based upon [Stewart Tall 1977] at 254 et seq.,
[Abbott 1969] at 8 et seq., [Halmos 1960] at 1 et seq., [Tiles 1989] at 118 et seq.,
[Rodgers 2000] at 277 et seq., and [Bourbaki 1968]:
256
MATH WITHOUT NUMBERS
A1. Axiom of Extent (or Extensionality). If A and B are equal,
then they have the same Elements. [Pinter 2014] at 25.
A2. Axiom of Class Construction. There exists a class S which
consists of all the Elements x which satisfy (x), where (x) is
a statement expressed entirely in terms of the symbols , , ,
, , , , brackets, and variables x, y, z, A, B, [Pinter 2014]
at 26-27, 223.
o Note: Axiom of Selection used in the Zermelo-Fraenkel
(ZF) axioms of Set Theory is more conservative: S= { x
U: (x) } [Pinter 2014] at 223-224. Cf., the Axiom of Class
Existence in [Stewart and Tall 1977] at 254-255.
o See also, the Property Axiom: There exists a set S such that
S = { x | x U and (x)}, where C is a Set and (x) is an
open statement. The notion of a Universal Set is implicit in
each property definition of a Set. See [Rodgers 2000] at 279.
o See also, the Axiom of Abstraction: [Abbott 1969] at 11.
o See also, the Axiom Schema of Separation [Suppes 1972]
at 56.
o See also, [Bourbaki 1968] at 347, stating that Elements of Sets
are capable of possessing properties.
o Axiom of Constructibility. [Pinter 2014] at 227, and
[Devlin 1993] at 37 and 120 et seq.
o Axiom of Subset Selection. [Devlin 1993] at 38.
Amongst the Sets in P(X), is the Set of all those Elements
x of X for which (x).
Set Axioms (i.e., axioms that establish the properties of Sets):
o A3. Intersection. Every subclass of a Set is a Set; the
Intersection of any two Sets is a Set. [Pinter 2014] at 45.
o A4. Null Set. is a Set. [Pinter 2014] at 45.
o A5. Pairs. If A and B are Sets, then { A, B } is a Set. [Pinter
2014] at 45.
o A6. Union. If A is a Set of Sets, then X
X A
is a Set. [Pinter
2014] See also, the Sum Set Axiom in [Tiles 1989] at 125.
o A7. Power Set. If X is a Set, then the Power Set of X, P(X),
is a Set. [Pinter 2014]
257
WILLIAM S. VEATCH
A8. Axiom of Foundation: If A is any Set, there is an Element a
A such that a A = . See [Pinter 2014] at 218. See also, the Axiom
of Regularity discussed in [Stoll 1963] at 305, used to establish that
{a} a= , and therefore that a a; or [Hrbacek and Jech 1999] at
259: there is a minimum Element. See [Devlin 1993] at 44 and
11: is a well-founded relation, meaning that every non-empty
subset has a minimal Element.
o See also, [Potter 2009] at 291:
A ( x A)(x is an individual/Atom or x
A=).
Note that Intersection is defined in terms of
belonging: x A=a is defined as a x and
a A. However, in MWN we require that the left
side of be an Atom and the right side be a Set.
We started with the assumption that x A, so
x must be an Atom. Therefore, a x is
undefined, and as a consequence x A=a is
undefined as well. Therefore, we can safely say that
x A=.
o See also, [Suppes 1972] at 53 for a very helpful discussion of
the Axiom of Regularity. The goal of the Axiom of
Regularity is to prevent infinite descending sequences of
Sets. This is also the goal of our assumption that Atoms
exist in MWN; for to have meaning, we need to assume
that at some point an Atom exists, i.e., an Element that is
not subdividable any further.
A9. Axiom of Replacement. If A is a Set and f: A B is a
surjective function, then B is a Set. See, [Pinter 2014] at 70. See also,
[Devlin 1993] at 41.
A10. Axiom of Choice. Every Set has a choice function. [Pinter
2014] at 113.
A11. Axiom of Infinity. There exists a successor Set. See, [Pinter
2014] at 125. See also, [Devlin 1993] at 42: X: X and {a} X
whenever a X.
A12. Axiom of Cardinality. See, [Pinter 2014] at 151.
A13. Axiom of Ordinality. See, [Pinter 2014] at 168.
Now, we outline the framework for MWN, and then in the Tables in the
Subchapters below we show where the Axioms of Set Theory come into play:
258
MATH WITHOUT NUMBERS
The Mathematics of Ideas
I. Domain: Atoms (X); Coatoms (Y)
A. The Universal Set:
The Universal Set of all Ideas exists.
The Universal Set has a single Element or Deemed
Atom.
B. Subdivision and Consolidation: We can Subdivide the
Universal Set ad infinitum; and once Subdivided, we can
Consolidate Atoms in any combination. (Note: by
Subdividing and Consolidating we are changing the number
of Deemed Atoms, but we are not creating Sets.)
C. Partitions: Once we have Subdivided the Universal Set to
arrive at the desired number of Deemed Atoms, we need
the ability to create Partitions of this Set of Atoms. We
assume the existence of Equivalence Relations, which are
equivalent to Partitions. Each Equivalence Class of the
Partition is a Set.
D. Coatoms: The moment we assume the existence of an
Atom a by Subdivision of U, we have also created a
Coatom, i.e., the rest of the Universe equal to not-a =
~a = 1. If there are more than two Atoms, then
not-a represents the Equivalence Class of a Partition
containing all Atoms other than a.
II. Domain: Power Set of Atoms P(X); Power Set of Coatoms P(Y)
A. The Universal Set and the Empty Set:
The Empty Set exists.
The Empty Set has no Elements.
P(U) = ( , U ); P()=( ).
B. Subdivision and Consolidation: Given a Set X of
Deemed Atoms of Dimension 1, we can create a Power Set
P(X) of Dimension 2. Subdivision and Consolidation
occur with reference to the Atoms in the Base Set X.
C. Set Operations: We define Set Operations or Laws of
Composition (i.e., Union, Symmetric Difference,
Intersection, Set Subtraction, and Complement) on the
Elements of our Power Set Domain. By defining Set
Operations on the Elements of the Power Set, we can create
mathematical structures such as a Group, Boolean Algebra,
Boolean Ring, Boolean Lattice, and Topology.
D. Coatoms: Defining a Power Set of Atoms necessarily
results in a Dual Inverse Power Set of Coatoms.
259
WILLIAM S. VEATCH
III. Domain: Power Set of Power Sets of Atoms P(P(X)); Power Set
of Coatoms P(P(Y))
A. The Universal Set and the Empty Set:
P(P(U)) = P( , U ) = ( , { , U} ).
P(P()) =( ).
Note that repeated application of the Power Set
Operation, as we define it in MWN, does not
change the number of Elements.
B. Subdivision and Consolidation: Given a Set X of
Deemed Atoms of Dimension 1, we can create a Power Set
P(X) of Dimension 2; and a Power Set of Power Sets
P(P(X) of Dimension 3.
C. Set Operations: We define Set Operations and Laws of
Composition (i.e., Union, Symmetric Difference,
Intersection, Set Subtraction, and Complement) on the
Elements of our Power Set of Power Sets Domain. By
defining Set Operations on the Elements of the Power Set
of Power Sets, we can create mathematical structures such
as a Group, Boolean Algebra, Boolean Ring, Boolean
Lattice, and Topology.
D. Coatoms: Defining a Power Set of Power Sets of Atoms
(Ideals) necessarily results in a Dual Inverse Power Set
of Power Sets of Coatoms (Filters).
IV. Converting Sets and Partition Equations from One Domain to
Another Within the World of Abstract Sets
Given a Set in any one of the six Domains, we can calculate the
corresponding Set in each of the other five Domains. For example,
the Power Set Unary Operation will convert Atoms and Partition
Equations in the Domain of Atoms, to Power Sets and Power Set
Partition Equations in the Domain of Power Sets.
V. Mappings across the Three Worlds: Physical, Ideas, and
Abstract Sets
A. In MWN, Atoms and Coatoms in the World of Abstract
Sets do not have properties or Attributes; Atoms and
Coatoms simply exist and are unique. We can, however,
create a Mapping from Attributes in the Physical World or
World of Ideas to the Power Set of Coatoms in the World
of Abstract Sets.
B. To illustrate the significance of the Mappings, in the World
of Abstract Sets all Chains are equal. In the Mapping from
260
MATH WITHOUT NUMBERS
the Physical World to the World of Abstract Sets, however,
the first Chain below has significance, whereas the second
does not:
(a: ab: abc: abcd) =
( {Dog}: {Dog, Cat}: { Dog, Cat, Snake }: {Dog,
Cat, Snake, Crab} ) =
(Dog: Mammal: Vertebrate: Animal)
(d: db: dbc: dbca) =
( {Crab}: { Crab, Cat}: { Crab, Cat, Snake }:
{ Crab, Cat, Snake, Dog} )
12.5. Domain: Atoms (X); Coatoms (Y)
12.5.1. The Universal Set
We start with the assumption that the Universal Set exists. See Table 12-1.
We can avoid much confusion in our study of the Mathematics of Ideas
if we always start with the Universal Set of all Ideas before creating
Subdivisions.
12.5.2. Subdivision and Consolidation
Once we have established the that Universal Set exists, we need to Subdivide
the Universal Set in order to create a number of Atoms that we can use to
generate a structure that matches what we observe in the Physical World or
imagine in the World of Ideas. Since Atoms in the World of Abstract Sets
have no properties or characteristics other than existence and uniqueness, the
only thing that matters at this point is the number of Atoms that we create
by Subdivision. Of course, once we create a Mapping to the Physical World
or World of Ideas, then we have put a limitation on our use of that particular
Atom, if we want future Mappings to be consistent with earlier ones. For
261
WILLIAM S. VEATCH
now, however, establishing the quantity of Deemed Atoms is the primary
goal.
Note that our MWN version of the Axiom of Extensionality is slightly
different from that of traditional Set Theory, in the we require t to be an
Atom (or Coatom), so that t X and t Y are well-defined.
In traditional Set Theory, we assume the existence of (A4. Null Set
Axiom) and a Successor Set (x) (A11. Axiom of Infinity). See [Pinter
2014] at 125 and [Stewart and Tall 1977] at 160 et seq. In this way, we can
create the Natural Numbers:
=0,
(0)={ }=1,
262
MATH WITHOUT NUMBERS
(1)={ ,{}}=2,
(2)={ , { }, {, {}} }=3,
See Subchapter 12.13 below.
In MWN we take a different approach, because we need to form a
Lattice of Objects, not just an Ordered Chain of numbers. An easy way to
do this is as follows:
Start with the existence of U, the Universe of all Ideas; and
Assume that we can subdivide U ad infinitum into sub-Atoms that
we refer to as Deemed Atoms, or simply Atoms for
convenience.
In other words, in MWN we always start with the Universal Set U
representing all Ideas. We create an infinite number of Atoms by
Subdivision of U. In a sense, this is the reverse of starting with the Empty
Set and generating an infinite number of Atoms by using the Successor
Set function.
We can reconcile generating natural numbers with the Successor
Function and Subdivision of the Universal Set as follows: let X equal
the Ordered Set starting with , where each succeeding Element is the Set
of what comes before (i.e., by applying the Successor Function):
X = ( , {}, {, {}}, { , {}, { , {}}} )
= ( , 1, 12, 123, )
By letting {} equal a, {{}} equal b, {{{}}} equal c,
etc., we get the following:
X = ( , a, ab, abc, ).
What we see is that we can use the Axiom of Infinity to generate a series
of Atoms. In each case, i.e., numbers or Atoms, we generate an Infinite
number of Objects that we can use to generate Sets. In summary, although
we prefer to use the concept of Subdivision of the Universal Set in MWN
in order to generate potentially Infinite Sets, we could also use the more
traditional application of Axiom of Infinity and the Successor Set to generate
an Infinite number of Objects.
263
WILLIAM S. VEATCH
12.5.3. Partitions and Equivalence Classes
We can create Partitions of the Atoms in the Universal Set.
Earlier, we noted that we can create a Partition of Atoms (or Deemed
Atoms) in the Universe. Creating Partitions is equivalent to an Equivalence
Relation that creates Equivalence Classes. A Partition does create Sets or
Classes, but it is not really an Operation in the sense of Union or
Intersection, i.e., it is not a Law of Composition that combines two Elements
to map to a third. Rather, a Partition simply divides a single Set of Atoms
into parts that we refer to as Equivalence Classes. We save Operations for
the Power Set Domain (or Power Set of Power Sets Domain), where we have
a Domain that is structured in such a way that we have one or more Closed
Operations.
12.5.4. Coatoms
As always, we start with the assumption that the Universal Set of all Ideas
exists. Next, suppose we postulate the existence of an Atom a. By
necessity, we have actually postulated the existence of two Atoms: a and
~a. The reason for this, assuming that U exists, is that the only way to
create a is to Subdivide U into two parts, a and ~a. We refer to ~a
(read not a) as a Coatom relative to a. In order to distinguish Atoms
from Coatoms more clearly, we relabel ~a as 1. If we subdivide the
Universal Set further, we label ~b as 2, ~c as 3, etc. See Fig. 12-4.
264
MATH WITHOUT NUMBERS
Note that when we Subdivide the Universe, we typically represent each
Atom with a lower case letter a, b, c,, but when we get to the last Atom,
we sometimes represent it with the capital letter R. The reason for this is
to remind us that there are always additional Ideas that may be represented
by Atoms, including Ideas that we have never spoken aloud, written down,
or even imagined. R represents the remainder or residual Atoms all
consolidated into a single Deemed Atom. As we saw earlier, this concept is
most helpful when we are combining Universes of Discourse. See Table 12-
4.
As discussed earlier, given that Coatoms can be derived from Atoms,
the reader might question whether we really need both Atoms and Coatoms.
The answer, however, is that Coatoms are extremely useful to us as a proxy
265
WILLIAM S. VEATCH
for properties in the Physical World or in the World of Ideas, which we
refer to in MWN as Attributes. Also, there are fewer Coatoms (than Atoms)
at the top of the Power Set Lattice, and fewer Atoms (than Coatoms) at the
bottom, so we can simplify calculations by using whichever Set is smaller.
12.6. Domain: Power Set of Atoms P(X); Power Set of Coatoms
P(Y)
In MWN, we treat the Power Set Operation as an Operation that converts
(or Maps) Atoms (in the Domain of Atoms, Dimension 1) to Sets of Atoms
or Compounds (in the Power Set Domain, Dimension 2). We illustrate in
Subchapter 12.6.2, that we can apply the Power Set Operation to complex
Partition Equations as well as simple Sets. We also can use the Power Set
Operation to convert (or Map) Sets of Atoms or Compounds (in the Power
Set Domain, Dimension 2) to Ideals (in the Domain of Power Sets,
Dimension 3).
12.6.1. The Universal Set and the Empty Set in the Power
Set Domain
We start with the assumption that the Universal Set and the Empty Set exist.
The Power Set Operation results in 2n Elements, including the Empty Set
(). Therefore, we need to assume that the Empty Set exists. See Table 12-
5.
In the Power Set Domain, the Universal Set and the Empty Set are duals
of one another, so by assuming the existence of one, we are, arguably,
assuming the existence of the other. Although we assume the existence of
the Universal Set and the Empty Set in the World of Abstract Sets, it is
difficult to truly understand the concepts of infinity and nothingness in the
Physical World. In any event, by starting with the concepts U and ,
we are essentially starting with the concepts of 1 and 0, which are
essential in Boolean Lattice theory, as the top or Supremum of the
Lattice is referred to as 1; and the bottom or Infimum of the Lattice
is referred to as 0. Table 12-6 illustrates some of the properties and
relationships of the Empty Set.
266
MATH WITHOUT NUMBERS
12.6.2. Subdivision and Consolidation in the Power Set
Domain
As in the Domain of Atoms, we can Subdivide and Consolidate Power Sets.
The easiest way to do so is by Subdividing or Consolidating the related Base
Sets of Atoms, and then converting to Power Sets.
267
WILLIAM S. VEATCH
In MWN, we view the Power Set (formed from a Base Set of Atoms) as
a separate Domain from regular Sets of Atoms (i.e., the Base Set), and in fact
there is a one-to-one Mapping from Partitions of regular Sets to Power Sets,
which we refer to as a Power Set Expansion or Power Set Contraction. This
Mapping relationship is extremely important, because if we are performing
calculations on Power Sets (e.g., Addition, Multiplication, etc.), it is much
268
MATH WITHOUT NUMBERS
easier to perform a Power Set Contraction (i.e., the inverse of a Power Set
Expansion) and then perform the calculations on the Sets of Atoms. When
the calculations are complete, we perform a Power Set Expansion to return
to Power Sets. In this way, we can simplify greatly, working with Power Sets,
which have the advantage of representing all Ideas in the Universe of Ideas,
whether Atomic or Compound.
12.6.3. Set Operations in the Power Set Domain
By defining a Power Set Domain consisting of all Subsets of a Base Set of
Atoms, we find that we can define Operations and Laws of Composition that
are Closed within the Domain, including Union, Intersection, and Symmetric
Difference. (In the Domain of Atoms, there were no Sets in the Domain,
just Atoms, so we could not define a Law of Composition.)
269
WILLIAM S. VEATCH
12.6.4. Coatoms in the Power Set Domain
Fig. 12-5 illustrates Atoms and Coatoms in a Lattice structure as we
Subdivide the Universe to create one, two, or three Atoms, along with their
related Coatoms.
270
MATH WITHOUT NUMBERS
In MWN, we show that we can map Attributes to the inverse dual Set of
Coatoms. The practical advantage is that it explains the distinction between Objects
and Attributes that we use in practice, and shows the inheritance of Attributes.
Therefore, whereas in Set Theory we talk of one Set, in MWN we always speak in
terms of dual Sets Atoms and Coatoms, and their related Power Sets.
Note that all of the rules and Axioms relating to Sets apply both to Sets
of Atoms and Sets of Coatoms.
271
WILLIAM S. VEATCH
12.7. Domain: Power Set of Power Sets of Atoms P(P(X)); Power
Set of Power Sets of Coatoms P(P(Y))
Applying the Power Set Operation a second time to a Power Set changes
each Element to a Power Set. For example, (a) becomes (, a ); (ab)
becomes (, a, b, ab). Note, however, that as we define the Power Set
Operation in MWN, applying it a second time does not increase the number
of Elements in the Domain; i.e., P(X) and P(P(X)) both have 2n elements,
where n is the number of Elements in the Base Set X.
12.7.1. The Universal Set and the Empty Set in the P(P(X))
Domain
Table 12-10 illustrates the Universal Set and the Empty Set in the P(P(X))
Domain. The key point here is the the number of Elements stays the same
when we move from the P(X) to the P(P(X)) Domain.
12.7.2. Subdivision and Consolidation in the P(P(X))
Domain
Subdivision and Consolidation always take place at the Atomic (Coatomic)
level. By applying the Power Set Operation once we can convert Atoms
(Coatoms) to a Power Set; and by applying the Power Set Operation twice
we can convert Atoms to Ideals and Filters.
272
MATH WITHOUT NUMBERS
12.7.3. Set Operations in the P(P(X)) Domain
We can define Operations of Addition and Multiplication for Ideals and
Filters. In the Power Set Domain, we Added and Multiplied Elements of the
Power Set, each of which was a Set; in the Power Set of Power Sets Domain,
we Add and Multiply Ideals, each of which is a Power Set. In each case, the
simplest approach is usually to apply a Power Set Contraction Operation (the
Inverse of the Power Set Expansion Operation), apply the Operation at issue
(i.e., Union, Symmetric Difference, Intersection etc.), and then apply the
Power Set Expansion Operation to convert back to a Power Set of Ideals.
In Subchapter 7.11.2, we saw an example of Set Operations in the
P(P(X)) in connection with our study of Complements. We could partition
elements in a Power Set to create an Ideal in the Dimension 2 Power Set.
However, to create Closed Operations of Addition ( ) and Multiplication
( ) of Ideals, we had to move to the Dimension 3 Power Set of Power Sets.
See Fig. 7-8.
12.7.4. Coatoms in the P(P(X)) Domain
As with the Power Set Domain, the Power Set of Power Sets Domain of
Ideals (or Downsets in Lattice Theory terminology) has a dual Power Set. In
the P(P(X)) Domain, we refer to the Dual Power Set as a Power Set of Filters
(or Upsets in Lattice Theory terminology).
As noted earlier, converting a Power Set Domain to a Power Set of
Power Sets Domain does not increase the number of Elements in the
Domain. Rather, it converts each Element from a Set to a Power Set.
12.7.5. Converting Sets and Partition Equations from One
Domain to Another Within the World of Abstract Sets
By applying the Power Set Expansion Operation, or its Inverse, the Power
Set Contraction Operation, we can convert from any one of our six principal
Domains to any other principal Domain. The key is to keep track of which
Domain we are working in, as it is important not to try to perform Operations
on a mix of Elements from different Domains.
As a practical matter, one of the most useful features of the Power Set
Expansion and Contraction Operations is that we can apply these Operations
not only to a Set of Atoms, but also to a Nested Partition Equation of Atoms.
273
WILLIAM S. VEATCH
This allows us to perform relatively complex Operations on Ideals and Filters
with great simplicity, by performing the Operations on the related Partitions
of Atoms and then converting back to Ideals and Filters.
Combining these techniques with the principles of Subdivision and
Consolidation of Atoms, we can further simplify Operations by
Consolidating Atoms and working with Deemed Atoms, and then later
Subdividing to return to our original Atoms. We start to see that there are
many techniques we can develop to simplify working with Ideas as Power
Sets, without losing any of the detail inherent in the Power Sets which grow
in size at an extraordinary rate as we add new Atoms.
12.8. Mappings Across the Three Worlds: Physical, Ideas, and
Abstract Sets
By creating a Mapping from the World of Abstract Sets to either the Physical
World or the World of Ideas, we can avoid the need for the Axiom of Class
Construction. Stated another way, in the World of Abstract Sets the concept
of a property of an Idea is represented by an Element in the Power Set of
Coatoms. As we discussed in Subchapter 2.4, each Atom in the Universe
of Ideas either has a given property, or does not have such property.
Therefore, since all possible subsets of Atoms in the Universe are represented
in the Power Set of Atoms/Coatoms, every property must correspond to
some Element in the Power Set. We map properties, which we refer to as
Attributes, to the Power Set of Coatoms (rather than Atoms), because the
Power Set of Coatoms allows us to track the Inheritance of Attributes. The
exercise of Class Construction becomes one of creating a Map from the
Physical World or World of Ideas to the World of Abstract Sets. Table 12-
11 summarizes the principles of the Mapping from the Physical World or
World of Ideas, to the World of Abstract Sets.
274
MATH WITHOUT NUMBERS
In mathematical terms, there is a Homomorphism among the
Domain of Atoms, the Power Set Domain, and the Power Set of Power Sets
Domain. (For a discussion of Homorphisms generally, see [Abbott 1969] at
117.) By establishing that there is a Homomorphism, we are assured that we
can convert from one Domain to another, and back again. Table 12-12
summarizes the six principal Domains for a three-Atom Universe of
Discourse, and illustrates the representation in the six Domains of the Idea
{ a, b } with two Atoms, assuming a Base Set X = { a, b, c, d }.
275
WILLIAM S. VEATCH
Note that by assuming the existence of a Universal Set that we can
Subdivide into Atoms, and treating properties as Mappings, we have
downplayed the concept of explaining Set Theory in terms of a language of
logical symbols. Cf. [Devlin 1993] at 30, describing Set Theory in terms of a
language describing Sets; and [Givant and Halmos 1998], using a language
of logical symbols to develop a Boolean logic.
12.9. What Exactly is an Object?
In the World of Ideas, each Idea is like a two-sided coin where one side is a
List Set of Objects which are examples of the Ideas; and the other side is a
Test Set of Attributes which are elements of the test for classifying an Idea.
What is important to note is that Objects and Attributes are related to one
another in an inverse fashion. In the World of Abstract Sets, we have Atoms
and Coatoms, which are also related to one another in an inverse fashion. It
is this similarity of Objects/Attributes and Atoms/Coatoms, and the
fact that in theory we can create a one-to-one mapping from
Objects/Attributes to Atoms/Coatoms, that allows us to define Objects
as Atoms or Compounds of Atoms (and Attributes as Coatoms or
Compounds of Coatoms) in the World of Abstract Sets.
12.10. What Exactly is an Attribute?; Demystifying the Concept
of a Property
In the study of Set Theory, it often is noted that we can form a Set by
grouping together Elements that share a common property. See e.g.,
[Hrbacek and Jech 1999] at 3. We take a different view in our study of MWN
by stating that Atoms (or Compounds of Atoms) in the World of Abstract
Sets do not have inherent properties, other than existence and uniqueness.
In the World of Abstract Sets, a property, which we refer to in MWN as
276
MATH WITHOUT NUMBERS
an Attribute, is simply a Partition of the Universe that creates a combination
of Atoms taken from the Power Set of Atoms for which the property is
True; and a combination of Atoms (i.e., the Complement of the True Set)
taken from the Power Set of Atoms for which the property is False. As
discussed earlier, every possible Attribute must, by definition be represented
by a single Element in the Power Set. The proof is obvious intuitively. Every
Object Atom in the World of Ideas (or the Physical World) either has or does
not have any given Attribute. This means, then, that there is a functional
relationship between any Attribute and the Power Set of Atoms, i.e., each
Attribute maps to a single member of the Power Set of Atoms consisting of
the largest Set of Atoms where each Atom exhibits the desired property. In
turn, each Element of the Power Set of Atoms maps to a single Element of
the Power Set of Coatoms. (We prefer to map Attributes to Coatoms,
because we can easily see the inheritance of Attributes.) This is a very
powerful observation, because it allows us to take the mystery out of
concepts like a property when we are studying pure math. In the Physical
World or the World of Ideas, we have properties such as color, size, shape,
etc., i.e., Attributes in MWN terminology, but in the World of Abstract Sets
we simply have Elements of a Power Set!
12.11. Russells Paradox
In MWN, the key to avoiding Russells Paradox is to assume that in the
statement a X, a must be an Atom (or Deemed Atom) and X must
be a Set. In this way, a Set can never be a member of itself (or any other Set
for that matter). Rather, only Atoms (or Coatoms) can be members of Sets
( ); whereas a Set can be a Subset of another Set ( ). Cf. [Stewart and
Tall 1977] at 253, where the authors use the Axiom of Class Existence to
avoid Russells paradox.
We avoid Russells paradox in MWN by: (i) assuming that our Universe
has a Finite number of Atoms (although the exact number can fluctuate), and
(ii) not allowing Atoms to have properties other than Existence and
Uniqueness, which in turn lead to Identifiability. When speaking of a
property of an Object, in MWN we interpret this to mean a Mapping from
an Idea in the World of Ideas to one of 2 n-m Elements of the Power Set of
Coatoms (where n is the number of Atoms in the Universe, and m is the
number of Atoms in the Atom or Compound in question). We also need to
state at all times, the Domain in which we are working. We interpret the
Belonging Relation a X to mean that an Atom a is a member of a Set
X.
277
WILLIAM S. VEATCH
Lets test this out with Russells Paradox: S = {x| x x}. Let x = S;
either S S or S S. Under the rules that we have established for MWN,
however, in the expression S S, the S on the left must be an Atom and
the S on the right must be a Set. Therefore, since S is a Set, neither the
expression S S nor the expression S S has any meaning, since S
cannot be both Set and an Atom.
Lets look further at how we would apply Russells Paradox in the
various principal Domains, starting with the Domain of Atoms. Fig. 12-6
illustrates the Domain of Atoms, the Power Set Domain, and the Power Set
of Power Sets Domain, where the Universe has three Atoms. While a three-
Atom Domain may seem like a simple example, it serves as a good model for
larger Domains, since we are assuming that the Domain has a Finite number
of Atoms.
In the Domain of Atoms, clearly there are only Atoms in the Domain,
so each Element must be an Atom. Table 12-13 provides some examples of
statements that are True in the Domain of Atoms.
278
MATH WITHOUT NUMBERS
Table 12-14 provides some examples of statements that are True in the
Power Set Domain.
279
WILLIAM S. VEATCH
Table 12-15 provides some examples of statements that are True in the
Power Set of Power Sets Domain.
280
MATH WITHOUT NUMBERS
12.12. What is a Number?
As we discuss the meaning of terms like property or Attribute in terms of
Set Theory, a natural question to ask is whether a number like 1, 2, 3 is
an Attribute? It turns out that we use numbers in two distinct ways: as
Attributes, but also as a way of counting the number of Elements in a Power
Set or at a Level in a Power Set. In this Subchapter, we look at the latter, and
then in Subchapter 12.13 we will look at numbers as Attributes.
To start, in a Power Set every Atom has a quantity of 1. When we
speak of a quantity of 2, what we are really saying is that we have two
distinct Objects that exist and are unique, i.e., two Atoms. As it turns out,
there is a way to interpret what a number is relative to a Power Set Lattice.
Natural numbers correspond to the number of Atoms in each Set at a Level
of a Power Set of Objects. Working from the bottom up in the P3 Power Set
with three Objects, we have 0 = ; 1 = a, b, or c; 2 = ab, ac, or bc; 3 = abc.
We are so accustomed to writing 2 + 2 =4, that we forget that implicitly
we are saying that there are four objects { a, b, c, d } and 2 + 2 really means
{ a, b } + { c, d}, or one of the other possible combinations such as { a, c }
281
WILLIAM S. VEATCH
+ { b, d }. Whichever combinations we add together, we end up with { a, b,
c, d } with four Elements, hence 2 + 2 = 4. If all we are concerned about is
quantity, then it makes sense to simply write 2+2=4; the other Attributes
of the Elements are irrelevant. In MWN, however, we do care about the
other Attributes, and sometimes the quantity as well.
As shown in Fig. 12-7, we can use a Partition Lattice to illustrate all
possible combinations of Elements the sum of which is four. See [Parrochia
and Neuville 2013] at 60.
Note that when we use numbers to describe the number of Atoms (or
Coatoms) in a Level of a Power Set, we are not using numbers as Attributes.
This is particularly clear when you notice that the number of Atoms in a Level
is not inherited by lower Levels, as would be the case with Attributes. This
is highlighted by the classic problem of the twelve Apostles. Whereas we can
say that each Apostle is pious, we cannot say that each Apostle is twelve.
Twelve is a property of the class of Apostles, but it is not an Attribute
inherited by each Apostle. See [Quine 1981] at 237. We know that there were
twelve Apostles, i.e., disciples of Christ. The number twelve might at first
appear to be an Attribute of the Idea Apostle, since there were twelve of
them. The problem, however, is that we know Attributes are inherited by
each Object in the Set. This means, therefore, that each Apostle, e.g., John,
would inherit the Attribute twelve, which makes no sense. In this example,
the number twelve reflects the number of Objects in the Set, not an
282
MATH WITHOUT NUMBERS
Attribute of such Objects.
12.13. Can a Number Also Be an Attribute?
In the previous Section 12.12, we saw that we can interpret the Natural
Numbers as the number of Atoms in each Set at a Level of a Power Set,
where the number increases by 1 as we move up the Levels in a Power Set.
Can a number also be an Attribute? For example, as illustrated in Table
12-16 for a P5 Power Set with five Atoms, we can envision a Chain of Sets as
follows: sets with at least four Objects, sets with at least three Objects,
sets with at least two Objects,
There are several important points to note about Table 12-16.
Constructed in this way, each number inherits the Attributes of the number
before it. For example, the number 4 inherits the following Attributes:
283
WILLIAM S. VEATCH
[4] = not sets with three Objects,
[3] = not sets with two Objects,
[2] = not sets with one Object, and
[1] = not sets with zero Objects.
Also, note that the sets beginning with 0=, 1={}, and proceeding
thereafter with the set of what comes before are actually sets of Attributes,
and not sets of Objects. Therefore, the Natural Numbers as we normally
think of them are actually Attributes rather than Objects. This makes sense,
as the Attribute is the Partition, and the Objects are the examples of sets that
have a certain number of Objects in them.
For those familiar with classical Set Theory, Column D in Table 12-
16 will be recognized as the classic way of generating the Natural Numbers
by starting with the Empty Set, and then using the Successor Function
( () ) first articulated by John von Neumann. See [Stewart and Tall 1977]
at 160-162. We create a sequence as follows: Starting with the Empty Set
, we assume that the Successor Set containing the Empty Set exists
{}; then we create the Successor of that Set consisting of the Set of the
previous Sets {{,{ }}; etc. We obtain the sequence set forth in Table
12-17.
Table 12-17 illustrates the classical Set Theory approach to generating
the Set of Natural Numbers from the Empty Set and the Successor Function
that is assumed to exist by the Axiom of Infinity. What the far right column
demonstrates is that we can also use the Axiom of Infinity to generate the
Set of Coatoms in our Universe, that we use to generate the Power Set. If
we treat the Natural Numbers as Attribute Coatoms, then we can establish
Inheritance as we move down the Chain. In other words, at the top of the
Chain are Sets that have any number of Atoms ([0]=no required minimum);
then, at least one Atom [1]; then, at least two Atoms; etc.
Once we have an interpretation for the Addition of Natural Numbers,
284
MATH WITHOUT NUMBERS
we can refer to the standard approaches in Set Theory of expanding to other
number systems, such as the Integers, and the Rational, Real, and Complex
Numbers. Since this is a book on Math Without Numbers, we do not want
to spend too much time discussing numbers! We simply discuss numbers so
as to distinguish them from other Ideas, and to show that Quantity (as used
to determine the number of Atoms at a Level in the Power Set) is not simply
an Attribute like Color, or Shape.
12.14. Update to the Definition of Idea
An Idea as a Set, Power Set, and Power Set of Power Sets: An Idea is
represented by a Set of Atoms (Coatoms) in the Domain of Atoms
(Coatoms) (Dimension 1); a Power Set in the Power Set Domain (Dimension
2); and a Power Set of Power Sets in the Power Set of Power Set Domain
(Dimension 3).
Numbers are relevant to the Mathematics of Ideas in two distinct ways.
First, we use numbers to count the number of Elements in a Set or Power
Set, or in each Row or Column of a Power Set. Second, a number can also
be an Attribute of an Object.
12.15. References, Historical Notes, and Further Reading
[Abbott 1969]: This is one of the most clearly written books that covers a
wide range of topics including the Axioms of Set Theory, Lattices, and
Boolean Algebras. At page 11 is a very clear explanation of how to use the
Axiom of Abstraction to avoid Russells Paradox. If you had to pick one
book to read for more background on the mathematics underlying the
Mathematics of Ideas, this would be a strong candidate.
Other helpful books on Set Theory include the following:
[Halmos 1960],
[Stoll 1963],
[Bourbaki 1968],
[Suppes 1972],
[Enderton 1977],
[Stewart and Tall 1977],
[Levy 1979],
[Tiles 1989],
[Devlin 1993],
[Hrbacek and Jech 1999],
[Rodgers 2000],
285
WILLIAM S. VEATCH
[Lawvere and Rosebrugh 2003],
[Jech 2003],
[Potter 2004],
[Kunen 2011], and
[Pinter 2014].
286
13. CONCLUSION
Now that we have completed Math Without Numbers Volume 1
Foundations, we take a moment to reflect upon our original goals and
objectives, whether we succeeded in meeting those goals and objectives, and
where to go from here.
13.1. Goals and Objectives in Developing a Mathematics of Ideas
Our original goals and objectives included the following:
To build upon the writings of the last 2,000 years, in order to develop
a Mathematics of Ideas based upon Set Theory.
More specifically:
o To analyze Ideas in terms of Sets.
o To develop concepts of addition and multiplication
that are meaningful in the context of Ideas rather than
numbers.
o To see how far we can extend concepts of Group Theory,
Boolean Algebras, Ring Theory, Lattice Theory, and
Topology to Ideas as opposed to numbers.
To explain the historical distinction in Philosophy and Logic
between Extension and Intension, which in MWN terminology
are Objects and Attributes, respectively.
To develop practical tools that everyone can use to analyze Ideas
and improve clarity of thinking and communication, even for those
who do not have a background in higher mathematics.
To apply the tools that we develop to the fields of Logic,
Argumentation, and Debate.
To demonstrate that it is possible to create a Knowledge
Representation Map of the Universe of Ideas.
To answer the question: Why is it important to study the
Mathematics of Ideas?
To inspire others to continue work in this field.
For a number of reasons, now is the perfect time to develop a
287
WILLIAM S. VEATCH
Mathematics of Ideas. For thousands of years, scholars have expended
enormous effort to sort, organize, and classify ideas. Prior to the advent of
computers, the thought of listing out the knowledge of mankind in a
mathematically precise manner was too enormous a task to give serious
consideration. As a result, the focus has been on developing the rules of
Logic, Argumentation, and Debate. Now, with the resources available to us,
we can do both: we can further develop the rules of Logic, Argumentation,
and Debate, but we can also develop a Knowledge Representation Map of
the Universe of Ideas.
Our knowledge and understanding of Mathematics has also increased
dramatically in recent years. With relatively recent advances in fields such as
Lattice Theory and Formal Concept Analysis, we are better equipped to
attempt to bridge the gap between the mathematics of numbers and the
Mathematics of Ideas.
13.2. Did We Succeed in Developing a Mathematics of Ideas?
If the reader has followed along with us from the beginning, she or he is
probably now thinking about the world in a different way. Every Idea is now
either an Atom or a Compound. Related Ideas form Chains, with Ideas lower
in the Chain inheriting Attributes (i.e., properties and characteristics) of Ideas
higher in the Chain. Dictionary definitions are recognized as lists of
Attributes, where each Attribute has the potential to be a new Level in a
Chain. Compound Ideas break down into their component Atoms or
Deemed Atoms, by forming Partitions where the elements of the Partition
are Mutually Exclusive.
We learned that an Idea is a Set, but not just one Set. Rather, an Idea is
made of up two Sets, like two sides of a coin: a Set of Objects and a Set of
Attributes. We also learned that each Idea-Set can be viewed in one
dimension (Atoms and Coatoms), two dimensions (Power Sets), or three
dimensions (Power Sets of Power Sets). We learned about Duality and how
Objects and Attributes (and their related Power Sets) are Inverse Duals of
one another.
We developed a means of adding and multiplying Ideas (based upon
Union and Intersection of Sets), whether articulated as Atoms/Coatoms,
elements of a Power Set, or elements of a Power Set of Power Sets.
We demonstrated that a Power Set constitutes a number of different
mathematical structures, including a Group, Boolean Algebra, Ring, Lattice,
and Topology. Mathematicians have studied each of these structures in detail
288
MATH WITHOUT NUMBERS
over the last 100 years with respect to numbers, but are just beginning to
explore them with respect to Ideas. Now that we have established that an
Idea is represented by a Power Set, all of the knowledge and understanding
of these structures applies equally to both numbers and Ideas.
With each step of constructing the foundations for our Mathematics of
Ideas, we gave practical examples of how to apply the mathematical concepts
to familiar Ideas relating to the Animal-Plant Kingdoms, or the Planets. The
examples, albeit simplistic in subject matter, illustrate techniques that we can
apply to any subject area. The techniques we developed include:
Chains: We use the is a relationship to form Chains,
Antichains/Partitions: We can form Partitions, by dividing a Set
of Objects into Exhaustive and Mutually Exclusive Equivalence
Classes,
Partition Equations: We can create a Partition Equation by
creating a string of Partitions: Partition Equation PE = (Partition A)
In (Partition B) In (Partition C) In (Partition D)
o We can use Cross Tables and Datasets to store large
amounts of data that are sorted by the Partition Equation.
Hierarchy Trees and Outlines: To create Hierarchy Trees or
Outlines, we select the Partitions that we want to use from the
Partition Equation, and the order in which we want to apply them.
Then, we use a Simplified Partition Equation to create a Hierarchy
Tree or Outline.
Throughout the application of each of the above techniques, we
highlighted the application to Ideas expressed both as Objects and as
Attributes. In particular, we emphasized the Inverse Dual Relationship
between Objects and Attributes. In this way, we have explained the historical
concepts of Extension and Intension, in mathematical terms as related
Power Sets of Atoms and Coatoms.
Through our discussion of Universes of Discourse, we demonstrated
that we can analyze a subject independently as a Universe of Discourse
(UD), and then later combine that UD with other UDs. In this way, we
begin to create a Knowledge Representation Map of the Universe of Ideas.
289
WILLIAM S. VEATCH
13.3. Next Steps: Where Do We Go From Here?
Although we have made great progress, there is much left to be done. As in
any academic endeavor, we build upon the work of others before us. In this
case, in the course of our journey we have opened our eyes to a rich universe
of scholarly literature from Aristotle (350 BCE), to Porphyry (260 CE), to
Boole (1854), to Venn (1894), to Russell (1920), to Stone (1936), and finally
to Halmos (1960), to name a few. Without this past work as a foundation,
we would not be where we are today. With this first Volume in the Math
Without Numbers series we have taken a step forward in developing a
Mathematics of Ideas, and, hopefully, have inspired others to continue the
journey.
For those ready to continue the journey, Volume 2 will explore Critical
Thinking and the Logic of Lattices. In particular, we will demonstrate how
we can build a system of Logic based on Set Theory, and that there is a Logic
inherent in a Knowledge Representation Structure.
Volume 3 will explore Argumentation, Debate, and Policy Making.
Specifically, we will address how we use a Value System to make arguments
and decisions about what to think and how to act.
Between the release of Volume 1 and Volume 2 in the Math Without
Numbers series, we will release a software application that is designed to
supplement the books. The application will operate much like a dictionary,
except that it will define Ideas in a precise manner in terms of Chains,
Partitions, and Lattices. We have seen that the Mathematics of Ideas is much
more than just Logic and the form that Arguments take; it also provides a
framework for mapping the knowledge of the World. Before the advent of
computers, the practical task of Mapping out a Universe of a billion or more
Ideas was beyond comprehension, but now, with the aid of computers, it
does not seem so far-fetched. The MWN Dictionary will be a first step
towards developing such a Knowledge Representation Map of the Universe
of Ideas.
After reading this Book, the benefits of applying Set Theory to analyze
the World of Ideas should be self-evident and clear. In particular, in the field
of education we can use the tools we have developed to outline subjects and
teach new concepts to students in a clear and precise manner. Admittedly,
the mathematics in Math Without Numbers Volume 1 is too complex for
some readers, but it is necessary to establish a firm foundation. Math
Without Numbers Volumes 2 and 3 will be more interesting to those who
290
MATH WITHOUT NUMBERS
want to apply the techniques of Math Without Numbers, without necessarily
understanding the intricacies of the underlying mathematics.
Why study the Mathematics of Ideas? For some, the answer is that it is
inherently interesting, and it serves our need to classify and understand the
knowledge of the World in which we live. For others, it is simply a practical
tool that we can use to communicate more clearly, and to answer the
question: why cant we have the same level of certainly and precision with
Ideas that we have with numbers? MWN opens up every area of study to
new analysis using MWN techniques. Hopefully, the reader has experienced
some of the excitement experienced by the author in attempting to link the
World of Mathematics to the World of Ideas.
The real test as to whether or not we have succeeded in our goals, will
be if others are inspired to continue and expand upon the work started here.
Hopefully, the excitement will be contagious, and rekindle interest in
Classical Logic, as well as spark a new interest in the Mathematics of Ideas.
291
WILLIAM S. VEATCH
APPENDIX A
DEFINITION OF IDEA
In this Appendix A, we summarize how the definition of Idea as evolved
over the course of the Chapters.
Chapter
1 Overview An Idea is any thought that can be conceived of in the mind.
2 Objects and Attributes: Each Idea has a dual nature as a Set of
Dual Objects Objects and a Set of Attributes.
and
Attributes
We can fully describe an Idea by a list of examples of the Idea,
called Objects.
We can fully describe an Idea by a list of properties or
characteristics of the Idea, called Attributes.
Objects in the Object Set (or List Set) are examples of the Idea (e.g.,
a Dog is a Mammal).
Attributes in the Attribute Set (or Test Set) are characteristics or
properties exhibited by each Object in the Object-Set (e.g., an
Animal can move and therefore is Mobile).
We can view an Idea as a single Attribute, which forms a Partition
of the Universe of Ideas into two Equivalence Classes.
Inverse Relationship: Objects and Attributes are related to one
another in an inverse relationship, in that the more Attributes an
Idea has, the fewer the number of Objects in the Universe there are
that have all of the required Attributes. (Note that each Object or
Attribute is itself an Idea that has its own Object-Set and Attribute
Set.)
Atom or Compound: An Idea is either an Atom (or Deemed
Atom) or a Compound made up of a Set of Atoms. Dually, an Idea
is either a Coatom (or Deemed Coatom) or a Compound made up
of a Set of Coatoms.
Power Set: The Power Set of Atoms represents all Atoms and all
possible Compound Ideas that we can form from the Set of Atoms.
Dually, the Power Set of Coatoms represents all Coatoms and all
possible Compound Ideas that we can form from the Set of
Coatoms.
292
MATH WITHOUT NUMBERS
3 U and : We assume the existence of both the Universal Set of all
Ideas as Ideas, and the Empty Set containing no Ideas.
Sets
Finite v. Infinite; Discrete v. Continuous: Ideas appear to be
Infinite and Continuous by their nature, but are easier to work with
in mathematical terms if we treat them as Finite and Discrete.
Subdivision and Consolidation: In theory, Ideas are capable of
Subdivision into more and more Atoms ad infinitum, as well as
Consolidation into any Element of the Power Set generated by such
Atoms.
Domains of X, P(X), and P(P(X)): Given a Set of Idea Atoms,
there are three Isomorphic Domains: X, the Power Set of X: P(X),
and the Power Set of Power Sets: P(P(X)).
Simplifying Assumption: In practice, we make a simplifying
assumption that the Universe of Ideas has a Finite number of Atoms
which we refer to as Deemed Atoms. If necessary, we can always
revisit our choice of Atoms and create a Finer Universe of Discourse
with more Atoms, or a Coarser Universe of Discourse with fewer
Atoms.
4 Five Set Relationships Among Objects: Any two Sets of Objects
Object Sets: within a specified Domain are related to each other in one of five
5 Set
Relationships ways:
and 5 Set Identical,
Operations
Disjoint,
Subset-Superset,
Superset-Subset, or
Partially Overlapping.
Within the Object Domain, we can create Partitions of the Atoms.
Five Principal Operations Between Objects: Within any
specified Object Power Set Domain, the five principal Operations
of Union, Intersection, Symmetric Difference, Set Subtraction, and
Complement apply.
293
WILLIAM S. VEATCH
5 Five Set Relationships Among Attributes: We can represent any
Attribute Attribute Idea as a Set of Coatoms in any one of several different
Sets:
5 Set Domains:
Relationships
and 5 Set The Domain of Attributes/Coatoms; or
Operations
The Domain of the Power Set of Attributes/Coatoms.
Any two Sets within a specified Attribute Domain are related to each
other in one of five ways:
Identical,
Disjoint,
Subset-Superset,
Superset-Subset, or
Partially Overlapping.
Within the Attribute Domain, we can create Partitions of the
Coatoms.
Five Principal Operations Between Attributes: Within any
specified Attribute Power Set Domain, the five principal Operations
of Union, Intersection, Symmetric Difference, Set Subtraction, and
Complement apply.
6 Ideas as Sets of Objects/Atoms: In the Domain of Atoms, an
Similarities Idea is represented by a list of Atoms, i.e., a single Atom if the Idea
and
Differences is an Atom or a List Set of Atoms if the Idea is a Compound Idea.
Between
Objects and Ideas as Power Sets of Objects/Atoms: In the Power Set
Attributes Domain of Objects/Atoms, an Idea is made up of the elements of
the Power Set of Atoms in the List Set referred to above.
Ideas as Sets or Power Sets of Attributes/Coatoms: By the
Principle of Duality, there are dual statements about Coatoms:
In the Domain of Attributes/Coatoms, an Idea is represented
by a list of Coatoms, i.e., a single Coatom or a Test Set of
Coatoms if the Idea is a Compound Idea, as to Coatoms.
In the Power Set Domain of Attributes/Coatoms, an Idea is
made up of the elements of the Power Set of Coatoms in the
Test Set referred to above.
294
MATH WITHOUT NUMBERS
7 Power Set of Power Sets of Atoms: There is a third Domain called
Power the Power Set of Power Sets Domain, in addition to the Domain of
Sets in
More Atoms and the Power Set of Atoms Domain. In the Power Set of
Detail Power Sets Domain, an Idea is made up of the elements of the Power
Set of Power Sets formed from the Base Set of Atoms, i.e., each
Element of the Domain is a Power Set.
We can use the Power Set Operation to transform a Set of Idea Atoms
to a Power Set, or a Power Set of Power Sets.
Therefore, we can represent an Idea by a Set, a Power Set, or a Power
Set of Power Sets of Atoms (or Coatoms).
Addition and Multiplication of Ideas: We can Add and Multiply
Ideas in any of the three Domains, where Addition is based on
Union (or Symmetric Difference in a Boolean Ring) and
Multiplication is based on Intersection. We can use either the
Atomic Approach or the DPE Approach to add or multiply Power
Sets, but we get the same result either way.
8 Chains: We can use each Attribute in the Test Set of an Idea to create
Chains a Level in a Chain.
Level of Detail: We can create Chains of varying Levels of Detail
from Coarse to Fine, by selecting fewer or more Attributes to use as
Levels in the Chain. If an Idea has m Attributes, then the applicable
Chain can have from 1 to m Levels.
The Levels of a Chain, expressed as Object Labels, form an
Exhaustive, Pairwise Disjoint Partition of the Attributes of an Idea.
9 Partitions: We can use each Object in the List Set of an Idea to create
Antichains an Equivalence Class in an Antichain / Partition.
or
Partitions
Level of Detail: We can create Antichains of varying Levels of Detail
from Coarse to Fine, by selecting fewer or more Objects to use as
Equivalence Classes in the Antichain. If an Idea has m Objects,
then the applicable Antichain can have from 1 to m Equivalence
Classes.
The Equivalence Classes of an Antichain, expressed as Attribute
Labels, form an Exhaustive, Pairwise Disjoint Partition of the Objects
of an Idea.
295
WILLIAM S. VEATCH
10 Partition Equations: Having defined Closed Binary Operations of
Partition Addition (based upon Union) and Multiplication (based upon
Equations
Intersection) in a Boolean Algebra, we can create Partition Equations,
as well as Nested Partition Equations that combine Partitions and
Chains in a way that reflects the relationships among Ideas in the
World of Ideas.
An Idea (Atomic or Compound) is defined by a Universe of Discourse
together with a Partition Equation that identifies how the Idea is
related to each other element of the Universe of Discourse.
Outlines, Hierarchy Trees, and Nested Lattice Diagrams: We
can represent Nested Partition Equations either as Outlines or
Hierarchy Trees, as wells as Nested Lattice Diagrams.
If there are n Partition Sets of a UD, then there are n! (read n
factorial) possible Partition Equations where order of the Partition
Sets makes a difference. Therefore, there are n! possible Hierarchy
Trees or Outlines.
11 Universe of Discourse: A Universe of Discourse is equivalent to a
Universe Set of Atoms and Coatoms, and their related Power Sets.
of
Discourse
We can represent an Idea by a Universe of Discourse or an Element
of a Universe of Discourse.
Relationships Between Universes of Discourse: Any two
Universes of Discourse are related in the same ways as any two Power
Sets, i.e., the related Base Sets of Atoms are related in one of the five
ways that any two Sets may be related: Identical, Disjoint, Subset-
Superset, Superset-Subset, and Partially-Overlapping.
12 An Idea as a Set, Power Set, and Power Set of Power Sets: An
Set Idea is represented by a Set of Atoms (Coatoms) in the Domain of
Theory
Revisited Atoms (Coatoms) (Dimension 1); a Power Set in the Power Set
Domain (Dimension 2); and a Power Set of Power Sets in the Power
Set of Power Set Domain (Dimension 3).
Numbers are relevant to the Mathematics of Ideas in two distinct
ways. First, we use numbers to count the number of Elements in a
Set or Power Set, or in each Row or Column of a Power Set. Second,
a number can also be an Attribute of an Object.
Appendix Ideas as Groups, Algebras, Rings, Lattices, and Topologies: An
B Idea is represented by Dual Power Sets of Atoms and Coatoms, each
Abstract
Algebra with Set Operations of Addition and Multiplication, based upon
Union (or Symmetric Difference in the case of a Ring) and
296
MATH WITHOUT NUMBERS
Intersection, respectively. Each Power Set constitutes:
An Additive Group,
A Multiplicative Subgroup,
A Boolean Algebra,
A Boolean Ring,
A Boolean Lattice, and
A Boolean Topology.
297
APPENDIX B
HOW THE MATHEMATICS OF IDEAS RELATES TO
THE STUDY OF ABSTRACT ALGEBRA AND
OTHER FIELDS OF MATHEMATICS
In this Appendix B, we look at how the Mathematics of Ideas fits in with
other more traditional areas of Mathematics. We have already seen how we
can interpret Ideas as Atoms or Sets of Atoms. From there we saw how by
creating a Power Set consisting of all of Subsets that we can form from a
Base Set of Atoms, we could create a Boolean Algebra or Boolean Ring. Now
we will provide a brief summary of those and other fields of Mathematics
that overlap with our Mathematics of Ideas.
In summary, we will demonstrate how the following areas of Mathematics
establish the foundation for the Mathematics of Ideas:
Equivalence Classes and Partitions: this is fundamental relationship
of Objects (i.e., Equivalence Classes) to Attributes (i.e., Partitions),
Group Theory: Group Theory explains the Symmetry of Power Sets,
Lattice Theory: three dimensional lattices are superior to Venn and
Euler Diagrams, as an aid to visualizing set relationships among four
or more Objects,
Boolean Algebra Theory: Boolean Algebra Theory explains how the
Operations of Union and Intersection constitute an Algebra,
Ring Theory: Ring Theory uses the Symmetric Difference and
Intersection Operations on Sets to create a Ring structure similar to
Addition and Multiplication of integers,
Topology: Topology allows us to Consolidate and Subdivide
structures with varying Levels of Detail, and
Formal Concept Analysis: Formal Concept Analysis brings us to the
latest thinking in terms of applying Lattice Theory to the analysis of
concepts and data.
What all of these areas of mathematical study have in common, is that the
Power Set plays a special role.
298
MATH WITHOUT NUMBERS
The importance of this exercise is that areas of Mathematics such as
Group Theory, Lattice Theory, Boolean Algebras, Ring Theory, and
Topology have been studied for many decades, and as a result there is a
wealth of general knowledge and specific theorems available to us. Once we
demonstrate that a structure of Ideas is a Group, Lattice, Algebra, Ring, or
Topology, then all of this knowledge built up over many decades is applicable
to our structure of Ideas. In addition, relatively new and exciting areas of
Mathematics such as Formal Concept Analysis (FCA), are at the cutting
edge of new thinking in Mathematics. We will see that MWN and FCA, while
not the same, are related, complementary fields of study.
Our goal in this Chapter is not to provide a lot of detail about any of these
areas of Mathematics, but rather to show how these different areas of
Mathematics relate to the Mathematics of Ideas. Once we establish that a
structure of Ideas forms a Group or Ring etc., we leave it to the interested
reader to seek out more information in the references listed at the end of this
Chapter.
1. Equivalence Relations and Partitions; Laws of Composition
As we have noted previously, it is extremely important when working in the
Mathematics of Ideas, to keep track of the Domain in which we are working.
In the Domain of Atoms (Dimension 1), the Elements of the Domain are all
Atoms. As a result, there are no closed Binary Operations based upon Union
and Intersection, since there are no Sets in the Domain. We can, however,
use Equivalence Relations to create Partitions:
One way to view a Relation R is as a Mapping from the Domain to
itself: X X. Another representation is as an Ordered Pair: (x, x) R,
where (x, x) is a Subset of the Direct Product: X X. We typically write
aRb, where a, b X, to mean that a and b are related by R.
In order to constitute an Equivalence Relation, the Relation must have
the following properties:
Reflexive: aRa;
Symmetric: If aRb, then bRa; and
Transitive: If aRb and bRc, then aRc.
See Fig. B-1.
299
WILLIAM S. VEATCH
Note that the Mappings in an Equivalence Relation are not unique. For
example, in Fig. B-1, both aRb and aRc are True statements. We sometimes
use the notation [a] to denote the Equivalence Class containing the
Element a. In the example given in Fig. B-1, [a]=[b]=[c].
Although we do not repeat the proof here, it is well-established that an
Equivalence Relation on a Set creates a Partition of the Set into Equivalence
Classes; similarly, a Partition creates an Equivalence Relation. See [Pinter
1990] at 119-125.
In the Power Set Domain (Dimension 2), we define Binary Operations
or Laws of Composition, including Union, Symmetric Difference, and
Intersection. One way to view such a Binary Operation/Law of
Composition is as a Mapping from two Elements in the Domain to a third
Element in the same Domain: P(X) x P(X) P(X). Note that for the Binary
300
MATH WITHOUT NUMBERS
Operation to be Closed, the result of the Operation must appear in the
same Domain. Union, Symmetric Difference, and Intersection are all Closed
Operations in the Power Set Domain.
Another possible representation is as an Ordered Triple: (x, x, x)
Operation, where (x, x, x) is a Subset of P(X) P(X) P(X). For example,
we typically write a b=ab, where a, b, ab P(X), to mean that the Union
of a and b is ab.
In the Power Set of Power Sets Domain (Dimension 3), as in the Power
Set Domain, we define Binary Operations or Laws of Composition, including
Union, Symmetric Difference, and Intersection. One way to view such a
Binary Operation/Law of Composition is as a Mapping from two Elements
in the Domain to a third Element in the same Domain: P(P(X)) x P(P(X))
P(P(X)). Note that for the Binary Operation to be Closed, the result of
the Operation must appear in the same Domain. Union, Symmetric
Difference, and Intersection are all Closed Operation in the Power Set of
Power Sets Domain.
Another possible representation is as an Ordered Triple: (x, x, x)
Operation, where (x, x, x) is a Subset of P(P(X)) P(P(X)) P(P(X)). For
example, we typically write (, a ) ( , b )=( , a, b, ab ), where (, a),
( , b ), ( , a, b, ab ) P(P(X)), to mean that the Sum (i.e., the Power Set
of the Union of Atoms) of (, a ) and ( , b ) is ( , a, b, ab ).
1.1. Notation
Before we move on to look in more detail at the various Binary Operations
in our three primary Domains, for convenience we repeat the table of
symbols (i.e., Tables 7-4 and 7-5) that we use in MWN for the various Binary
Operations. Table B-1 illustrates the various symbols used in our books on
MWN for the Addition and Multiplication of Elements of Boolean Algebras
and Boolean Rings for generic Algebra and Ring structures, the Power Set
Domain, and the Power Set of Power Sets Domain.
301
WILLIAM S. VEATCH
As noted earlier in Subchapter 7.5, we have developed a simple
notation that allows us to show that we are taking the Direct Product and
following up with the application of a Binary Operation to each Ordered Pair
(or N-Tuple) resulting from the Direct Product. Table B-2 illustrates this
concept.
Note in Table B-2 that we use a colon, :, before the parentheses to
indicate Intersection rather than Union or Symmetric Difference; and we
use :: to indicate Symmetric Difference. Although we have covered
Operations in the Power Set Domain, for the moment we have not specified
combined Direct Product-Operation notation for the Power Set of Power
Sets Domain. We can do this later if it becomes necessary. Usually it is not
an issue, because we convert to Sets of Atoms, apply Operations, and then
302
MATH WITHOUT NUMBERS
convert back to Power Sets.
There is one exception to Table B-2. For clarity and simplicity of
notation in our use of Partition Equations (see Chapter 10), we use +
instead of , along with In to indicate Direct Product followed by
Intersection ( ). We do this because the + symbol is easier to read.
2. Group Theory Symmetry
2.1. Characteristics of a Group
When we hear Group Theory, we think symmetry. We can use Group
Theory in Mathematics to describe symmetrical structures in nature like
snowflakes or crystals. The Power Set Lattice that we use to represent Ideas
also has a symmetrical structure, which explains at an intuitive level why a
Power Set together with a single Binary Operation such as Addition or
Multiplication, constitutes a Group (or Semigroup in the case of
Intersection as Multiplication).
2.2. Definition of a Group
A Group has one Binary Operation, an Identity Element, and each Element
in the Domain has an Inverse. The generic definition of Group is as
follows:
Definition: Group: (G, ) = G; , e, -a=a , where:
Domain: A Set G. In MWN, our Domain is either the Power Set of
X, P(X), where X is a Base Set of Atoms; or the Power Set of Power
Sets P(P(X)).
Binary Operation: : There must be a Binary Operation that is Closed
over G, and that is associative. In the case of a Commutative Group
(also known as an Abelian Group), the Operation is also commutative.
The Operation may be Additive or Multiplicative, and a single Domain
may have more than one Operation associated with it.
Identity: There is an Element e such that a e=a, and e a=a
for every Element a in G. We write e as 0 in an Additive Group,
and as 1 in a Multiplicative Group.
Inverse: There is an element -a in an Additive Group, or a-1 in a
Multiplicative Group, such that a+a=0 or aa=1.
303
WILLIAM S. VEATCH
A Group G is a Mapping : G G G
In the context of a Power Set, there is one Group based upon Addition, and
one Semigroup based upon Multiplication.
Definition: Additive Power Set Group: (P(X), ) =
P(X); , , -a=a , where:
Domain: P(X), where Base Set X Atoms = {a,b,c,d,}.
Binary Operation: is based upon Symmetric Difference (which is
equivalent to Union minus Intersection) of the Elements in the
Domain, and is associative and commutative (Abelian). See Chapter 4.
Identity: 0=:
Inverse: -a=a: Each element is its own Inverse (Idempotent); a (-a)=
a a=0.
Definition: Multiplicative Power Set Semigroup: G = P(X); , 1 ,
where:
Domain: P(X), where Base Set X Atoms = {a,b,c,d,}.
Binary Operation: is Intersection, and is Associative and
Commutative (Abelian). See Chapter 4.
Identity: 1=X.
Inverse?: Elements of P(X), other than the Element X, do not have
an Inverse, so is a Semigroup with Identity, but not a Group. For
example, in the P3 Power Set with Elements = ( , a, b, c, ab, ac, bc,
abc), the Intersection of any Power Set Element with any other Element
can never be larger than the smaller of the two Elements. Therefore,
only X can have an Inverse, i.e., abc abc = abc=1.
As with the Power Set Domain, in the context of a Power Set of Power
Sets, P(P(X)), there is one Group based upon Addition, and one Semigroup
based upon Multiplication.
Definition: Additive P(P(X)) Group: (P(P(X)), ) = P(P(X)); , ,
-a=a , where:
Domain: P(P(X)), where Base Set X Atoms = {a,b,c,d,}.
Binary Operation: is based upon the Symmetric Difference of the
related Base Set Elements (which is equivalent to Union minus
Intersection), and is associative and commutative (Abelian). See
Chapter 4.
Identity: 0=:
304
MATH WITHOUT NUMBERS
Inverse: -a=a: Each element is its own Inverse (Idempotent); a (-a)=
a a=0
Definition: Multiplicative P(P(X)) Semigroup: G = P(P(X)); , 1 ,
where:
Domain: P(P(X)), where Base Set X Atoms = {a,b,c,d,}.
Binary Operation: is based upon Intersection of the underlying
Base Set Elements, and is associative and commutative (Abelian). See
Chapter 4.
Identity: 1=X.
Inverse?: Elements of P(P(X)), other than the Element P(X), do not
have an Inverse, so is a Semigroup with Identity, but not a Group.
The Groups that we can form that are relevant to MWN include the
following:
Power Set of Atoms / Coatoms
o Additive Group
o Multiplicative Semigroup
Power Set of Ideals / Filters
o Additive Group
o Multiplicative Semigroup
Logic Operations (to be discussed in MWN Volume 2)
o Additive
o Multiplicative
3. Boolean Algebra
When working in the Mathematics of Ideas, we often work with Boolean
Algebras instead of Boolean Rings. The reason for this, at least in part, is that
Boolean Algebras are easier to understand intuitively, because Addition is
Union. Addition for Boolean Rings is Symmetric Difference, which is a
brilliant way to create an Inverse for every Element in the Domain (i.e., by
making every Element its own Inverse), but it creates some odd results. We
normally do not think that adding an Idea to itself should result in . (It
can also lead to confusion over the meaning of subtraction as either adding
an Inverse versus Set Subtraction.) In any event, we can easily convert a
Boolean Algebra to a Boolean Ring if we need to (e.g., to take advantage of
the properties of Rings), so there is no harm in working with Boolean
Algebras.
305
WILLIAM S. VEATCH
3.1. Characteristics of a Boolean Algebra
A Boolean Algebra has two Operations, Addition and Multiplication. The
example with which most people are familiar, at least to some degree, is a Set
of Objects together with the Operations of Union, as Addition, and
Intersection, as Multiplication. Another important example of a Boolean
Algebra is found in the field of Logic, where the and (conjunction) and
or (disjunction) Binary Operations, together with the negation Unary
Operation, form a Boolean Algebra. See, [Givant and Halmos 1998].
3.2. Definition of a Boolean Algebra
Definition: Generic Boolean Algebra based on Sets:
A Boolean Algebra is defined as U; , , 0, 1, _ where:
Domain: U is a Base Set, with Elements A, B, C, ,
Addition: is defined as Set Union,
Multiplication: is defined as Set Intersection,
Additive Identity: 0 is the Empty Set = ,
Multiplicative Identity: 1 is the Universal Set = U, and
Complement: _ is the Complement Unary Operation.
See, [Abbott 1969] at 183.
A Boolean Algebra U is a Mapping : G G G
Definition: Power Set Boolean Algebra:
The Boolean Algebra of all Subsets of X is defined as
P(X); , , 0, 1, _ where:
Domain: P(X), where Base Set X Atoms = {a,b,c,d,}.
Addition: is Union, and is associative and commutative (Abelian).
See Chapter 4.
Additive Identity: 0 is the Empty Set = :
Multiplication: is Intersection, and is associative and
commutative (Abelian). See Chapter 4.
Multiplicative Identity: 1 is the Universal Set = X.
Complement: _ is the Complement Unary Operation.
Distributive: Multiplication is Distributive over Addition.
306
MATH WITHOUT NUMBERS
Definition: P(P(X)) Boolean Algebra: P(P(X)); , , 0, 1, _
where:
Domain: P(P(X)), where Base Set X Atoms = {a,b,c,d,}.
Addition: is based upon the Union of the related Base Set
Elements, and is associative and commutative (Abelian). See Chapter 4.
Additive Identity: 0=:
Multiplication: is based upon Intersection of the underlying Base
Set Elements, and is associative and commutative (Abelian). See Chapter
4.
Identity: 1=X.
Distributive: Multiplication is Distributive over Addition.
What makes the Boolean Algebra particularly usesful is that we do not need
to have an Additive Inverse. As mentioned earlier, however, we can convert
any Boolean Algebra to a Boolean Ring if we need to, by using Symmetric
Difference (instead of Union) for Addition.
4. Ring Theory
We will look at three types of Rings: a Generic Ring, a Ring in the Power
Set Domain, and a Ring in the Power Set of Power Sets Domain.
4.1. Characteristics of a Ring
When we think of a Ring, we think of a generalized Set that resembles the
Set of Integers: there are two Operations, Addition and Multiplication, an
Additive Identity (0), a Multiplicative Identity (1), and each Element has an
Additive Inverse (i.e., negative numbers in the case of Integers). (The analogy
is not perfect, however, because the Integers also constitute an Integral
Domain, whereas a Boolean Ring does not, so the Integers have some
properties that do not carry through to Boolean Rings. An Integral Domain
is a Commutative Ring with Unity that has no divisors of zero. See [Pinter
1990] at 200. Boolean Rings have divisors of zero, because the Intersection
of any Disjoint elements is zero.) Another key characteristic of a Boolean
Ring is that we can convert any Boolean Ring into a Boolean Algebra, and
vice versa.
307
WILLIAM S. VEATCH
4.2. Definition of a Ring
The generic definition of Ring with Identity (which we shall refer to as a
Ring) is as follows:
Definition: Generic Ring: A Ring is an Ordered Quadruple
R; +, , e consisting of a nonempty Set R, two Binary Operations
Closed on R, and one Nullary Operation such that:
Domain: R. In MWN, our Domain is either the Power Set of X,
P(X); or the Power Set of Power Sets, P(P(X)).
Addition: + is a Commutative Group. There must be a Binary
Operation that is Closed over G, and that is associative. In the case
of a Commutative Group (also known as an Abelian Group), the
Operation is also commutative.
Multiplication: is a Semigroup.
Distributive: is Distributive (on both sides) over the Operation
+.
Multiplicative Identity: e = 1. There is an Element e such that
a e=a, and e a=a for every Element a in G. We write e
as 1 in a Multiplicative Group.
See [Burton 1970] at 1.
A Ring R is a Mapping : G G G
In the context of a Power Set, there is a Ring comprised of one Group based
upon Addition, and one Semigroup based upon Multiplication.
Definition: Power Set Ring is defined as: P(X); , , e , where:
Domain: P(X), where Base Set X Atoms = {a,b,c,d,}.
Addition: A B=(A B ) ( A B)=(A B)-(A B). is
based upon Symmetric Difference (which is equivalent to Union minus
Intersection) of the Elements of the Domain, and is associative and
commutative (Abelian). See Chapter 4.
Multiplication: A B = A B. is Intersection, and is associative
and commutative (abelian). See Chapter 4.
Multiplicative Identity: e=1=X.
Multiplicative Inverse?: Elements of P(X), other than the Element
X, do not have a Multiplicative Inverse, so is a Semigroup with
Identity, but not a Group.
308
MATH WITHOUT NUMBERS
Distributive: Multiplication is Distributive over Addition.
See [Abbott 1969] at 178 and 185. Cf. [Burton 1970] at 204 et seq.,
discussing direct sums of Rings.
We can also define a Ring with Identity as an Additive Abelian Group,
together with a Commutative Multiplicative Semigroup on the same Domain
Set, such that Multiplication is Distributive over Addition.
As with the Power Set Domain, in the context of a Power Set of Power
Sets, P(P(X)), there is a Ring comprised of one Group based upon Addition,
and one Semigroup based upon Multiplication.
Definition: P(P(X)) Ring: P(P(X)); , , , X , where:
Domain: P(P(X)), where Base Set X Atoms = {a,b,c,d,}.
Addition: is based upon the Symmetric Difference of the related
Base Set Elements (which is equivalent to Union minus Intersection),
and is associative and commutative (Abelian). See Chapter 4.
Additive Identity: 0=:
Additive Inverse: -a=a: Each element is its own Inverse (Idempotent);
a (-a)= a a=0
Multiplication: is based upon Intersection of the underlying Base
Set Elements, and is associative and commutative (Abelian). See Chapter
4.
Identity: 1=X.
Inverse?: Elements of P(P(X)), other than the Element P(X), do not
have an Inverse, so is a Semigroup with Identity, but not a Group.
Distributive: Multiplication is Distributive over Addition.
Adding and Multiplying Power Sets is where the action is, for those who
love math. The challenge when trying to apply mathematics to Ideas, is to see
how far we can go to make the rules of Algebra apply to abstract Ideas. In
MWN, we look at three different Sets relating to Ideas, along with certain
Operations on those Sets, each of which constitutes an Algebraic structure:
Ideas as Sets,
Propositions (covered in MWN Volume 2), and
Power Sets of Ideas.
The most interesting is the Power Sets of Ideas (sometimes referred to
as Ideals or Filters), as developing rules for adding and multiplying Power
Sets provides us with a practical technique for working with complex Sets of
309
WILLIAM S. VEATCH
Ideas that would otherwise be extremely cumbersome.
In a MWN Volume 2, we will explore using the Equivalence or
Biconditional Operation as Multiplication instead of Intersection, i.e., the
operation: (p q) (~p ~q), which is the Complement or Dual of
Symmetric Difference. The Equivalence Operation can be thought of as the
Intersection of p and q, together with any elements of the Domain not
included in p or q or the elements in both sets or in neither. See [Abbott
1969] at 50. Each element becomes its own Multiplicative Inverse.
5. Lattice Theory
Throughout this book, we have made extensive use of Lattice diagrams, often
referred to as Hasse Diagrams. Lattice diagrams help us to visualize Chains,
AntiChains (or Partitions), and Power Sets. Lattice Theory is a vast topic
covering a wide variety of Lattice types. Our focus, however, will be limited
to Boolean Lattices, as well as Chains and Antichains, since we have
demonstrated earlier (see Subchapter 7.12) that a Boolean Lattice can be
viewed as a sum of Chains, or a sum of Antichains.
5.1. Characteristics of a Boolean Lattice
While there are many different types of Lattices, as mentioned above, our
focus is on Boolean Lattices, along with their component parts: Chains and
Antichains (or Partitions). In fact, a Boolean Lattice is equivalent to a
Boolean Algebra, a Boolean Ring, and a Discrete Topology.
5.2. Definition of a Boolean Lattice
A Lattice L is a Poset X, where every pair of Elements of X has
a Supremum and an Infimum. See [Bergman 2012] at 22. We will not
go into any detail in this Book to explain these terms, but many excellent
reference books on Lattices are listed at the end of this Chapter. Our goal is
simply to point out that the Power Set structures that we have been
examining constitute Lattices, and then provide the interested reader with
further resources.
A Lattice is Bounded if both 0L and 1L exist. See [Bergman 2012] at 22.
Elements x and y in L are Complements if x y =0 and x y =1. See
[Bergman 2012] at 29.
310
MATH WITHOUT NUMBERS
L is Complemented if every Element has a Complement. See [Bergman
2012] at 29.
L is a Boolean Lattice if it is Bounded, Distributive and Complemented.
See [Bergman 2012] at 29.
A Boolean Algebra is an Algebra X; , , _ , 0, 1 where
X; , is a Distributive Lattice with bounds 0 and 1 and for every x
X, x is a Complement of x . See [Bergman 2012] at 29.
L is a Complete Lattice if for every Subset X of L, both Supremum X
and Infimum X exist. See [Bergman 2012] at 30.
6. Topology Manipulating Power Sets in 3D
Books on the introduction to Topology often provide images of teapots and
doughnuts, or Mobius strips. In the context of the Mathematics of Ideas,
however, Topology helps us to visualize Power Sets in three-dimensional
space, and in particular the Consolidation of Sets into Coarser Topologies;
or the Subdivision of Sets into Finer Topologies.
6.1. Characteristics of a Topology
A Boolean Topology is a Subset of the Power Set P(X), such that the chosen
Elements, referred to as Open Sets are closed under Union and finite
Intersection. For our purposes in the Mathematics of Ideas, we focus
primarily on the Topology formed by the entire Power Set, referred to as the
Discrete Topology.
6.2. Definition of a Topology
Definition: Topology: A Topology on a non-empty Base Set X is a
collection or class T of subsets of X, called the Open Sets, such that:
Any Union of any number of Elements of T belongs to T, i.e., the
Union of any number of Open Sets is an Open Set;
Any finite Intersection of Elements of T belong to T, i.e., the
Intersection of any finite number of Open Sets is an Open Set.
and X belong to T, i.e., each of and X is an Open Set.
We say that (X, T) is a Topological Space. See [Willard 1970] at 23.
311
WILLIAM S. VEATCH
See also, [Lipschutz 2012] at 66; [Dugundji 1966] at 62-63; and [Eves 1990]
at 234. A Boolean Space is a Compact, Totally Disconnected, Hausdorff
space. See, [Abbott 1969] at 204 and 268.
If an Open Set G contains a point p X, then G is called an Open
Neighborhood of p. G\{p} is called a Deleted Open Neighborhood of p.
The following is a list of all Topologies on a Base Set with two Elements,
X = { a, b }:
{ , X } = { , ab }
{ , a, X } = { , a, ab }
{ , b, X } = { , b, ab }
{ , a, b, X } = { , a, b, ab }
Definition: Discrete Topology: Power Set of X = {a,b,c,d}
Let X = { abcd }, and T = {, a, b, c, d, ab, ac, ad, bc, bd, cd, abc, bcd, cda,
dab, X=abcd }. Then, T is a Topology on X. |T| = 16.
It is the largest, or Finest, Topology on X.
( X, T ) is a Discrete Topological Space.
Definition: Indiscrete Topology
C = {X, } is the Indiscrete Topology, and ( X, C ) is an Indiscrete
Topological Space or an Indiscrete Space.
It is the smallest, or Coarsest, Topology on X. See, [Willard 1970] at 24 and
[Lipschutz 2012] at 66.
While we just touch on Topology in this Book, we will explore it in more
detail in later volumes. The significance of Topology to our Mathematics of
Ideas is that Topology provides the mathematical foundation for expanding
and contracting Power Set Lattices to represent various Levels of Detail,
from Coarse to Fine.
7. Formal Concept Analysis Analysis of Data
One of the newest and most exciting areas of research in mathematics is
Formal Concept Analysis (FCA). FCA began with the work of Bernhard
312
MATH WITHOUT NUMBERS
Ganter and Rudolf Wille, in the seminal book [Ganter and Wille 1999]:
Formal Concept Analysis Mathematical Foundations. FCA is much like MWN,
but starting from a different perspective. FCA focuses on analyzing data
through the use of Complete Lattices, where MWN focuses more on Power
Sets and the philosophical foundations of the Mathematics of Ideas.
Mathematicians and researchers in the field of FCA have made great
strides over of the last few years in analyzing data in a wide variety of
contexts. There remain, however, vast untapped areas of research and
application of FCA principles by combining the techniques of FCA with the
techniques of MWN outlined in this book.
Formal Concept Analysis (FCA) is a means of selecting a subset of the
Power Set that includes only those elements that have significance when
modelling a Set of Ideas in our real world.
7.1. Characteristics of a Formal Concept
In FCA, we proceed from a Cross Table Formal Context Formal
Concept Concept Lattice. Table B-3 shows a sample Cross Table, and
the following Table B-4 compares FCA to our Mathematics of Ideas and
illustrates how they are similar and how they differ.
313
WILLIAM S. VEATCH
314
MATH WITHOUT NUMBERS
Both FCA and MWN analyze Ideas, but from a different perspective. FCA
uses Complete Lattices to analyze data, whereas MWN focuses primarily on
Boolean Lattices. FCA and MWN are complementary, not competing, ways
of analyzing the same subject. In a later Volume in the MWN series, we will
examine the similarities and differences between FCA and MWN in more
detail.
315
WILLIAM S. VEATCH
7.2. Definition of a Formal Concept
Definition: Formal Context K := (G,M,I) where:
G is a Set of Objects.
M is a Set of Attributes.
I is a relation between G and M.
We write gIm or (g,m) I and read it as the object g has the attribute
m. Note that I is a subset of the Direct Product of G and M, i.e., I GxM.
See, [Carpineto and Romano 2004] at 10.
Note that in a Formal Context (as opposed to a Concept), it is possible that
g I m or (g,m) I.
Definition: A :={m M|gIm for all g A}, which is the set of Attributes
common to the Objects in A.
In the terminology of our Mathematics of Ideas, A = the List Set of Objects
and A = the related Test Set of Attributes, including inherited Attributes.
Definition: B :={g G|gIm for all m B}, which is the set of Objects
which have all Attributes in B.
Similarly, in the terminology of our Mathematics of Ideas, B = the Test Set
of Attributes and = B the related List Set of Objects.
Definition: A Formal Concept of the context (G,M,I) is a pair (A,B) with
A G, B M, A = B and B = A. In MWN terminology, A is the
Extent, Set of Objects, or List Set; and B is the Intent, Set of Attributes,
or Test Set. The pair (A,B) represents a single Element in the Dual Power
Set of Atoms and Coatoms.
8. Update to the Definition of Idea
An Idea is represented by Dual Power Sets of Atoms and Coatoms, each with
Set Operations of Addition and Multiplication, based upon Union (or
Symmetric Difference in the case of a Ring) and Intersection, respectively.
Each Power Set, and therefore each Idea, constitutes:
316
MATH WITHOUT NUMBERS
An Additive Group,
A Multiplicative Subgroup,
A Boolean Algebra,
A Boolean Ring,
A Boolean Lattice, and
A Boolean Topology.
9. References, Historical Notes, and Further Reading
Our purpose in this Chapter is not to provide a detailed analysis or
explanation of Groups, Algebras, Rings, etc., but only to show how Ideas
form structures that constitute Groups, Algebras, Rings, etc. Having made
this connection, the interested reader can read further to explore how we can
apply the rules and theorems in these various fields of mathematics, to the
Mathematics of Ideas.
Overview
[Abbott 1969]: This is a great book to start with for an overview of Lattices
and Boolean Algebras.
See also, [Pinter 1990] at 119 et seq. for a discussion of Partitions and
Equivalence Relations.
Power Sets as Groups
For a discussion of Power Sets as Groups under Addition, see [Pinter 1990]
at 30-31.
Group Theory
See [Eves 1990] at 127 for a discussion of the fundamental theorems of
Groups, and at 128 for a discussion of the significance of Groups in Algebra
and Geometry.
See generally:
[Wallace 1998],
[Pinter 1990], and
[Clark 1984].
317
WILLIAM S. VEATCH
Boolean Algebras
For an introductory level explanation of a Boolean Algebra, see [Eves 1990]
at 216 et seq.
See generally:
[Givant and Halmos 2009]
[Abbott 1969]
Ring Theory
See generally:
[Wallace 1998]
[Pinter 1990]
[Clark 1984]
Lattice Theory
See generally:
[Caspard, Leclerc, and Monjardet 2012]
[Roman 2008]
[Schrder 2003]
[Grtzer 2003]
[Davey and Priestley 2002]
[Crawley and Dilworth 1973]
[Birkhoff 1967]
Ideal Theory
See generally:
[Northcott 1968]
Topology
For an introductory level explanation of a Topological Space, see [Eves 1990]
at 230 et seq.
318
MATH WITHOUT NUMBERS
See generally:
[Willard 1970]
[Abbott 1969]
[Dugundji 1966]
Formal Concept Analysis
See generally:
[Ganter, Stumme, and Wille (Eds.) 2005]
[Carpineto and Romano 2004]
[Ganter and Wille 1999]
319
Bibliography
The year listed is the year of the latest revision by the author, but not
necessarily the latest printing if no revisions were made. That way, the reader
can get a sense of the historical perspective of the Book or article. (Note that
the labels in the Category column form a valid Partition of the Book Titles!)
Abbreviation Book Category
[Aarts 2011] Aarts, Bas, Oxford Modern English Knowledge Base:
Grammar, Oxford University Press English Grammar
2011.
[Abbott 1969] Abbott, James C., Sets, Lattices, and Abstract Algebra
Boolean Algebras, Boston, Allyn and
Bacon, Inc. 1969.
[Aristotle 350 Richard McKeon (ed.), The Basic Works Classical Logic.
B.C.E.] of Aristotle, New York, Random House
1941.
[Auslander and Auslander, Maurice and David A. Group Theory, Ring
Buchsbaum 1974] Buchsbaum, Groups, Rings, Modules, Theory
Mineola, New York, Dover 1974.
[Bachhuber 1957] Bachhuber, Andrew H., Introduction to Logic
Logic, New York Appleton-Century-
Crofts, Inc. 1957,
[Baynes 1861] Baynes, Thomas Spencer, The Port- Logic
Royal Logic, London, Hamilton, Adams,
and Co. 1861. Reprinted by Kessinger
Legacy Reprints.
[Beason and Lester Beason, Larry, and Mark Lester, English Knowledge Base:
2013] Grammar and Usage, New York, English Grammar
McGraw Hill 2013.
[Bergman 2012] Bergman, Clifford, Universal Algebra Abstract Algebra
Fundamentals and Selected Topics, New
York, CRC Press, 2012.
[Birkhoff 1995] Birkhoff, Garrett, Lattice Theory, Lattice Theory
Providence, Rhode Island, American
Mathematical Society 1995.
[Black 1952] Black, Max, Critical Thinking, New Logic
Jersey, Prentice-Hall, 1952.
[Black 1968] Black, Max, The Labyrinth Of Language, Philosophy of Language
New York, Encyclopedia Britannica
1968.
[Boole 1854] Boole, George, An Investigation of The Logic
Laws of Thought, On Which are Founded
The Mathematical Theories of Logic and
Probabilities, New York, Dover Reprint
of 1854 Ed.
[Bourbaki 1968] Bourbaki, Nicolas, Elements of Set Theory
Mathematics, Theory of Sets, Menlo Park,
320
MATH WITHOUT NUMBERS
Abbreviation Book Category
California, Addison-Wesley Publishing
Co. 1968.
[Brachman and Brachman, Ronald J. and Hector J. Knowledge
Levesque 2004] Levesque, Knowledge Representation and Representation
Reasoning, San Francisco, California,
Elsevier 2004.
[Burris and Burris, Stanley and H.P. Universal Algebra
Sankappanavar Sankappanavar, A Course in Universal
1981] Algebra, New York, Springer 1981.
[Burton 1970] Burton, David M., A First Course in Ring Theory
Rings and Ideals, Menlo Park, California,
Addison-Wesley Publishing Company,
Inc. 1970.
[CQ14 2014] CQ Researcher, Issues for Debate in Knowledge Base:
American Public Policy Selections Current Issues
From CQ Researcher, 14th Edition
(2014)
[CQ15 2015] CQ Researcher, Issues for Debate in Current Issues
American Public Policy, 15th Ed. (CQ Knowledge Base:
Press 2015)
[Carpineto and Carpineto, Claudio, and Giovanni Formal Concept
Romano 2004] Romano, Concept and Data Analysis Analysis
Theory and Applications, Chichester, West
Sussex, England, John Wiley & Sons
Inc. 2004.
[Case, Funke, and Case, Christine L., Berdell R. Funke, Knowledge Base:
Tortora 2016] and Gerard J. Tortora, Microbiology-An Microbiology
Introduction, Essex, England, Pearson
Education 2016.
[Caspard, Leclerc, Caspard, Nathalie, Leclerc, Bruno, and Lattice Theory
and Monjardet Monjardet, Bernard, Finite Ordered Sets,
2012] New York, Cambridge University Press
2012.
[Chao ed. 1993] Chao, Liu Hai, Shaolin Gong-Fu A Knowledge Base:
Course in Tradition al Forms, Henan Martial Arts
Scientific and Technical Publishing
House 1994.
[Chen and Koh Chen, Chuan-Chong and Koh, Khee- Combinatorics
1992] Meng, Principles and Techniques in
Combinatorics, Singapore, World
Scientific Publishing 1992.
[Chilvers 2013] Chilvers, Ian, Chief Consultant, Art that Knowledge Base: Art
Changed the World, New York, Dorling
Kindersley (DK) 2013.
[CLA 2006] Yahia, Sadok Ben, Engelbert Mephu Lattice Theory
Nguifo, Radim Belohlavek (Eds.),
Fourth International Conference, on Concept
Lattices and Their Applications (CLA),
New York, Springer 2008.
321
WILLIAM S. VEATCH
Abbreviation Book Category
[Clark 1970] Clark, Allan, Elements of Abstract Algebra, Abstract Algebra
New York, Dover 1984.
[Cogeval, Patry, Cogeval, Guy, Patry, Sylvie, and Knowledge Base: Art
and Guegan 2010] Guegan, Stephane, Van Gogh, Gauguin,
Cezanne, and Beyond, New York, Del
Monico Books -Prestel 2010.
[Cogeval, Guegan, Cogeval, Guy, Guegan, Stephane, and Knowledge Base: Art
and Thomine- Thomine-Berrada, Alice, Birth of
Barrada 2010] Impressionism, New York, Del Monico
Books -Prestel 2010.
[Cohn 1999] Cohn, Paul M., An Introduction to Ring Ring Theory
Theory, New York, Springer 2000.
[Coile 2005] Coile, D. Caroline, Encyclopedia of Dog Knowledge Base: Dogs
Breeds, New York, Barrons 2005.
[Cothran Book I Cothran, Martin, Traditional Logic Logic
2000] Introduction to Formal Logic, Memoria
Press 2000.
[Cothran Book II Cothran, Martin, Traditional Logic Logic
2000] Advanced Formal Logic, Memoria Press
2000.
[Cothran 2006] Cothran, Martin, Material Logic, Logic
Memoria Press 2006.
[Crawley and Crawley, P. and Dilworth, R.., Algebraic Lattice Theory
Dilworth 1973] Theory of Lattices, New Jersey, Prentice-
Hall, Inc. 1973.
[Curry 1976] Curry, Haskell B., Foundations of Mathematical Logic
Mathematical Logic, New York, Dover
1977.
[Davey and Davey, B.A., and Priestley, H.A., Lattice Theory
Priestley 2001] Introduction to Lattices and Order (2d Ed.),
Cambridge University Press 2002.
[Denney, Duncan, Denney, Joseph Villiers, Carson S. Argumentation and
and McKinney Duncan, and Frank C. McKinney, Debate
1910] Argumentation and Debate, New York,
American Book Company 1910.
Reprinted by Forgotten Books 2012.
[Devlin 1993] Devlin, Keith, The Joy of Sets, New York, Set Theory
Springer 1993.
[Dugundji 1966] Dugundji, James, Topology, Boston, Topology
Allyn and Bacon, Inc. 1966.
[Durant Vol.1 Durant, Will, The Story of Civilization: 1, Knowledge Base:
1935] Our Oriental Heritage, New York, Simon History
and Schuster 1935.
[Durant Vol.2 Durant, Will, The Story of Civilization: 2, Knowledge Base:
1939] The Life of Greece, New York, Simon and History
Schuster 1939.
[Durant Vol.3 Durant, Will, The Story of Civilization: 3, Knowledge Base:
1944] Caesar and Christ, New York, Simon and History
Schuster 1944.
[Durant Vol.4 Durant, Will, The Story of Civilization: 4, Knowledge Base:
322
MATH WITHOUT NUMBERS
Abbreviation Book Category
1950] The Age of Faith, New York, Simon and History
Schuster 1950.
[Durant Vol.5 Durant, Will, The Story of Civilization: 5, Knowledge Base:
1953] The Rennaissance, New York, Simon and History
Schuster 1953.
[Durant Vol.6 Durant, Will, The Story of Civilization: 6, Knowledge Base:
1957] The Reformation, New York, Simon and History
Schuster 1957.
[Durant Vol.7 Durant, Will and Ariel Durant, The Story Knowledge Base:
1961] of Civilization: 7, The Age of Reason Begins, History
New York, Simon and Schuster 1961.
[Durant Vol.8 Durant, Will and Ariel Durant, The Story Knowledge Base:
1963] of Civilization: 8, The Age of Louis XIV, History
New York, Simon and Schuster 1963.
[Durant Vol.9 Durant, Will and Ariel Durant, The Story Knowledge Base:
1965] of Civilization: 9, The Age of Voltaire, New History
York, Simon and Schuster 1965.
[Durant Vol. 10 Durant, Will and Ariel Durant, The Story Knowledge Base:
1967] of Civilization: 10, Rousseau and Revolution, History
New York, Simon and Schuster 1967.
[Durant Vol.11 Durant, Will and Ariel Durant, The Story Knowledge Base:
1975] of Civilization: 11, The Age of Napoleon, History
New York, MJF Books 1975.
[Ehrlich 1991] Ehrlich, Gertrude, Fundamental Concepts Abstract Algebra
of Abstract Algebra, Mineola, New York
Dover 1991.
[Enderton 1977] Enderton, Herbert B., Elements of Set Set Theory
Theory, New York, Academic Press
1977.
[Eves 1990] Eves, Howard, Foundations and Mathematics:
Fundamental Concepts of Mathematics, 3rd Foundations
ed., New York, Dover 1997.
[Freeley and Freeley, Austin J., and David L. Argumentation and
Steinberg 2009] Steinberg, Argumentation and Debate, Debate
Boston, Massachusetts, Wadsworth
Cengage Learning 2009.
[Frisch 1969] Frisch, Joseph C., Extension and Logic
Comprehension in Logic, New York,
Philosophical Library 1969.
[Funakoshi 1956] Funakoshi, Gichin, Karate-Do Kyohan Knowledge Base:
The Master Text, New York, Kodansha Martial Arts
International 1973.
[Ganter and Wille Ganter, Bernhard and Rudolf Wille, Formal Concept
1999] Formal Concept Analysis, New York, Analysis
Springer 1999.
[Ganter, Stumme, Ganter, Bernhard, Stumme, Gerd, and Formal Concept
and Wille 2005] Wille, Rudolf (Eds.), Formal Concept Analysis
Analysis Foundations and Applications,
Springer 2005.
[Givant and Givant, Steven and Paul Halmos, Logic Logic
Halmos 1998] as Algebra, The Mathematical
323
WILLIAM S. VEATCH
Abbreviation Book Category
Association of America 1998.
[Givant and Givant, Steven and Paul Halmos, Boolean Algebras and
Halmos 2009] Introduction to Boolean Algebras, New Rings
York, Springer 2009.
[Grtzer 2003] Grtzer, George, General Lattice Theory Lattice Theory
2d Ed., Boston, Birkhuser 2003.
[Grtzer 2008] Grtzer, George, Universal Algebra 2d Universal Algebra
Ed., New York, Springer 2008.
[Halmos 1960] Halmos, Paul R., Naive Set Theory, Set Theory
Princeton, Van Nostrand 1960.
[Hamilton] Hamilton, George Heard, The Library of Knowledge Base: Art
Art History: 19th and 20th Century Art,
New York, Harry N. Abrams, Inc.
[Hilbert and Hilbert, D. and W. Ackermann, Mathematical Logic
Ackermann 1938] Principles of Mathematical Logic,
Providence, Rhode Island, AMS
Chelsea Publishing 2008.
[Hill and Leeman Hill, Bill, and Richard W. Leeman, The Argumentation and
1977] Art and Practice of Argumentation and Debate
Debate, Mountain View, California
Mayfield Publishing Company 1997.
[Hitzler and Schrfe Hitzler, Pascal and Henrik Schrfe Formal Concept
2009] (Eds.), Conceptual Structures in Practice, Analysis
Baca Raton, Florida, CRC Press 2009.
[Hrbacek and Jech Hrbacek, Karel, and Thomas Jech, Set Theory
1999] Introduction to Set Theory, 3d Ed., Boca
Raton, Florida, CRC Press 1999.
[ICFCA 2004] Eklund, Peter (Ed.), Formal Concept Formal Concept
Analysis, 2nd International Conference, Analysis
International Conference on Formal Concept
Analysis (ICFCA), New York, Springer
2004.
[ICFCA 2009] Ferr, Sbastien, and Sebastian Formal Concept
Rudolph (Eds.), Formal Concept Analysis, Analysis
7th International Conference, ICFCA, New
York, Springer 2009.
[ICFCA 2011] Valtchev, Petko, and Robert Jschke Formal Concept
(Eds.), Formal Concept Analysis, 9th Analysis
International Conference, ICFCA, New
York, Springer 2011.
[ICFCA 2012] Domenach, Florent, Dmitry I. Ignatov, Formal Concept
Jonas Poelmans (Eds.), Formal Concept Analysis
Analysis, 10th International Conference,
ICFCA, New York, Springer 2012.
[ICFCA 2013] Cellier, Peggy, Felix Distel, Bernhard Formal Concept
Ganter (Eds.), Formal Concept Analysis, Analysis
11th International Conference, ICFCA,
New York, Springer 2013.
[ICFCA 2014] Glodeanu, Cynthia Vera, Mehdi Formal Concept
Kaytoue, Christian Sacarea (Eds.), Analysis
324
MATH WITHOUT NUMBERS
Abbreviation Book Category
Formal Concept Analysis, 12th International
Conference, ICFCA, New York, Springer
2014.
[Isaacs 2008] Isaacs, I. Martin, Finite Group Theory, Group Theory
Providence, Rhode Island, American
Mathematical Society 2008.
[Jech 2003] Jech, Thomas, Set Theory 3d Millennium Set Theory
Ed., New York Springer 2003.
[Jevons 1877] Jevons, W. S., Elementary Lessons in Logic, Logic
Macmillan and Co. 1877. Reprint by
Forgotten Books.
[Kaburlasos 2006] Kaburlasos, Vassilis G., Towards a Lattice Theory
Unified Modeling and Knowledge-
Representation based on Lattice Theory, New
York, Springer 2006.
[Kamke 1950] Kamke, E., Theory of Sets, New York, Set Theory
Dover, 1950.
[Klammer, Schulz, Klammer, Thomas P., Muriel R. Knowledge Base:
and Volpe 2013] Schulz, and Angela Della Volpe, English Grammar
Analyzing English Grammar 7th Ed., New
York, Pearson Education, Inc. 2013.
[Kneale and Kneale Kneale, William, and Martha Kneale, Logic
1971] The Development of Logic, Oxford,
Clarendon Press 1971.
[Kneebone 1963] Kneebone, G. T., Mathematical Logic and Mathematical Logic
the Foundations of Mathematics, New
York, D. Van Nostrand Company Ltd.
1963.
[Kolln, Gray, and Kolln, Martha, Loretta Gray, and Knowledge Base:
Salvatore 2016] Joseph Salvatore, Understanding English English Grammar
Grammar, New York, Pearson 2016.
[Kreeft 2010] Kreeft, P., Socratic Logic, South Bend, Logic
Indiana, St Augustines Press 2010.
[Kunen 2011] Kunen, Kenneth, Set Theory, London, Set Theory
College Publications 2011.
[Langer 1966] Langer, Susanne K., An Introduction to Logic
Symbolic Logic, New York, Dover 1967.
[Lawvere and Lawvere, F. William, and Rosebrugh, Set Theory
Rosebrugh 2003] Robert, Sets for Mathematics, University
of Cambridge 2003.
[Lee and Ricke Lee, Soon Man and Ricke, Gaetane, Knowledge Base:
1999] Modern Taekwondo, New York, Sterling Martial Arts
Publishing Co., Inc. 1999.
[Levy 1979] Levy, Azriel, Basic Set Theory, Mineola, Set Theory
New York, Dover 1979.
[Lipschutz 2012] Lipschutz, Seymour, General Topology, Topology
New York, McGraw Hill 2012.
[Maritain 1946] Maritain, Jacques, Formal Logic, New Logic
York, Seed & Ward 1946.
[Markman 1998] Markman, Arthur B., Knowledge Knowledge
325
WILLIAM S. VEATCH
Abbreviation Book Category
Representation, New York, Psychology Representation
Press 1998.
[McInerny 2004] McInerny, Dennis, La Logique Facile, Logic
Paris, Eyrolles 2004.
[Mill 1872] Mill, John Stuart, System of Logic Vol. II, Logic
London, Longmans, Green, Reader,
and Dyer 1872. Reprint by Forgotten
Books 2012.
[Minto 1893] Mito, William, Logic Inductive and Unabridged facsimile of
Deductive, Boston, Adamant Media the edition published in
2005. 1893 by John Murray,
London.
[Moore 2013] Moore, Gregory H., Zermelos Axiom of Set Theory
Choice, Mineola, New York, Dover
2013.
[Monk 1969] Monk, J. Donald, Introduction to Set Set Theory
Theory, New York McGraw-Hill, Inc.
1969.
[Nakayama, Nakayama, M., Best Karate 2 Knowledge Base: martial
Fundamentals Fundamentals, New York, Kodansha arts
1978] International 1978.
[Northcott 1968] Northcott, D.G., Ideal Theory, Ring Theory, Ideal
Cambridge University Press 1968. Theory
[Parker and Veatch Parker, Francis H. and Henry B. Logic
1959] Veatch, Logic as a Human Instrument,
New York, Harper & Brothers 1959.
[Parrochia and Parrochia, Daniel and Neuville, Pierre, Universal Logic,
Neuville 2013] Towards a General Theory of Classifications, Classification Theory
Springer Basel, 2013.
[Pinter 1990] Pinter, Charles C., A Book of Abstract Abstract Algebra
Algebra, 2d ed., Mineola, New York,
Dover, 1990.
[Pinter 2014] Pinter, Charles C., A Book of Set Theory, Set Theory
Mineola, New York, Dover 2014.
[Porphyry 260 C.E.] Porphyry, Introduction, Translated by Logic
Jonathan Barnes, New York, Oxford
University Press 2003.
[Potter 2009] Potter, Michael, Set Theory and its Set Theory
Philosophy, Oxford University Press
2009.
[Quine 1969] Quine, Willard Van Orman, Set Theory Set Theory, Logic
and Its Logic, Cambridge,
Massachusetts, Harvard University
Press 1969.
[Quine 1981] Quine, Willard Van Orman, Mathematical Logic
Mathematical Logic, Cambridge, Mass.,
Harvard University Press, 1981.
[Quine 1982] Quine, Willard Van Orman, Methods of Logic
326
MATH WITHOUT NUMBERS
Abbreviation Book Category
Logic, Cambridge, Massachusetts,
Harvard University Press 1982.
[Quine 1986] Quine, Willard Van Orman, Philosophy Logic
of Logic, Cambridge, Massachusetts,
Harvard University Press 1970.
[Read 1909] Read, Carveth, Logic Deductive and Logic
Inductive, London, Alexander Moring
Limited 1909. Reprinted by Forgotten
Books 2012.
[Rees 2012] Rees, Martin Ed., Universe, New York, Knowledge Base: Space
Dorling Kindersley (DK Smithsonian) Science
2012.
[Rigdon 1903] Rigdon, Jonathan, Grammar of the Knowledge Base:
English Sentence, London, Forgotten Astronomy and the
Books 2013. Planets
[Roberts and Zweig Roberts, Edgar V., and Robert Zweig, Knowledge Base:
2012] Literature An Introduction to Reading and literature
Writing, New York, Longman 5th
Compact Ed. 2012.
[Rodgers 2000] Rodgers, Nancy, Learning to Reason, Set Theory, Logic
New York, John Wiley & Sons, Inc.
2000.
[Roman 2008] Roman, Steven, Lattices and Ordered Sets, Lattice Theory
New York, Springer 2008.
[Russell 1920] Russell, Bertrand, Introduction to Philosophy
Mathematical Philosophy, New York,
Dover ed. 1993.
[Russell 1921] Russell, Bertrand, The Analysis of Mind, Philosophy
Mineola, New York, Dover 2005.
[Schrder 2003] Schrder, Bernd S. W., Ordered Sets, Lattice Theory
Boston, Birkhauser 2003.
[Shenefelt and Shenefelt, Michael, and Heidi White, If
White 2013] A, then B - How the World Discovered Logic, Logic
New York, Columbia University Press
2013.
[Shertzer 1986] Shertzer, Margaret D., The Elements of Knowledge Base:
Grammar, New York, Macmillan English Grammar
Publishing Company 1986.
[Sikorski 1964] Sikorski, Roman, Boolean Algebras, New Boolean Algebra
York, Springer-Verlag 1964.
[Singh 1959] Singh, Jagjit, Great Ideas of Modern Mathematics
Mathematics: Their Nature and Use, New
York, Dover Publications 1959.
[Smullyan 2014] Smullyan, Raymond M., A Beginners Mathematical Logic
Guide to Mathematical Logic, New York,
Dover Publications 2014.
[Solomon 1990] Solomon, A. D., The Essentials of Boolean Boolean Algebra
Algebra, 1990.
[Sowa 1984] Sowa, John F., Conceptual Structures: Knowledge
327
WILLIAM S. VEATCH
Abbreviation Book Category
Information Processing in Mind and Machine, Representation
Menlo Park, California, Addison-
Wesley Publishing Company 1984.
[Sowa 2000] Sowa, John F., Knowledge Representation, Knowledge
Pacific Grove, California, Brooks/Cole Representation
2000.
[Stewart 2004] Stewart, Ian, Galois Theory, New York, Galois Theory
Chapman & Hall/CRC 2004.
[Stewart and Tall Stewart, Ian and Tall, David, The Set Theory
1977] Foundations of Mathematics, Oxford
University Press 1977.
[Stoll 1963] Stoll, R..R., Set Theory and Logic, New Set Theory
York, Dover 1963.
[Stone 1936] Stone, M.H., The Theory of Boolean Algebras
Representation for Boolean Algebras,
Trans. Amer. Math. Soc. 40, 37-111
(1936).
[Stone 1937] Stone, M.H., Applications of the Boolean Algebras
Theory of Boolean Rings to General
Topology, Trans. Amer. Math. Soc.
41, 375-481 (1937).
[Strunk and White Strunk Jr., William and E.B. White, The Knowledge Base:
2000] Elements of Style, Boston, Allyn and English Grammar
Bacon 2000.
[Sullivan 2005] Sullivan, Scott M., An Introduction to Logic
Traditional Logic, North Charleston, SC,
Booksurge Publishing 2006.
[Suppes 1972] Suppes, Patrick, Axiomatic Set Theory, Set Theory
New York, Dover 1972.
[Tall 2014] Tall, Aliou, From Mathematics in Logic, to Logic
Logic in Mathematics, Boston, Docent
Press 2014.
[Thurman 2003] Thurman, Susan, The Only Grammar Knowledge Base:
Book Youll Ever Need, Avon English Grammar
Massachusetts, Adams Media 2003.
[Tiles 1989] Tiles, Mary, The Philosophy of Set Theory, Set Theory
Mineola, New York, Dover 1989.
[Toulmin 2003] Toulmin, Stephen E., The Uses of Argumentation
Argument, New York, Cambridge
University Press 2008.
[Toulmin, Rieke, Toulmin, Stephen, Richard Rieke, and Argumentation
and Janik 1979] Allan Janik, An Introduction to Reasoning,
New York, Macmillan Publishing Co.
1979.
[van der Waerden Van der Waerden, B.L., Algebra Vol. I, Abstract Algebra
1966] New York, Springer 1970.
[van der Waerden Van der Waerden, B.L., Algebra Vol. II, Abstract Algebra
328
MATH WITHOUT NUMBERS
Abbreviation Book Category
1967] New York, Springer 1970.
[Venn 1894] Venn, John, Symbolic Logic, London, Logic
Macmillan and Co., Forgotten Books
ed. 2012.
[Wallace 1998] Wallace, D.A.R., Groups, Rings and Group Theory, Ring
Fields, London, Springer 1998. Theory
[Walton 2016] Walton, Douglas, Argument Evaluation Argumentation
and Evidence, New York, Springer 2016.
[Warren 2013] Warren, Rebecca, Knowledge Knowledge Base:
Encyclopedia, New York, Dorling General
Kindersley (DK Smithsonian) 2013.
[Whately 1836] Whately, Richard, Elements of Logic, New Logic
York, William Jackson 1836.
[Willard 1970] Willard, Stephen, General Topology, Topology
Mineola, New York, Dover 1970.
[Woodward 2013] Woodward, John, Geography A Visual Knowledge Base:
Encyclopedia, New York, Dorling Geography
Kindersley (DK Smithsonian) 2013.
329
WILLIAM S. VEATCH
Index
A G
Absolute Complement, 85 Galois Theory, 121
Antichain, 195 Group, 303
Atom, 9
Attribute, 10
Attribute Signature, 177
Axiom, 248
B H
Base Set, 130 Hasse Diagram, 310
Binomial Theorem,127 Hierarchy Tree, 207
Boolean Algebra, 131
Boolean Ring, 132
C I
Chain, 169 Ideal, 132
Classical Logic, 16 Idea Signature, 41, 111, 113
Coatom, 9, 88 Infinite, 62
Commutative Ring, 308 Inheritance, 40
Complement, 37, 84 Intension, 10
Compound Idea, 9 Intersection, 81, 102
Cover, 84 Is a Relation, 252
Cover Plus a Difference, 45
D L
Direct Product, 129 Lattice, 73
Disjoint, 78, 94 List Set, 24
Distributive Property, 85
Domain, 46, 254, 261, 266, 272
Down Set, 144
Duality, 110, 116
E M
Equivalence Class, 29, 32 Mutually Exclusive, 26
Equivalence Relation, 22, 113 MWN, i
Euler Diagram, 15
Exhaustive, 190
Extension, 10
Extent, 21
F N
Filter, 108, 132 Natural numbers, 262
Finite, 62 Nested Lattice, 186, 200
Nested Partition Equation, 87, 154
330
MATH WITHOUT NUMBERS
O
Object, 10
Order, 73
Outline, 219
P
Pairwise Disjoint,192
Paradoxes,
Russells, 277
Partition, 190
Partition Equation, 153, 202
Power Set, 9, 122
Power Set Contraction, 50, 151, 268
Power Set Expansion, 50, 151, 268
Proposition, 7, 11
R
Relative Complement, 85
Ring, 308
S
Set, 54 et seq.
Simplified Partition Equation, 290
Substitution Principle, 157
Successor Set, 258, 262
Sum Decomposition, 139
Syllogism, 8, 11, 16
Symmetric Difference, 83, 103
T
Test Set, 27
Topology, 311
U
Union, 81, 102
Up-Set, 108, 144
V
von Neumann, 256, 284
331
ABOUT THE AUTHOR
William S. Veatch is a practicing attorney, living in San Francisco, California.
He obtained his B.A. degree in History at the University of Winnipeg,
Winnipeg, Manitoba, Canada (1985), where he also studied Mathematics and
Philosophy; his LL.B. law degree at the University of Manitoba, Winnipeg,
Canada (1985); and his J.D. law degree from the University of California,
Hastings College of the Law, San Francisco, California (1987).
332
You might also like
- Th. F. Stcherbatsky - Buddhist Logic - in Two Volumes. 1-Dover (1962)100% (1)Th. F. Stcherbatsky - Buddhist Logic - in Two Volumes. 1-Dover (1962)573 pages
- (Synthese Historical Library 12) E. J. Ashworth (Auth.) - Language and Logic in The Post-Medieval Period-Springer Netherlands (1974)100% (1)(Synthese Historical Library 12) E. J. Ashworth (Auth.) - Language and Logic in The Post-Medieval Period-Springer Netherlands (1974)320 pages
- Cognition and Communication Within The Oral MajorityNo ratings yetCognition and Communication Within The Oral Majority27 pages
- Complete Reasoning Study Notes For Bank ExamNo ratings yetComplete Reasoning Study Notes For Bank Exam3 pages
- Annotated Bibliography of Neo-Scholasticism PDFNo ratings yetAnnotated Bibliography of Neo-Scholasticism PDF396 pages
- An Introduction To Islamic Decoloniality: DR Syed Mustafa Ali The Open University (UK) S.m.ali@open - Ac.uk100% (1)An Introduction To Islamic Decoloniality: DR Syed Mustafa Ali The Open University (UK) S.m.ali@open - Ac.uk66 pages
- PhA 122 - David The Invincible - Commentary On Aristotle's Prior Analytics PDFNo ratings yetPhA 122 - David The Invincible - Commentary On Aristotle's Prior Analytics PDF232 pages
- Legal Reasoning And Methodlogy: Nhận Định - True FalseNo ratings yetLegal Reasoning And Methodlogy: Nhận Định - True False15 pages
- Test Bank For Principles of Comparative Politics 3rd Edition Golder Clark 1506318126 9781506318127 - Download Today With Full Content100% (5)Test Bank For Principles of Comparative Politics 3rd Edition Golder Clark 1506318126 9781506318127 - Download Today With Full Content42 pages
- Idea Essence Existence and Archetype On PDFNo ratings yetIdea Essence Existence and Archetype On PDF21 pages
- Gaisi Introduction To Axiomatic Set Theory100% (1)Gaisi Introduction To Axiomatic Set Theory258 pages
- Political Objectives Military Objectives WAR: The Relationship BetweenNo ratings yetPolitical Objectives Military Objectives WAR: The Relationship Between84 pages
- Getting Through - Legal Technique and Logic Summary (Aquino's Book)No ratings yetGetting Through - Legal Technique and Logic Summary (Aquino's Book)13 pages
- Apologetics: Night 2 Presented by Eric DoumaNo ratings yetApologetics: Night 2 Presented by Eric Douma8 pages
- How To Think: A Brief Summary of How Thought Works: IndividualNo ratings yetHow To Think: A Brief Summary of How Thought Works: Individual11 pages
- Gillespie, The Aristotelian Categories, 1925No ratings yetGillespie, The Aristotelian Categories, 192511 pages
- Week 3-4 Gec Chapter 3 PROBLEM SOLVING AND REASONINGNo ratings yetWeek 3-4 Gec Chapter 3 PROBLEM SOLVING AND REASONING7 pages
- Philosophy: Logic Final Examination I. On Categorical SyllogismNo ratings yetPhilosophy: Logic Final Examination I. On Categorical Syllogism2 pages
- Neoplatonism and The Philosophy of NaturNo ratings yetNeoplatonism and The Philosophy of Natur9 pages
- 2015.189506.probability Statistics and Truth - Text PDF100% (2)2015.189506.probability Statistics and Truth - Text PDF263 pages
- Burgess & Rosen-A Subject With No Object - Strategies For Nominalistic Interpretation of Mathematics - Oxford University Press, USA (1997)No ratings yetBurgess & Rosen-A Subject With No Object - Strategies For Nominalistic Interpretation of Mathematics - Oxford University Press, USA (1997)140 pages
- Feferman - Infinity in Mathematics - Is Cantor NecessaryNo ratings yetFeferman - Infinity in Mathematics - Is Cantor Necessary66 pages
- An Introduction To Finite Projective Planes by Albert A.A., Sandler RNo ratings yetAn Introduction To Finite Projective Planes by Albert A.A., Sandler R112 pages
- Woodin - The Axiom of Determinacy, Forcing Axioms and The Non-Stationary Ideal100% (1)Woodin - The Axiom of Determinacy, Forcing Axioms and The Non-Stationary Ideal859 pages
- Paolo Mancosu - Philosophy of Mathematics and Mathematical Practice in The Seventeenth Century (1996)100% (1)Paolo Mancosu - Philosophy of Mathematics and Mathematical Practice in The Seventeenth Century (1996)285 pages
- Universal Arithmetick - Raphson - 1707 - Isaac NewtonNo ratings yetUniversal Arithmetick - Raphson - 1707 - Isaac Newton286 pages
- An Episodic History of Mathematics Mathematical Culture Through Problem Solving, MAANo ratings yetAn Episodic History of Mathematics Mathematical Culture Through Problem Solving, MAA483 pages
- (Ideas in Context) Reviel Netz-The SHAPING of DEDUCTION in GREEK MATHEMATICS A Study in Cognitive History-Cambridge University Press (2003)100% (1)(Ideas in Context) Reviel Netz-The SHAPING of DEDUCTION in GREEK MATHEMATICS A Study in Cognitive History-Cambridge University Press (2003)396 pages
- Proof, Logic, and Conjecture - The Mathematician's Toolbox PDFNo ratings yetProof, Logic, and Conjecture - The Mathematician's Toolbox PDF481 pages
- Rota's Philosophy in Mathematical ContextNo ratings yetRota's Philosophy in Mathematical Context40 pages
- LOGIC CHAPTER-5 (Fallacies) - Short VersionNo ratings yetLOGIC CHAPTER-5 (Fallacies) - Short Version55 pages
- Ramsey - Foundations of Mathematics PDFNo ratings yetRamsey - Foundations of Mathematics PDF94 pages
- Historical Changes in The Concepts of Number PDFNo ratings yetHistorical Changes in The Concepts of Number PDF29 pages
- Eurocentrism in The History and Philosopny of MathematicsNo ratings yetEurocentrism in The History and Philosopny of Mathematics6 pages
- V Sankrithi Krishnan An Introduction To Category Theory North Holland 1981100% (2)V Sankrithi Krishnan An Introduction To Category Theory North Holland 1981181 pages
- The Way To Geometry: [Being Necessary and Usefull for Astronomers, Enginees, Geographers,. Architects, Land-meaters, Carpenters, Sea-men & Etc.]From EverandThe Way To Geometry: [Being Necessary and Usefull for Astronomers, Enginees, Geographers,. Architects, Land-meaters, Carpenters, Sea-men & Etc.]No ratings yet
- Does Mathematical Study Develop Logical ThinkingNo ratings yetDoes Mathematical Study Develop Logical Thinking5 pages
- Heisenberg - Abstraction in Modern Science100% (1)Heisenberg - Abstraction in Modern Science15 pages
- Set Theory: An Introduction To The World of Large Cardinals: Expanded Notes of A Course in Four Lectures Delivered at The100% (1)Set Theory: An Introduction To The World of Large Cardinals: Expanded Notes of A Course in Four Lectures Delivered at The49 pages
- Constructive Formalism. Essays On The Foundations of Mathematics (University College Leicester, 1951) by R. L. Goodstein PDFNo ratings yetConstructive Formalism. Essays On The Foundations of Mathematics (University College Leicester, 1951) by R. L. Goodstein PDF89 pages
- DM Johnson - The Problem of The Invariance of Dimension in The Growth of Modern TopologyNo ratings yetDM Johnson - The Problem of The Invariance of Dimension in The Growth of Modern Topology92 pages
- Mill S A System of Logic Critical Appraisals 1st Edition Antis Loizides 2024 Scribd Download100% (3)Mill S A System of Logic Critical Appraisals 1st Edition Antis Loizides 2024 Scribd Download45 pages
- Giaquinto - Hilbert's Philosophy of Mathematics100% (1)Giaquinto - Hilbert's Philosophy of Mathematics15 pages
- Undergraduate Texts in Mathematics: S. AxlerNo ratings yetUndergraduate Texts in Mathematics: S. Axler11 pages
- Chapter 7, 3 Variations of The Categorial SyllogismNo ratings yetChapter 7, 3 Variations of The Categorial Syllogism2 pages
- The Realm of The Infinite: W. Hugh WoodinNo ratings yetThe Realm of The Infinite: W. Hugh Woodin35 pages
- Projective Geometry and TransformationsNo ratings yetProjective Geometry and Transformations37 pages
- Travels With John Conway, in 258 Septillion DimensionsNo ratings yetTravels With John Conway, in 258 Septillion Dimensions9 pages
- Restitution of the Two Books on Plane Loci of Apollonius of Perga together with On Spheric Contacts by Pierre de Fermat (1601-1665)From EverandRestitution of the Two Books on Plane Loci of Apollonius of Perga together with On Spheric Contacts by Pierre de Fermat (1601-1665)No ratings yet
- The Actual Infinity in Cantor's Set Theory (G.mpantes)100% (1)The Actual Infinity in Cantor's Set Theory (G.mpantes)13 pages
- Incompleteness The Proof and Paradox of Kurt Godel PDFNo ratings yetIncompleteness The Proof and Paradox of Kurt Godel PDF9 pages
- H. Vic Dannon - Cantor's Set, The Cardinality of The Reals, and The Continuum HypothesisNo ratings yetH. Vic Dannon - Cantor's Set, The Cardinality of The Reals, and The Continuum Hypothesis16 pages
- From Euclidean Geometry To Transformation GeometryNo ratings yetFrom Euclidean Geometry To Transformation Geometry10 pages
- Does Mathematics Need New Axioms?: Solomon FefermanNo ratings yetDoes Mathematics Need New Axioms?: Solomon Feferman19 pages
