
Citation 
 Permanent Link:
 http://ufdc.ufl.edu/AA00004701/00001
Material Information
 Title:
 Subset systems and generalized distributive lattices
 Creator:
 Zenk, Eric R
 Publication Date:
 2004
 Language:
 English
 Physical Description:
 vi, 110 leaves : ; 29 cm.
Subjects
 Subjects / Keywords:
 Adjoints ( jstor )
Algebra ( jstor ) Conceptual lattices ( jstor ) Factorization ( jstor ) Functors ( jstor ) Isomorphism ( jstor ) Mathematics ( jstor ) Monads ( jstor ) Partially ordered sets ( jstor ) Universal algebra ( jstor ) Dissertations, Academic  Mathematics  UF ( lcsh ) Mathematics thesis, Ph. D ( lcsh )
 Genre:
 bibliography ( marcgt )
theses ( marcgt ) nonfiction ( marcgt )
Notes
 Thesis:
 Thesis (Ph. D.)University of Florida, 2004.
 Bibliography:
 Includes bibliographical references.
 General Note:
 Printout.
 General Note:
 Vita.
 Statement of Responsibility:
 by Eric R. Zenk.
Record Information
 Source Institution:
 University of Florida
 Holding Location:
 University of Florida
 Rights Management:
 Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for nonprofit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
 Resource Identifier:
 3101123 ( ALEPH )
AA00004701_00001 ( sobekcm ) 767968358 ( OCLC )

Downloads 
This item has the following downloads:

Full Text 
SUBSET SYSTEMS AND GENERALIZED DISTRIBUTIVE LATTICES
By
ERIC R. ZENK
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
2004
SUBSET SYSTEMS AND GENERALIZED DISTRIBUTIVE LATTICES
By
ERIC R. ZENK
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
2004
ACKNOWLEDGMENTS
I would like to thank 
Jorge Martinez. Sometimes words are inadaquate. He has been kind, generous, patient,
and interesting to work with. I have grown under his guidance.
The people who inhabited the University of Florida math department from 1999 to
2004. They make the department the wonderful place it is.
My friend Cielo: earth is wonderful, when one can see into the sky.
My M. L.
My mathematical siblings other graduate students who worked with Jorge, especially
Ricardo Carrera.
My family Wayne, Phyllis, Margie, Jeff, and Rob.
Those teachers who, by challenging me, caused me to improve.
Participants in the fRings and Ordered Algebraic Structures conferences in Gaines
ville and Nashville during my time in graduate school. Because of these conferences, I
feel like I am joining a family of researchers, rather than just getting a degree. It is
an honor to know them.
Those who helped proofread this document: constructive comments were made by
Jorge Martinez, Scott McCullough, and Pham Tiep.
Anyone taking the time to read these words: a dissertation, like any book, is meant to
be read.
u
TABLE OF CONTENTS
ACKNOWLEDGEMENTS
ABSTRACT v
CHAPTERS
1 INTRODUCTION 1
1.1 Distributive Lattices 1
1.2 Subset Systems
1.3 Methods and Results 5
2 PRIMER ON CATEGORIES AND POSETS 8
2.1 Distinguished Maps
2.2 Bounds 12
2.3 Natural Transformations 17
2.4 (Co)Limits 19
2.5 Adjoint Functors 23
3 ALGEBRAS OF A MONAD 28
3.1 Categories of Algebras 28
3.2 Adjoint Connections induce Monads 32
3.3 Detecting Categories of Algebras 34
3.4 Distributive Laws 37
4 GENERATING SUBMONADS 42
4.1 Subfunctors 42
4.2 Meseguers Lemmas 50
4.3 A Partial Algebra Which Does Not Extend 59
5 FREE ALGEBRAS 63
5.1 Complete semilattices 64
5.2 Completely Distributive Complete Lattices 68
5.3 Some categories of algebras 71
6 COEQUALIZERS 82
6.1 Epis and Equalizers in P 82
iii
6.2 Factorization of Maps Using Preorders 85
6.3 Factorization of Meetsemilattice maps 88
6.4 Coequalizers in DU, 90
7 (j,m)SPACES 93
7.1 Spatial/Sober Functorial Galois Connection 94
7.2 The Skula Topology and Extremal Monos 97
7.3 Computing Limits 100
7.4 Quotients, Extremal and Regular Epis 101
7.5 Flat Spectra 104
7.6 Epicomplete objects in 105
108
REFERENCES
BIOGRAPHICAL SKETCH
no
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
SUBSET SYSTEMS AND GENERALIZED DISTRIBUTIVE LATTICES
By
Eric R. Zenk
August 2004
Chairman: Jorge Martinez
Major Department: Mathematics
Distributive lattices alone, or with enriched structure are mathematical objects
of fundamental importance. This text studies generalized distributive lattices: the general
ization is that certain infinite meets and joins are required to exist. Subset systems (natural
rules which select a family of subsets of each poset) j and m label which sets have joins and
meets, respectively.
A calculus of subfunctors is developed: using this calculus, it is shown that any
subfunctor F of a monad (containing the image of the unit) generates a submonad F. Under
suitable conditions, any partial Falgebra extends to an Falgebra. The monad F for the free
distributive (j. m)complete lattice is the submonad of the completely distributive complete
lattice monad generated by a subfunctor obtained from j and m.
The category DP^ of (j. m)complete lattices which can be embedded in a completely
distributive complete lattice is a full subcategory of Falgebras. DP(n is complete and has
coequalizers.
(j, m)complete families of subsets of a set (generalized topological spaces) are inves
tigated in analogy to classical pointset topology. Assuming suitable restrictions on j and
m, subspaces can be defined. Assuming these restrictions, there are wellbehaved categories
corresponding to To and sober spaces.
CHAPTER 1
INTRODUCTION
1.1 Distributive Lattices
A semilattice is a set with an operation A satisfying the following universally quantified
equations
a A a = a,
a A b b A a, and
(a A b) A c = a A (b A c).
A and V, which are related by the condition
a V b b.
< b to mean a A b = a; with this order a A b
is the largest thing smaller than both a and b and a V b is the smallest thing larger than both
a and b.
A lattice is distributive if either of the following, equivalent, universally quantified
equations hold
a A (b V c) (a A b) V (a A c),
a V (b A c) = (a V b) A (a V c).
Let us consider the concept and its relevance.
A distributive lattice bears some resemblance to ordinary arithemetic where A and
V correspond to addition and multiplication; the principal difficulty with this view is that
a A a = a V a = a, which does not hold in arithmetic. There is more symmetry in the
A lattice is a set with two semilattice operations
that
a A b = a 4=>
A lattice may be partially ordered by defining a
1
2
equations defining distributive lattices than in ordinary arithmetic; formally A and V are
interchangable, and switching them reverses the order. Distributive lattices are interesting
algebraic structures in the same right as rings structures with + and sensibly defined.
Another perspective is that distributive lattices are models of logic, with A repre
senting and and V representing or. Obviously, the connectives and and or both
satisfy the semilattice rules. In this interpretation, the distributive laws are tautologies and
a < b means the proposition a implies b. Conventional logics are often described by dis
tributive lattices obeying extra equations, which correspond to additional tautologies to be
modeled.
A perspective particularly revelant to the author is that distributive lattices (with
some additional structure) describe topological situations. Intuitively, a topological space
is an amorphous blob, from which certain pieces can be cleanly removed. The removable
pieces are called closed parts and the complements (i.e, things left over after a closed part
has been removed) are called open parts. The usual definition of a topological space is a set
X, together with a designated family of open subsets, such that
X and the empty subset are open,
if U and V are open, so is the intersection U ft V, and
if (Ui) is a family of open subsets, then the union Jt f/ is also open.
The lattice of open sets encodes how a space is woven together. Real analysis provides some
justification for the usual definition of a topological space. However, the author wondered
how the concept of a topology changes if one varies the definition by requiring either fewer
unions of open sets be open, or more intersections of open sets be open.
1.2 Subset Systems
A category is an abstract class of objects with structure, and maps (or, homomor
phisms) which preserve the relevant structure. The maps in a category allow comparisons
3
between objects. Category theory allows formal comparisons between various theories of
algebraic objects: e.g., one can compare the category of all rings with the category of all
distributive lattices.
Many categories of partially ordered sets (with additional structure) fit into a simple,
general pattern. The objects are partially ordered sets in which certain intentionally distin
guished subsets have suprema or nfima while the maps are order preserving functions which
preserve said nfima and suprema. P denotes the category of all posets and order preserving
maps.
The challenge is how one selects subsets which have nfima and suprema. We follow
Thatcher, Wright and Wagner [31], who introduced the useful (but blandly named) concept
of a subset system.
Definition and Remarks 1.2.1. A subset system Z, is a rule which assigns a family Z(A)
of subsets to any poset A, such that for any order preserving map / : A > B,
{,f(S) : 5 G Z(A)} C Z(B).
Zcomplete posets are posets A in which each set S G Z(A) has a supremum. (See Erne [7]
for more extensive bibliography regarding subset systems.)
The following examples of subset systems may convey the generality and usefulness
of the concept.
1. If k is a cardinal, we use n to denote the subset system which selects all subsets
with cardinality less than k. We use oo for the functor which places no restriction on
cardinality. This is a subset system because for any function / : A > B, S C A, with
A < k implies /(A) < n.
2. A subset S C A is (upward) directed if for each x, y G S, there exists u G S such that
x < u and y < it. The rule dir which selects all directed subsets of a poset is a subset
system: if f{x),f(y) G /(S') and S is directed, then there is u G S such that x < u
and y < u. Thus f(u) is a common upper bound for f(x),f(y).
4
3. A subset S C A is (upward) compatible if for any x,y 6 S there is u A such that
x < u and y < u. Compatible differs from directed because for the former u 6 S,
while for the later we only require u G A. Similar arguments show that compat, which
selects compatible subsets is a subset system.
4. A subset of S C A is a chain if x, y S implies x < y or y < x. The rule ch which
selects all chains in a poset is a subset system.
5.We say a subset 5 C A is (upward) selfbounded if there is s 6 S such that for all
x S, x < s; a selfbounded set contains a maximum element. The rule sb, which
selects all selfbounded sets, is a subset system. Note that any order preserving map
preserves joins of upward selfbounded sets.
6. A nonexample: An antichain is a set of pairwise incomparable elements. The rule
ac which selects antichains is not a subset system, because there is an order preserving
surjection / : D > 2, where D is the twopoint antichain and 2 is the two point chain.
7. Generating examples: Let Q be any class of posets closed under order preserving
surjections. One may define a subset system Zq by
Za(A) = {S C A : S e Q}.
This construction shows there is a great multitude of subset systems. The subset sys
tem, compat, described above, is not generated this way, because one cannot determine
if S C A is compatible merely by looking at the poset S with its induced order.
We use subset systems, which we generically call j and m, to select which subsets
have joins and meets, respectively. Now we enumerate some categories of interest in this
discussion.
 the category of all (j, m)complete posets: that is, posets in which jsuprema and
TOinfima exist and are preserved by all maps.
5
DP^ the category of all completely distributive complete lattices.
DP^ the full subcategory of P^ containing objects which can be P^embedded in a
completely distributive complete lattice. DP^ is discussed in Section 5.3.
M, the full subcategory of P^ containing posets with an F structure, where F is
the monad defined in Section 5.3.
SpFJm the full subcategory of P, containing spatial posets. See Chapter 7.
 the category of generalized spaces with (j, m)complete families of distinguished
subsets: see Chapter 7.
FÂ£, the full subcategory of SpH?Jm containing spatial posets with flat spectrum: A is
defined to have flat spectrum if the maps A + 2 are trivially ordered: see Section 7.5.
1.3 Methods and Results
The discussion of DP^ and Mjn, which offer generalizations of distributive lattices,
uses the language of category theory. A quick introduction occurs in Chapter 2. The crucial
notion of a free object is formalized by monads, which are introduced in Chapter 3.
Chapter 4 describes a theory of subfunctors. The class of subfunctors of a functor
F bears a strong similarity to the power set lattice of a set X. Given a monad, with
functor part T, Meseguers Lemmas 4.2.1, 4.2.2, 4.2.6, and 4.2.7, show that any subfunctor
F of T containing all constants has a monadic closure, i.e., a smallest submonad F of T
which exceeds F. (Meseguers Lemmas were formulated and proved by the author, but the
technique is similar to one in Meseguer [25].) Intuitively, TX is the full set of polynomials (in
the sense of universal algebra) with variables in X, FX is a natural subset of polynomials,
and FX is the smallest natural subset of polynomials which is closed under composition
and contains FX. An algebra structure for T is a way of evaluating all polynomials; a
partial algebra structure for F is a way of evaluating polynomials in F." Under suitable
conditions, partial algebras extend to Falgebras.
6
In Chapter 5, Meseguers Lemmas are brought to bear upon monads for free complete
semilattices and free completely distributive complete lattices. Given any subset systems j
and m, there is a submonad F of the free completely distributive complete lattice monad.
The category MÂ£, of Falgebras offers a (somewhat mysterious) generalization of the category
of distributive lattices. The subcategory BP^, containing all (j, m)complete posets which
may be embedded in a completely distributive lattice is somewhat easier to understand and
still well behaved.
The existence of free objects in OP^ contrasts with the nonexistence of free objects
in PÂ£Â£ [9] and the category of complete Boolean algebras [8, 9], A fundamental difference
between these categories and DP, is the requirement that joins and meets obey a distributive
law.
The power of category theory comes as much from what it ignores as what it examines.
Significant conclusions are often obtained without examining the grubby details of what
is going on. But this innocence of grubby details limits the scope of investigation. In
the case of this document, several nicely behaved categories M?m, PI,, and SpÂ¥Jm are
introduced. For general subset systems j and m, the author does not even know if these
categories differ! The end of Section 5.3 from Corollary 5.3.10 onwards describes most
of the authors knowledge on the relationship between these categories.
Chapter 6 explores congruences, quotients and coequalizers in P, P3m and OH,. Much
classical algebra (ring theory, lattice theory, group theory, etc.) is simplified by the fact that
any surjection is a regular epimorphism. For the categories introduced here, the situation is
not so simple. Example 6.2.5 shows that a jjoin preserving surjective image of a jcomplete
poset need not be jcomplete.
The results of Chapter 7 predate the other results presented here. Herrlich [12]
contains a detailed examination of reflections (and coreflections) in categories of topological
spaces. This dissertation aimed to generalize results summarized in Herrlich [12], for (j,m)
spaces. A (j, m)space consists of an underlying set and a family of open subsets which are
7
closed under junions and mintersections. Continuous maps of (j, m)spaces are functions
such that preimages of open sets are open. The initial aim was to find reflections and
coreflections of the category of (j, m)spaces (obeying a Tostyle separation axiom), and
study how the existence and properties of reflections and coreflections varied depending
upon the subset systems j and m. An obvious prerequisite to such a project is knowledge
of factorizations of continuous maps.
The chapter contains a description of (j, m)subspaces and (j, m)quotients. In addi
tion, Section 7.5 describes a reflection of (j, m)spaces that corresponds to the 7) reflection
of topological spaces. Lastly, Section 7.6 describes epicomplete (j, m)spaces.
CHAPTER 2
PRIMER ON CATEGORIES AND POSETS
The text assumes a familiarity with the theory of sets typically used in mathematical
arguments. So familiar constructions unions, intersections, cartesian products, quotients
by equivalence classes, functions, Zorns Lemma, and transfinite induction are used without
further comment. (See Halmos [10] if this background is needed.) A basic familarity with
general topology is helpful.
Also some comfort with category theory is assumed. Namely, the reader can fill in
the blanks in the following informal definitions.
A category A consists of a class of objects Obj(A) and maps Map(A), such that each
object has an identity map, and there is an associative notion of composition of maps.
The set of Amaps from A\ to A2 is denoted A(Ai, A2).
A functor F : A > B assigns each A G Obj(A) an object F(A) Obj(B) and
each Amap / : Aj A2 a Bmap Ff : F(A1) > F(A2). The assignment respects
composition and identity arrows.
If A is a category, Aop is the category with the same objects as A, but all arrows
reversed. For a category theoretic concept C, the dual is obtained by applying C to
Ap.
A contravariant functor A + B is a functor A Bop.
Diagrams are used to display the behavior of a collection of maps; a diagram commutes
if any composites with the same domain and codomain are equal. For example, the
8
9
diagram
W
f
h
Y
commutes if and only if gf ih.
i
Recall the following properties of functors:
9
Definition 2.0.1. Let F : A > B be a functor. For each A\,A2 E A, F gives a function
from the homset A(Ai, A?) into rB(FAi, FA2) by
(/ : At A2) ~ (Ff : Ax A2).
If, for each Ai and A2 this map is onto, then F is said to be full. If, for each A\ and A2 this
map is onetoone, then F is said to be faithful.
A full subcategory of a category A is a category B such that Obj(B) C Obj(A) and
all / : A > B with A, B G Obj(B) are Bmaps. B C A is full if, and only if, the inclusion
functor is full.
Good general references for category theory are MacLane [21], Borceux [6], and Her
rlich and Strecker [13]. MacLane [21] gives a concise, high level summary of most category
theory and includes a chapter on monads. Herrlich and Strecker [13] is quite user friendly and
concretely describes many examples of adjoint functors. Borceux [6] covers a large amount
of material; the exposition is clear and very detailed.
2.1 Distinguished Maps
Definition and Remarks 2.1.1. Begin by defining a dual pair of concepts which coincide
with the notions injective and surjective in the category Set.
1. A map / : A\ > A2 is epi, a.k.a epic (in noun form, an epimorphism) if
whenever g and h are maps A2 A3 such that gf = hf, then g = h.
10
2. A map / : A2 > A3 is mono, a.k.a. monic (in noun form, a monomorphism)
if whenever g and h are maps A\ > A2 and such that fg = fh, then g h.
One may verify that a composition of epimorphisms (resp. monomorphisms) is epi (resp.
mono). Moreover, if f = ab is epi (resp. mono), then a is also epi (resp. b is also mono).
Definition 2.1.2. A map / : A > B is an isomorphism, if there is g : B A such that
id^ = gf and idB = fg.
In most categories of sets with structure: a map is mono if and only if it is injective,
surjective maps are epi, but epimorphisms may not be surjective.
Example 2.1.3. Consider tfAb the category of torsionfree abelian groups, i.e., abelian
groups such that
na = 0 => a = 0
for any natural number n and group element a, together with group homomorphisms. The
inclusion i : Z Q of the integers in the rational numbers is epi, but not onto.
In categories of sets with relational structure, bijective maps are not necessarily
isomorphisms.
Example 2.1.4. Consider Top the category of topological spaces and continuous maps.
The identity function i : > R from the reals (with discrete topology) to the reals (with
the usual topology), is a continuous bijection. However, the inverse function i_1 is not
continuous.
Example 2.1.5. Consider P the category of partially ordered sets and order preserving
maps. Either bijection
two elements is order preserving. But the inverse function 4>_1 is not order preserving.
For further discussion and more examples of epimorphisms and monomorphisms see
Herrlich and Strecker [13, Section 6] and Borceux [6, Volume 1, Sections 1.7 and 1.8].
11
Definition and Remarks 2.1.6. Consider a pair of maps /, 5 : Ai > A2. A map i : Aq >
A\ rightidentifies f and g if fi = gi. A map i : Aq > A\ is called the equalizer of / and g
if:
(eql) i rightidentifies / and g, and
(eq2) i has the feature that whenever j : B0 > A\ rightidentifies / and g, there
is a unique map e : Bo > Aq such that j ie.
The definite article is used for equalizers, because (eq2) implies that if i : Ao A and
i' : A'0 > A axe equalizers for / and g, then there is an isomorphism j : Aq Aq such that
i' ij. Notation: i = eq(f,g).
If there are / and g such that i : A0 > A\ is the equalizer of / and g, then i is regular
mono. Regular monomorphisms are monomorphisms. If / is epi and regular mono, then / is
an isomorphism. For proofs of the assertions in this paragraph and a discussion of examples,
see Borceux [6, Volume 1, Section 2] or Herrlich and Strecker [13, Section 16].
The definitions of coequalizer and regular epi are dual to equalizer and regular
mono, but are repeated for emphasis, e : A2 > A3 leftidentifies f and g if ef eg. A
map e : A2 A3 is the coequalizer of / and g if
(coeql) e leftidentifies / and g, and
(coeq2) e has the feature that whenever d : A2 B3 leftidentifies / and g, there is
a unique map c : A3 B3 such that e = cd.
Notation: e = coeq(/, 5). If e is the coequalizer of some pair of maps, then e is called
regular epi The duals of all basic properties of regular monomorphisms hold for regular
epimorphisms.
Definition 2.1.7. An epimorphism / is extremal if whenever f = gh and g is mono, then g
is an isomorphism. Dually, a monomorphism / is extremal if whenever / = gh and h is epi,
then h is an isomorphism. For more detailed discussions, see Borceux [6, Volume 1, Section
4.3] and Herrlich and Strecker [13, Section 17].
12
Definition 2.1.8. The map / : A B is split mono if there exists g : B A such that
id = gf. The map g : B > A is split epi if there exists / : A > B such that id^ = gf. For
more information, see Herrlich and Strecker [13, Section 5]. Note that our terminology differs
slightly from the reference; section and split mono are synonyms, as are retraction
and split epi.
Lemma 2.1.9. For A either mono or epi, consider the following statements.
1. f is splitA.
2. f is regularA.
3. f is extremalA.
4. f is A.
The implications 1 => 2 => 3 => 4 always hold. None of the converses generally hold.
For proof see: (1 2) Herrlich and Strecker [13, 16.15], (2 => 3) Herrlich and Strecker
[13, 17.11] or Borceux [6, Volume 1, 4.3.3(1)], (3 => 4) holds by definition.
2.2 Bounds
Recall that a partial order on a set A is a relation < satisfying:
(pol) For all a 6 A, a < a.
(po2) Whenever a < b and b < c, a < c.
(po3) Whenever a < b and b < a, a = b.
A preorder is a relation that satisfies (pol) and (po2). If ^ is a preorder on A, define an
equivalence relation ~ on A by
a ~ b <==> a ^b and b < a.
The relation is a partial order on A/ ~. A set A with a partial order (resp. preorder) is
called a partially ordered set (resp. preordered set).
13
Definition and Remarks 2.2.1. Let A be a preordered set. Define the up and down
closures of x 6 A by
 x = {a A : a < x}
and
\ x = {a A : x < a}.
More generally, if 5 C A define
I 5 U{ i:iGS} = {ayl:3sGS,aXs}
and
 5 = U{f i:iGS} = {a6A:3iSS^ a}.
If 5 C A an upper (resp. lower) bound for 5 is a Â£ A such that for all s G 5, s < a (resp.
a < s). Use the following notation for the set of upper bounds of 5,
1(5) = n{f x : x 5} = {a G A : Vs 6 5, s < a}
and similar notation for the set of lower bounds
1(5) = n{j x : x 5} = (a G A : Vs G 5, a < s}.
If a G A, S C A and  a = 1(5), then a is a least upper bound, a.k.a. join, a.k.a.
supremum of 5. An equivalent way to say this is
a < x <=> V s 5, s < x.
If both a and a' are suprema of 5, then a,a' 1(5). Thus a ~ a'. In a partially ordered
set, the (unique) supremum of 5 is written \/ 5. If a A, S C A and j a 1(5), then a is
a greatest lower bound, a.k.a. meet, a.k.a. infimum of 5. An equivalent way to say a is a
least upper bound is
x < a V s 5, x < s.
14
In a partially ordered set, the (unique) infimum of 5 is denoted /\ 5. (In using this notion
with a preordered set (A, :<), one refers to the associated poset A/ For example, \J S
then denotes the equivalence class containing all suprema of 5.) If a = V 5 or a = f\ 5, a is
an optimum bound for 5.
The following are elementary properties of the boundoperators and optimum bounds.
Lemma 2.2.2. For any preordered set A (in particular any poset), the following properties
hold. Use IB to denote either operator or B.
1. For x,y A, x ^ y [ x C j y <==> ] y C] x.
2. For S,T C A, S CT => B(T) C 1(5).
3. For SC A, 1(5) = I( 5), 1(5) 1(T 5).
4 For (5j)j6/ a family of subsets of A, B(u5) = nB(5).
5. Iff: A^B is order preserving, then /(B,4(5)) C Bs(/(5)).
Definition and Remarks 2.2.3. Let A and B be preordered sets. A Galois connection
between A and B is a pair of functions / : A > B and g : B > A such that
(gel) / and g are order reversing.
(gc2) For all a A and b G B, a ^ g(f{a)) and b < f(g(b)).
Below, basic properties of Galois connections are listed. Symmetry in the definition
allows symmetry in proofs. For any true statement about Galois connections, then the
statement obtained by switching the roles of / and g, along with A and B is also a true
statement about Galois connections.
Note that / and g respect ~. The notation / ~ g means for all a A, f(a) ~ g(b)."
1 9 gfg Suppose a Â£ A. By (gc2), a r< f(g{a)) and g(a) < g(f(g(a))). Using
a ^ f(g{a)) and (gel), g(f(g(a))) ^ g(a).
15
2. f(A/ ~) is dually order isomorphic to {6 E B : b ~ fg(b)}: This follows from the
preceding statement. Since / ~ fgf, any member f(a) E f(A/ ~) is equivalent to
f(g(f(a))). Moreover, if b E B and b ~ f(g(b)), then b E f(A/ ~) because b ~ f(g(b)).
3. a < g{b) b < /(a): if a ^ g(b), then b < f(g(b)) < f{a). The converse is proved
similarly.
4. For any S C A, Bf(S) = ^_1B(5). Calculate
x E B(f(S)) Vsg5,iS/(s)
<=^ Vs S,s ^ g(x)
*=> g(x)eB(S)
x p_1(5)
5.If A and B are posets, the previous item implies /\/(5) = f(\/ S).
It is helpful to rephrase condition 3 for posets: ug(b) is the largest a with b < /(a).
In symbols,
9(b) = \/{a E A:b < /(a)}.
Basic information about Galois connections has been wellknown sincethe 1940s;
Raney [28] contains a bibliography of this early literature. The basic properties and def
inition are listed in Herrlich and Strecker [13, Exercise 27Q]. The particular summary here
is by the author.
The concept of a Galois connection is symmetric, and allows one to transfer a great
deal of information between preordered sets. However, the fact that the functions involved
are order reversing is sometimes inconvenient. The concept of an adjoint connection between
posets is obtained by formally reversing one of the posets involved.
16
Definition and Remarks 2.2.4. Suppose A and B are preordered sets. A pair / : A > B,
g : B > A of order preserving functions is an adjoint connection between A and B if
(adl) For all a 6 A, a < g(f(a)).
(ad2) For all b e B, f(g(b)) < b.
Each basic property for Galois connections corresponds to a basic property of adjoint
connections. The basic properties of adjoint connections are listed below; proofs are omitted.
1 / fgf and g ~ gfg
2. f(A/ ~) = {b G B : b ~ g(f{b))} and g{B/ ~) = {a 6 A : a ~ f(g(a))}.
3. /(a) ^ b <=> a X g(b)
4. For any SC A, Bg(S) /_1B(5) and B/(S) = g~1M(S)
5. If A and B are posets, f(/\ S) = f\ f(S) and g(\J S) = \/ g(S).
Because of the asymmetry between / and g, and property 3, / is called the left adjoint
and g the right adjoint. Again there is an interpretation of 3 in words, /(a) is the smallest
b such that a < g{b)\ g(b) is the largest a such that /(a) ^ 6.
The definition and basic properties of adjoint connections are folklore. For another
discussion of them see Johnstone [16, Chapter I, Paragraph 3].
Definition 2.2.5. A poset A is complete if each subset SC A (including the empty set)
has a supremum. Since the supremum of set of lower bounds for S is a lower bound for S,
A is complete if and only if each subset S C A (including the empty set) has an infimum.
There is a criterion for testing when a given order preserving (resp. reversing) map
between posets is part of an adjoint (resp. Galois) connection.
Theorem 2.2.6. Adjoint Existence Suppose A and B are posets.
17
1. Suppose f : A > B is order preserving and A is complete. Then f is a left adjoint if
and only if f(\J S) = \f f(S) for all S C A.
2. Suppose g : A > B is order preserving and A is complete. Then g is a right adjoint if
and only if g(f\ S) = /\ g(S) for all S C A.
3. Suppose f : A > B is order reversing and A is complete. Then f is part of a Galois
connection if and only if f(\J S) = f\ f(S) for all SC A.
This theorem is the poset version of the adjoint functor theorem. Folklore: see Johnstone
[16, Chapter I, Section 4, Paragraph 2] or Joyal and Tierney [17, Chapter 1, Section 1].
Proof. A proof for 1 follows; the other items are similar. Define g : B > A by
9P) = \J{aeA: f(a) < b}.
Since / preserves suprema, f(g(b)) < b. Thus, g(b) is the largest a such that /(a)
In adjoint (resp. Galois) connections, /(a) and g(b) can be defined as suprema or
intima. There is a sort of converse to this; one can view the supremum as an adjoint to a
particular map. See Lemma 5.1.1.
2.3 Natural Transformations
Maps compare objects in a category, functors compare categories, and natural trans
formations compare functors. This section contains no new results; results and some expo
sition are paraphrased from Herrlich and Strecker [13, Section 13].
Definition 2.3.1. Let F, G : A > 3 be functors. A natural transformation a : F > G is
a rule which assigns a map aA : FA > GA to each A A such that if / : A > B is an
Amap the following diagram commutes.
Ff
FA >FB
r\A
r]B
GA
Gf
GB
18
Construction 2.3.2. Let F, G, H : X > A be functors, a : F > G and f3 : G > H natural
transformations. (The situation is drawn in the diagram below.)
F
G
Q
H
' J
0
Then the assignment (/3a)A = (/3A)(aA) is a natural transformation.
Proof. Let / : A > B be a map. Since a and (3 are natural, each square below commutes.
FA
Ff
*FB
GA
Gf
GB
HA
Hf
HB
Therefore, the outside rectangle commutes; hence, (3a is natural.
Call /3a the vertical composition of a and (3.
Construction 2.3.3. If F : A > B and G, H : 3 G are functors and a : G > H is a
natural transformation, then (aF)A := a(FA) is a natural transformation aF : GF > HF.
Construction 2.3.4. If F : H > C and G, H : A H are functors and a : G > H is a
natural transformation, then (Fa)A := F(aA) is a natural transformation Fa : FG > FH.
Construction 2.3.5. Suppose F, G : A > B and H,J:T>*G are functors, and a : F *
G, (3 : H > J are natural transformations. (The situation is drawn below.)
F H
0 e
19
Then for each object A Obj(.A) the following square commutes.
FHA FPA > FJA
aHA
aJA
GHAG0A >GJA
The assignment (/3 a)A := (aJA)(F/3A) (Gf3A)(aHA) is a natural transformation.
Proof. The square commutes because a is a natural transformation; to see this, one applies
the natural property at the map (3A : HA + JA. To prove the natural property of 0a,
we use the squares which define 0a A and 0aB. Compare the corners of the squares using
maps obtained from / by application of the functors FH, GH, FJ, and GJ. The resulting
commutative cube shows /3a is a natural transformation.
Call 0 a the horizontal composition of a and 0.
Definition 2.3.6. A natural equivalence a : F G is a natural transformation such that
each component aA is an isomorphism. Given functors F and G are naturally equivalent if
there is a natural equivalence a : F G.
Categories A and B are equivalent if there are functors F : A > T and G : T> > A
such that FG is naturally equivalent to ids and GF is naturally equivalent to id^.
A and are dual if there exist contravariant functors F : A B and G : B > A
such that FG is naturally equivalent to ids and GF is naturally equivalent to id^.
2.4 (Co)Limits
For intuition, it is useful to view the objects of a category as a preordered class, with
the preorder A < B if and only if there is a map f : A B. Categories are more complex
than preordered classes, because there may be many maps / which manifest A < B.
Definition and Remarks 2.4.1. Let A be a category. A diagram D in A (a.k.a. a small
subcategory of A) is a set of objects Obj(D) C Obj(A) and maps Map(D) C Map(A)
20
between them; for technical reasons one requires that if f,g G Map(D) and fg is defined
then fg G Map(D), and that for all A G Obj(D), idA Map(D). This definition of a diagram
is equivalent to, but differs from the one in the majority of the literature; see MacLane [21],
Borceux [6], or Herrlich and Strecker [13] for the standard definition.
Suppose D is a diagram. A source (S, {s^ : S A : A G Obj(Â£))}) for D consists of
S G Obj(A) and maps sA : S > A such that if / : A > A' is a map in D, fsA = sA>. A source
for D is a lower bound compatible with all maps in D. A source (L, {Â£A : A G Obj(Z))})
for D is the limit of D if whenever (5, : A G Obj(D)}) is a source, there is a unique map
c : S L such that for each A G Obj(D), sA iAc.
Dually, define a sink, or cosource for D to be an object S together with maps iA :
A > S (for A G Obj(D)) such that for each f : A> A' in Map(D), iA = iA'f.
A colimit (C, {Â¡jla : A G Obj(D)}) for D is a sink for D, such that if (5, {m : A G
Obj(D)}) is any sink there is a unique map c : C > S such that for each A G Obj(D),
Ha = ciA.
Note that limits and colimits of D are unique up to compatible isomorphism. To
prove this (for limits), suppose (L,Â£A) and (L',Â£'A) are both limits for D. The limit property
guarantees that there are unique compatible maps c : L > L' and d : L' > L. But
cd : L' > L' and dc : L > L are both compatible maps. By uniqueness, cd = id'Â£ and
dc = idL. From now on, we shall use the definite article when writing about limits and
colimits.
Example 2.4.2. Let us consider Set. If X is a set of sets, form a diagram containing
all members of X and no functions. The limit for this diagram is the Cartesian product
(nx,7rx).
Recall that Â¡Q X contains all functions / : X > J X,with the feature that f(X) G X
for all X G X. Such / are called choice functions, because they choose one member of each
X G X. The functions nx 'WX X are defined by nx(f) = f(X).
21
If there is a source (S, {s* : S > X : X X}), there is a unique function c : S > f] ^
making
commute for each X. It is defined by c(a)(X) sx(a).
Motivated by this example, one defines the categorytheoretic product of a set X of
objects (in any category) as the limit of the diagram containing all members of X and no
maps. In most categories of sets with structure products (exist and) look like products in
Set, with suitable structure added.
In Set, the colimit of the diagram containing all members of X and no functions,
is the disjoint union ]J X, with inclusion maps px X i JJ X. The colimit property is
satisfied by (JJ X, px) because if there is a sink (C, ix), the function c : ]J X C defined
by c(x) = ix(x) for the unique X X containing x is the unique compatible map.
In categories other than Set, coproducts are defined identically. Usually the coprod
uct of a set X of ^.objects is the *Aobject freely generated by JJ X.
Example 2.4.3. The [co]equalizer of /, g : A > B is the [co]limit of the diagram containing
objects A and B along with maps / and g. The notation for equalizers is customarily
simplified by omitting the source map to B. Previously, i : E > A was defined to be the
equalizer if fi = gi and i factors through any other map which right identifies / and g. i
is the source map to A. The source map to B is redundant: it must be fi = gi : E B.
Similar notational economy is applied to coequalizers.
22
Example 2.4.4. Let us consider the diagram
A
f
B *C.
(The diagram also contains identities for all objects, but for brevity these are omitted.) The
limit (B Xc A, nA, ttb) of this diagram is called the pullback of / along g, or the pullback of g
along /. To be explicit, fnA = nBg and if (Q,qA : Q A,qB : Q > B) satisfies fqA = qBg,
then there is a unique map i : Q > B Xp A.
In Set,
B xc A = {(b, c) e B x A: g(b) = /(c)},
and the projection maps n are the restrictions of the projections from the cartesian product.
Pullbacks in any concrete category A equipped with a limit preserving faithful functor
U : A Set are computed identically. Two particular instances of pullbacks deserve
special attention.
First, let / : A C be any map and g : S > C be a subset inclusion of S C C.
Note that (s, a) S xc A if and only if g(s) = /(a); suppressing mention of g, this reads
S xc A {(s,6) : f(b) = s}. Thus, this pullback is canonically isomorphic to the preimage
of S under /. This example partially motivates the name pullback.
Second, let / : A > B be any map and consider the pullback of the diagram,
A
f
A *B
i.e., the pullback of / along itself. Using the computation for pullbacks in Set, given above,
A xB A {(a, a') A x A : f(a) = /(a')}.
23
This relation on A is often called the kernel of /. Thus, one calls (A x# A,nleft,nright) the
kernel pair of /. More indepth discussion of pullbacks is given in Borceux [6, Volume 1,
Section 2.5] and Herrlich and Strecker [13, Section 21].
Definition and Remarks 2.4.5. A category is said to be complete if each diagram has a
limit; it is said to be cocomplete if each diagram has a colimit.
Unlike the situation for posets, a category may be complete without being cocomplete
and vice versa. (See Herrlich and Strecher [13, Section 23, pl61ff] for a detailed discussion
of this and related issues.)
Consider a functor F : A > B. If D is a diagram in A, there is a diagram FD with
Obj (FD) = {FA : A e Obj(D)} and Map (FD) = {Ff : f G Map(D)}.
Let D be a diagram in A. A functor F : A > B preserves the limit of D, if whenever
the limit (L,Â£a) exists in A, (DL, D(iA)) is the limit of FD. F preserves limits if for any
diagram D, F preserves the limit of D.
If G : A B is contravariant, D is a diagram in A, and (L, {Â£A : A G Obj(D)}) is
the limit of D, then (GL, {GiA : GA > GL}) is a sink for GD. If (GL, {GÂ£A : GA GL})
is the colimit for GD, then G takes limits to colimits. Analogously, may G take colimits to
limits.
2.5 Adjoint Functors
There are several useful concepts of adjoint connections between categories. There is
a strong analogy between preordered sets and categories; any category may be preordered
by
A^B <=> 3/ : A > B.
Categories are more complex, because many maps could manifest A A B. We begin with
the concept analogous to 2.2.3.
Definition and Remarks 2.5.1. A (functorial) Galois connection between categories A
and 'B consists of contravariant functors T : A > G : B > A, together with natural
24
transformations
77 : id* GF
and
e:W3 > FG
such that (Fr]A)(eFA) = id^ and (GeB)(r]GB) = ides for each A 6 Obj(A) and B
Obj(33); these equations are the socalled triangle identities. (Alternate terminology: if
(F,G,r],e) is a functorial Galois connection, then F and G are adjoint on the right.)
1. The categories fix(rÂ¡) and fix(e) containing all objects such that r]A (resp eB) is an
isomorphism are dual. [3, Section 4, Lemma 1]
2. There is a natural bijection A(A, GB) (5, FA) given by
(f : A GB) H> F(f)(eB) : B > FA.
The inverse map is
(g : B > FA) >> G(f)(vA) : A > GB.
There are two (identical) calculations required to check that the functions are mutually
inverse. One is summarized by the diagram below.
GFA (GFf) > GFGB GB
The square commutes because 77 is natural. The triangle commutes because of the
identity ides = (GeB)(rÂ¡GB). The reader may formulate and check what is meant by
naturality of the bijection.
3.F and G both take colimits to limits.
25
4. Each rjA has the following universal property: if / : A GB, there is a unique map
/ = F(f)(eB) : B EM which makes the following diagram commute:
A GFA
Each eB has the analogous universal property.
Much of the literature just deals with adjoint connections, where both functors are covariant.
By duality, any such result can be translated in terms of Galois connections. Banaschewski
and Bruns [3] includes a reasonably thorough expository section on functors which are adjoint
on the right.
The (functorial) Galois connection concept is symmetric, but the contravariance of
the functors involved is sometimes awkward. An analogous asymmetric concept, with the
functors both covariant is described below.
Definition and Remarks 2.5.2. A (functorial) adjoint connection between categories A
and 3 consists of functors F : A > 3 and G : 3 A and natural transformations
r] : idyi GF and e : FG > id such that (GeB)(rÂ¡GB) = ides and (eFA)(FrÂ¡A) = id^;
these equations are the socalled triangle identities. In this situation, one also says F and
G are adjoint functors, F is the left adjoint and G is the right adjoint. The basic
properties of functorial adjoint connections closely correspond to the basic properties of
functorial Galois connections.
1. The categories fix(r/) and fix(e) containing all objects such that 7/.4 (resp eB) is an
isomorphism are equivalent.
2. There is a natural bijection A(A, GB) 3(FA, B) given by
(f : A* GB) t> (eB)F(f) : FA > B.
26
The inverse map is
(g : B FA) G(/)foi4) :A^GB.
3. F preserves colimits; G preserves limits. Borceux [6, Volume 1, 3.2.2]
4. Each rjA has the following universal property: if / : A > GB, there is a unique map
/ = (eB)F(f) : FA B which makes the following diagram commute:
A >GFA
Each eB has the analogous universal property: if / : FA > B, there is a unique
/ = G(f)(rjA) : A > GB such that
commutes. Any functor for which there is a natural transformation rj with the above
universal property is part of an adjoint connection.
For discussions of adjoint functors, see MacLane [21, Chapter IV], Herrlich and Strecker [13,
Sections 26, 27, 28], or Borceux [6, Volume 1, Chapter 3].
There is a criterion for determining when functors are adjoints, which corresponds to
Theorem 2.2.6.
Theorem 2.5.3. (Adjoint Functor Theorem, [21, V.6.2],) Let A be a complete category.
A functor G : A > B has a left adjoint if and only if
1. G preserves limits, and
27
2. (solution set condition) for each B 3 there is a set I and an Iindexed family of
maps fi'.B> GAi such that any map f : B > GA can be written as h = (Gt)fi for
some i 6 I and t : AÂ¡ > A.
CHAPTER 3
ALGEBRAS OF A MONAD
Monads (a.k.a. triples, a.k.a. standard constructions) and their (EilenbergMoore)
algebras provide a concise formulation of many important categorical aspects of universal
algebra. The category of algebras for a monad has special properties, which are summarized
in Section 3.1; a category of algebras is often complete and cocomplete, and always has
free objects. Each (functorial) adjoint connection induces a monad; the correspondence
between adjunctions and monads is discussed in Section 3.2. In Section 3.3, the question of
when an adjoint connection connects A to a category of algebras is addressed. In Section
3.4, the question when does the composite of two monadic adjunctions yield a monadic
adjunction? is discussed.
The results in this chapter are reasonably well known. MacLane [21, Chapter VI],
Borceux [6, Volume 2, Chapter 4], Barr and Wells [4, Chapters 3 and 9], Manes [23] and
the introduction to the seminar notes [1] contain good expositions of monads from various
perspectives.
3.1 Categories of Algebras
Definition 3.1.1. A monad T = (T, rj, /i) on A consists of a functor
T:A^A,
a natural transformation
V i T,
and a natural transformation
/i: T2 > T,
28
29
such that the following identities (expressed by commutative diagrams) hold: the unit laws
T2
(In these diagrams, Tn denotes the nfold composite of T with itself.)
Intuitively, TA is the free object on A; r]A is the insertion of variables map; fjA
is the semantic composition, i.e., a map which allows one to view a polynomial with
polynomial variables as a polynomial. See Example 3.2.2 for a concrete illustration of the
roles of T, 77 and Â¡1.
Definition and Remarks 3.1.2. Let T be a monad on A. A Talgebra (A, a) consists of
A Obj(A) and a : TA > A (the socalled structure map) such that a(r)A) kU (unit
law) and
T2A *TA
M a
TA
30
(associative law) commutes. A homomorphism / : (A, a) > (B, b) of Talgebras is an Amap
/ such that
Tf
TA *TB
a b
commutes. The category of all Talgebras and Talgebra homomorphisms is denoted AT.
There is a forgetful functor UT : AT + A; it is defined by UT(A,a) = A and
UT(f) = f. UT has a left adjoint FT : A AT defined by
Ft(A) = (TA,pA)
and
FT(f) = Tf.
The associated natural transformations are
r]T = r,: id^ UtFt = T
and
eT : FtUt id^r : eT(A, a) := a.
[21, VI.2.Theorem 1]
Limits in Ar are computed in A in the following sense.
Proposition 3.1.3. [4, 3.3.4], [6, Volume 2, 4.3.1] Suppose (T,r/,/x) is a monad on A. If D
is a diagram in Ar such that UT(D) has a limit (L,pu(A)), then there is a unique structure
map I : TL L such that each pu(A) a Talgebra homomorphism.
Proof. Let D be a diagram in Ar, such that UTD has a limit (L, IA). For each A G Obj(T),
name the structure map sA : TA + A. The goal is to produce a structure map s :TL > L
such that each tA : L A is a Talgebra map.
31
The requirement that each is Talgebra map amounts to: for each A, the diagram
below commutes.
FL
SA
Since (L,Â£a) is a source for UD, (FL, sa(FIa)) is a source for UD. Thus, there is a unique
A map s : FL > L making each diagram above commute.
Checking the unit law, s(iL) = id: the diagram
commutes using the definition of s, and that i is a natural transformation. Thus, for each
A, a as(L). Prom the uniqueness of the map from the limit of D to any other source
for D, it follows that id = s(iL).
A similar comparision of squares diagram can be used to verify that the algebra
associative law holds.
Corollary 3.1.4. If A is complete and T is a monad on A, then AT is complete.
If A is cocomplete, AT is often also cocomplete. The following theorem was originally
proved with fewer hypotheses in Linton [20], Other expositions are given in Borceux [6,
Volume 2, 4.3.4] and Barr and Wells [4, Section 9.3].
Theorem 3.1.5. Let A be cocomplete and T be a monad on T. AT is cocomplete if and
only if AJ has coequalizers.
32
3.2 Adjoint Connections induce Monads
The following proposition gives a correspondence between monads and adjoint con
nections. It should be emphasized that the correspondence is not bijective. Each monad
gives rise to a unique adjoint connection, but in general many adjunctions induce the same
monad.
Proposition 3.2.1. Correspondence between monads and adjoint connections
1. Let F : A > 3, G : CB > A, rÂ¡ : kU > GF, e : FG > id be an adjoint connection.
Then (GF,rÂ¡,GeF) is a monad on A.
2. If T = (T, T], /i) is a monad, then
GtFt = T,
T
T1 =1h
and
GTeTFT = [i.
The proof of the preceding Proposition consists of verifying identities: the triangle
identities for adjoint connections imply the unit laws; the associative law holds because
the square defining horizontal compositions commutes. Details are given in MacLane [21,
VI. 2. Theorem 1],
This correspondence allows construction of many examples of monads. Often, but not
always, a naturally occurring adjoint connection corresponds to the category of algebras
over the induced monads.
Example 3.2.2. Consider the category Grp of groups. The forgetful functor Ucrp Grp
Set and the free group functor FgtP : Grp Set form an adjoint connection between Grp
and Set. Recall that ForpiX) is the set of all reduced words
33
where x 6 X, s 6 Z; a word is reduced if for all i with 2 < i < n, x,_i / x. The operation
on TGrp(A') is concatenation of words, followed by reduction. T = ^Grp^Grp is the functor
part of the induced monad; 77 : idset > T is the natural transformation whose component at
X sends x X to word x1; if w*1 u>n 6 T2(X) and for each z, = x^1 then
(ma)K' ...<) = (*5y~*Ksri
_6i,m(m
x
n.m(n)\a,
n,m(n)
Each group G is an Talgebra; the group multiplication and inversion give a map from the free
group on the underlying set of G to G. Conversely, each Talgebra structure gives a group
multiplication and inversion. Grp is equivalent to the category of Setr. (This example is
typical in the sense that any category of finitary algebras in the sense of universal algebra
 is also a category of monad algebras. The parts T, rj, and Â¡1 of a monad generally have the
same roles as in this example.)
Example 3.2.3. Some (nontrivial) adjunctions involving Set induce the trivial monad.
For example, consider the category Top of topological spaces and continuous maps. The
forgetful functor
i/rbp : Top > Set
has a left adjoint
Ftop : Set > Top,
which sends a set X to the discrete topological space with underlying set X. One easily
checks that T = GTopFTop is the identity functor on Set, and that all associated natural
transformations are identities.
Definition 3.2.4. Suppose F : A > G:B* A, rj : id^ 7 GF, and e : FG > id
forms an adjoint connection. By Proposition 3.2.1, the adjoint connection induces a monad
(T = GF, 77, GfiF). If B is equivalent to AT, then we say the adjoint connection (sometimes
just the right adjoint G) is monadic. Often a concrete category B has a canonical forgetful
functor G : B > A\ in this case, one may even say B is monadic over A omitting mention
of G.
34
These examples illustrate the qualitatively different behaviors of monadic and non
monadic adjoint connections. Monadic categories are determined by the combinatorial struc
ture of the free algebra functor FT. Most relational categories like Top, P, the category
of graphs, etc. have free functors, but these free functors do not add any structure to the
underlying set; they merely attach the most discrete possible relation to the given set.
3.3 Detecting Categories of Algebras
Because of the special properties of AT, the question of when an adjoint connection
is monadic has great practical importance.
Remark 3.3.1. Let (A, a) be a Talgebra. Because of the associative and unit laws, and
because 77 is natural, the following equations hold:
(nat) (Ta)(r[TA) (rp4)a,
(unitl) (jA)(tÂ¡TA) = idTA,
(unit2) a(r)A) = id, and
(assoc) a(Ta) = a(fiA).
These equations imply that a coeq(Â¡iA,Ta)\ for if b : TA B rightidentifies Â¡1A and Ta,
then b(r]A) : A B and
b(rÂ¡A)a = b(Ta){T]TA) by (nat)
= b(fiA)(r]TA) since b right identifies
b by (unitl).
One verifies that / = b(rÂ¡A) the unique map / : A > B with fa = 6; for if fa = b, then
/ = fa(rjA) = b{rÂ¡A).
These equations give a great deal of information about a. Thus, the key hypothesis
in Becks theorem the criterion for when an adjunction is monadic is the preservation
and reflection coequalizers obeying the equations described above.
Using the above described equations for motivation we offer the following definitions.
35
Definition and Remarks 3.3.2. Consider maps f,g:A>B.
1. If e : B + C has the feature that for any functor F, Fe = coeq(Ff, Fg) then we say e
is an absolute coequalizer.
2. If there are maps e : B > C, sc C > B and sb : B > A such that
fsB
kU = gsB,
idc = esc,
ef = eg,
then we say e is a split coequalizer.
Note that for every Talgebra {A, a), a is a split coequalizer of Ta and pA. It is also easy to
see that every split coequalizer is absolute.
The following theorem is due to Beck (unpublished). Linton [19] contains a detailed
discussion of variations on the theorem. The theorem was originally phrased in terms of
split coequalizers only; Pare [26] refined the theorem to include the absolute coequalizer
condition. Many variations on the hypotheses exist; the version stated here is found in
MacLane [21, VI.7.1], Other expositions of the theorem may be found in Barr and Wells [4,
Section 3.3] and Borceux [6, Volume 2, Section 4.4].
Theorem 3.3.3. Let (F : A > 25, G : 3 > A, r?, e) be an adjoint connection, and T = GF,
rj = g, and g = GeF be the associated monad. The following conditions are equivalent:
1. The adjunction (F,G,r],e) is monadic.
2. If f, g : A B Map(!B) and the pair (Gf,Gg) has an absolute coequalizer e!, then
e = coeq(f,g) exists and Ge e'.
36
3. If f,g : A > B e Map(!B) and the pair (Gf,Gg) has a split coequalizer e', then
e = coeq(/, g) exists and Ge = e'.
Recognizing monadic adjunctions is complex because compositions of monadic func
tors are not generally monadic.
Example 3.3.4. The category Ab of abelian groups is monadic over Set, and tfAb (see
Example 2.1.3) is monadic over Ab, because any reflection is monadic. Each free abelian
group is torsion free, so the monad on Set induced by the free abelian group functor is the
same as the monad induced by the free torsion free abelian group functor. Thus, if tfAb
were monadic over set, then tfAb and Ab would be equivalent categories both equivalent
to a suitable category of monad algebras. The categories tfAb and Ab are obviously not
equivalent. (The preceding example is summarized from Borceux [6, Volume 2, Example
4.6.4].)
If G : A Set, the hypotheses of Becks theorem can be reformulated to make it
easier to check whether G is monadic.
Theorem 3.3.5. [11, Theorem 4.2] Let G : A > Set be a functor with a left adjoint.
Suppose that A is complete and has coequalizers. The following are equivalent:
1. G is monadic.
2. G satisfies the following conditions
G preserves and reflects regular epimorphisms;
if f : GA > X is an isomorphism, then there is a unique map g : A > B such
that Gg f;
G reflects kernel pairs.
The following lemma provides useful information concerning when a functor between
two categories of algebras is monadic.
37
Lemma 3.3.6. [6, Volume 2, Corollary 4.5.7] Suppose U : 23 > A, V : G > A, and
Q : B > C are functors. IfU VQ, U and V are monadic, and 23 has coequalizers, then Q
is monadic. In particular, Q has a left adjoint.
3.4 Distributive Laws
Compositions of monadic adjunctions are not generally monadic; when a composition
of monadic adjunctions is monadic, it indicates a distributive law between the two structures.
The following section summarizes the results later needed from Becks [5]; this exposition
follows Becks notation, except that composition of functions here will read righttoleft.
For the duration of the section, assume there are two monads T = (T, t]t, pT) and S =
(S, r]s, ps) over some base category A. For any monad, we use F and U, with superscripts for
the name of the monad, to denote the free algebra functor and forgetful functor associated
with a monad, respectively. See 3.1.2 for the definitions of F and U.
This material is rather abstract, so it helps to have an example in mind: after each
definition and theorem, we will illustrate what it means using S the free monoid monad
over Set and T the free abelian group monad over Set; the composite monoid TS gives
the free ring. For a set X, SX consists of all strings from X, with concatenation as the
operation: 5 acts on functions by
Sf(xix2 xn) = f(x1)f(x2) f(xn).
TX consists of all formal (finite) linear combinations of elements of X, with integer coeffi
cients. T acts on functions in the expected way.
Despite the concrete illustrations in terms of these monads, all theorems and defintions
will apply to any monads S and T.
Definition and Remarks 3.4.1. A distributive law of S over T is a natural transformation
Â£ : ST>TS
38
satisfying the compatibility conditions:
T S
If there is a distributive law Â£ : ST TS of S over T, then the composite monad is TS =
(TS,r]TS := r]TT]s,pTS := pT ps(TÂ£S)). In the definitions of r)TS and pTS, juxtaposition
denotes horizontal composition of natural transformations. The verification that TS actually
defines a monad is omitted; for more information see Beck [5],
In the case when S=free monoid and T=free abelian group, the natural trans
formation Â£ : ST > TS expresses a productofsums as a sumofproducts in the usual
way:
n 5^ aik n ac(k),k
keK iei c keK
where c ranges over all choice functions c : K > I. It is a somewhat enlightening exercise
to check that defined this way is a natural transformation ST TS.
A distributive law may of T over S also be viewed as a way of lifting T to a monad
over S.
39
Definition 3.4.2. T has a lifting into ./Is, if there is a monad T such that TUS = UST,
UsrjT r]TUs and Usp^ = pTUs. To sketch the situation, T is a lifting of T onto if
commutes.
Theorem 3.4.3. There is a bijective correspondence between liftings of T to /Is and dis
tributive laws of S over T.
The proof is outlined in Beck [5]. The correspondence is defined as follows: if i is a
distributive law, and (A, a) is an 5algebra, then T is defined as follows:
T(A,a) = (TA, (Ta){iA)),
r)T{A, a) = t]TA : (A, a) + T(A, a),
and
fiT(A, a) = pTA : TT(A, a) T(A, a).
Consider the following diagram:
the square commutes because rÂ¡T is a natural transformation; the triangle commutes because
l is compatible with rf. Since the perimeter of the diagram commutes, rf A is an 5algebra
40
map, A similar diagram, which uses the compatibility between I and pT, shows pF is an
5algebra map. Since the underlying maps are defined in A and obey the monad laws, tjt
and pF also obey the monad laws. If T is a lifting of T over As, I is defined to be the
following composition:
ST STS = STUsFs = USFSUSTFS uSfFfs USTFS = ST.
(In the above equation, es is the counit of the adjunction (Fs, Us, r/5, es); es(A. a) = a.)
After some detailed computation, one verfies that this is a distributive law, and that the
correspondences described are mutually inverse.
Corollary 3.4.4. If T has a lifting to As, then there is a composite monad TS.
Theorem 3.4.5. Suppose I: ST > TS is a distributive law. The categories ATS and (AS)T
are equivalent.
Let us consider this in more detail. An object in (AS)T consists of an Aobject,
A, along with an 5structure as : SA > A, Tstructure ax : TA + A, and 5structure
t: ST A > A for TA such that the following diagram commutes.
SciT
ST A 1 SA
t aS
TA >A
Given a T5algbra {A, a), A has a Tstructure aT map defined by
aT = a(Tr]SA) :TA>A
and an 5structure map as defined by
as = a(r]TSA) : SA > A.
41
The following diagram which expresses the distributivity of the structure maps commutes.
Thus we have a map ATS > (AS)T. given by,
(A, a) e ATS i (A, ar, as,Tas(Â£A))
which is functorial, because both rj1' and rf are natural.
The inverse functor {AS)T > ATS, maps (A,ar,as,t) to asS(ar) art. The verifi
cations for these assertions is given in Beck [5],
CHAPTER 4
GENERATING SUBMONADS
In this chapter, a technique for creating monads is laid out; before proceeding formally,
let us consider a rough outline of the technique. Suppose (T, r], /r) is a monad on a reasonable
category, and that / : F T is a subfunctor of T. The goal is to extend F to a submonad
of T.
Intuitively speaking, F is a natural collection of polynomials. One should therefore
require that for all A, (r/A)(A) C F(A) i.e., that for each A and a G A, FA contains the
constant polynomial with value a. F is not generally closed under composition; i.e., it
is not generally true that (^,A)(F2A) C A. To correct this problem, we start with F and
iteratively add polynomials obtained by composing members of FA.
The purpose of the chapter is to formalize the preceding vague outline. The author
abstracted and clarified [25, Proposition 3.4], which may be viewed as a special case of this
result. Thus, the main results of this chapter are called Meseguers Lemmas 4.2.1, 4.2.2,
4.2.6, and 4.2.7 to acknowledge the analogy which prompted the technique. The author
believes the general formulation of the technique is new.
Section 4.1 details requirements on a reasonable category, explains precisely what
is meant by subfunctor, and gives methods for constructing subfunctors. Section 4.2 gives
a proof of Meseguers Lemmas. Section 4.3 describes an example showing the necessity of a
technical hypothesis of the lemmas.
4.1 Subfunctors
This section describes a general theory of subfunctors. A subfunctor of F : X > Set
is a natural transformation r] : E + F such that each component is a subset inclusion. The
42
43
concept subfunctor is quite useful, but in some categories such as P, Top, and Locale
(explained in Example 4.1.2), there is either
no obvious meaning for subset inclusion (in Locale), or
more than one possible structure on each subset (in P and Top).
The following section expounds a theory in which subfunctor means natural transforma
tion whose components are all extremal mono.
Axioms 4.1.1. Throughout, the base category A is assumed to have the following features:
1.A is complete.
2.A is an (epi, extremal mono) category. Recall that an (epi, extremal mono) category
is a category with the features that
For each map /, it is possible to factor f as f me where e is epi and m is
extremal mono. This factorization is unique, in the sense that if / = m'e' is
another way of writing / as an epi followed by an extremal mono, there is an
isomorphism i such that
e
commutes.
if a and b are extremal mono and ab is defined, then ab is extremal mono.
3.A is extremallywellpowered. This means: for each A, there is a set
{/a S\> A}
of extremal monomorphisms, such that if m : T A is any extremal monomorphism,
there is an index A and isomorphism i such that f\ mi.
44
If A has these features, then it is an SFcategory. (SF stands for subfunctor.)
Example 4.1.1. By Herrlich and Strecker [13, 34.5], any wellpowered category with in
tersections and equalizers is an (epi, extremal mono) category. So if A is complete and
wellpowered then A is an SFcategory.
It follows that P, Set, Top and practically any reasonable category of topological
spaces, is an SFcategory.
Example 4.1.2. The category of frames i.e., complete lattices in which
a a\J S = \J {a A s : s E S}
holds for all elements a and subsets S is complete and cocomplete. This category also has
(regular epi, mono) factorizations. Thus, Locale, the dual category to the category of frames,
is an SFcategory. Locale is extremallywellpowered, but not wellpowered. The authors
interest in locales motivated him to use the given definition (which only requires extremal
wellpoweredness) for SFcategory rather than defining SFcategory to mean complete and
wellpowered, so that the theory of subfunctors would apply to localic subfunctors in addition
to the previous examples. (For background on frames and their relation to pointset topology,
see Isbell [15], Johnstone [16], Joyal and Tierney[17] and Madden [22].)
Lemma 4.1.3. Assume that A is an (epi, extremal mono) category. Then
1. (diagonalization) [13, 33.3] If ge = mf, where e is epi and m is extremal mono, then
there exists k such that mk = g and ke = f.
2. [13, 34.2(2)] Any intersection of extremal subobjects is extremal.
3.[13, 34.2(3)] Pullbacks of extremal monomorphisms are extremal mono.
Definition and Remarks 4.1.4. Defining Sub(A) the lattice of extremal subob
jects: In an SFcategory A, the lattice of (equivalence classes) of extremal subobjects has
particularly nice structural features.
45
An extremal subobject of A is an extremal monomorphism s : S > A. Define a
preorder on the class of extremal subobjects by
(s : S A) C (t : T + A) 4= 3c : S T, s tc.
For brevity, one often only mentions the map or object part of an extremal subobject. To
avoid confusion, the same letter will be used to denote both parts, with the lower case letter
used for the function.
Since t is mono, there is at most one map c which manifests s C t. If s C t and t C s,
then there are cx and c2 such that s tc\ and t sc2. So t = fc!c2 and s = sc2cj. Since
c2ci shows s C s, c2Ci = idg. Similarly, cic2 = idy. One identifies extremal subobjects s and
t if s C t and f C s, or equivalently, when there is an isomorphism c such that s tc. The
set of equi valance classes under this relation is denoted Sub (A).
Any map c, which exhibits s C t, is extremal mono. For if ca cb, then sa =
tea = teb = sb, so (since s is mono) a = 6; so c is mono. If c = me, where e is epi, then
s = tc = tme. Since s is extremal mono, e must be an isomorphism. Thus c is extremal
mono.
Recall that the intersection of a set M of monomorphisms with a common codomain
is the limit of the diagram generated by M. Since A is complete and intersections of extremal
monomorphisms are extremal, Sub(A) is a complete lattice, with meet operation f) and
join operation J. In general, joins in Sub(A) are not disjoint unions; in general, (JA is
computed using
(jAi = f){A,:Vi, ACA'}.
(See Borceux [6, Volume 1, 4.2.2, 4.2.3, 4.2.4].)
A map f : A B induces an adjoint connection between Sub(A) and Sub(B). (See
Borceux [6, Volume 1, 4.4.6]; Borceuxs results are phrased in terms of strong monomor
phisms. Under our assumptions a map is strong mono if and only if it is extremal mono.)
If s : S A Sub(A), use the (epi, extremal mono) factorization to obtain a unique
46
s' 6 Sub(Â£?) such that
A
f
B
S
S *f+1(S)
commutes. The notation /+1(S) is used for the image of S, to remind the reader of the
analogy to ordinary set theoretic images of subsets under maps. The map (s : S > A) i
(s' : S' = f+1(S) > B) is the left adjoint. The right adjoint f~l : Sub(S) > Sub(A) is
obtained by taking the pullback of s : S > B along /. Thus,
f+1{S) CT ^ SCf\T),
/+l(Us<)=lbw(s<>'
and
r1(f]si) = f]r1(si).
Definition and Remarks 4.1.5. Let F : X A be a functor. A subfunctor E of F is a
rule that selects an extremal subobject eX : EX > FX for each X G Obj(X) such that if
f : X >Y, then
(Ff)+1(EX) C EY.
Any such assignment E of F gives rise to a functor E : X > A. Ef : EX > EY is defined
to be the composition EX (Ff)+1(EX) C EY.
Since eY : EY > FY is mono, Ef is uniquely determined by the condition that
Ff
FX >FY
eX
eY
EX *EY
commutes. Thus, (Ff)+1(EX) C EY implies that Ef can be defined to make the above
square commute.
47
The converse also holds: if there is a map Ef such that the square commutes, then
(Ff)+1(EX) C EY. To prove this, suppose there is a map Ef which makes the square
commute. Use the unique factorization Ef = ab where a : A EY is extremal mono and b
is epi. Note that (Ff)(eX) = (eY)(Ef) (eY)ab gives a factorization of (Ff)(eX) into an
epi b followed by an extremal mono (eY)a. Thus,
[{eY)a : A > FT] [(Ff)+1(EX) * FT]
and (Ff)+1(EX) C EY. Thus, a subfunctor is exactly a natural transformation whose
components are all extremal monomorphisms.
Let Subfun(F) denote the class of subfuctors of F : X > A. Define a preorder on
Subfun(F) by
E\ C F2 <=> W1 Obj(.A), Ei{A) C E2(A).
If Ei C E2, then, for each A, there is a unique extremal monomorphism cA such that
e\A (e2A)(cA). One easily verifies that cA are the components of a natural transformation
c : E\ > E2; in fact, c : E\ > E2 is a subfunctor.
The following constructions show that the class Subfun(F) behaves very much like
Sub(j4).
Construction 4.1.6. Subfun(F) is complete. If {Ei : i 1} is any class of sub functors,
there is a supremum (JF and infimum P) Et. The supremum is given by (IJEi)(A)
UFj(j4). The infimum is given by (PF)(A) = P Ei(A).
Proof. Since A is extremally wellpowered, for each A, the class {EfA) : i 1} has a
representative set. Thus, the objectbyobject definitions for J F, and fj Ex make sense. It
is obvious that if U F and p) F are subfunctors that they are the optimum bounds in the
subfunctor lattice.
Let f : X *Y be any map. A h> P EX(A) is a subfunctor because for each i I,
(F/)+1(P E(A)) C (Ff)+\Ei(A)) C Ei(B),
48
so (F/nai^c a Â£(Â£?).
>11 Ei(A) is a subfunctor because,
(F/)+1(LJf(A)) = U(Ff)+l(EfiA)) C U^(5).
i i
a
Construction 4.1.7. Suppose a : F + G is a natural transformation. Then there is an
(order theoretic) adjoint conection
a+1 : Subfun(F) > Subfun(G) : a1.
The left adjoint a+1 is defined by (a+1E)(A) = (aA)+1(EA). The right adjoint a1 is defined
by (a~1E)(A) = a~1(EA).
Proof The order on Subfun(F) is defined objectbyobject, and whenever
f : A B,
f+1 and f~l form an adjoint connection between Sub(.4) and Sub(F). So it suffices to check
that a+1E and a~lE define subfunctors.
Let E Subfun(F), / : X > Y be a map, and consider the following diagram.
(aA)+1(EA) (aB)+1(EB)
The trapezoids (i) and (iii) are obtained by factoring (aB)(eB) and (aA)(eA), respectively;
in each case 2 is the epi part and e' is the extremal mono part. The square (iv) expresses
the naturality of a. The trapezoid (ii) expresses that e : E F is a subfunctor.
To show a+1E is a subfunctor of G, it suffices to show there is a map
k : (aA)+\EA) * (aB)+1(EB)
that makes the top trapezoid commute. For this, use the diagonalization property from
Lemma 4.1.3. Define / = zB(Ef), g = (Gf)e'A, e = zA and m = e'B; note that ge mf
with e epi and m extremal mono. Thus, the diagonalization property guarantees the desired
k exists.
Assume / : A B. A similar diagram is used to verify that a_1E is a subfunctor
of F whenever E e Subfun(G). The missing map (aA) 1{EA) > (aB) X{EB) is obtained
from the pullback property of (aB)~1(EB).
50
Corollary 4.1.8. Let a : F > G be a natural transformation, and e : E > F be a subfunc
tor of F. Let e' : a+1E > F be the subfunctor described in Construction 4.1.7. There is a
natural transformation z : a+1E > F such that for each object A, (aA) (e'A)(zA) is the
(epi, extremal mono)factorization of aA.
Proof. One defines e' and 2 as in the preceding proof. Examining the comparison of squares
diagram used to produce a+1E/, shows that 2 and e! are natural.
Construction 4.1.9. Assume E,F,G,H are functors A > A. Suppose that
1. a : E > F and (3 : G > H are sub functors,
2. Either E or F preserves extremal monomorphisms.
then the horizontal composition (3a : EG FH is a subfunctor.
Proof. The horizontal composition of natural transformations is always natural. Hypothesis
2 implies that, for any A, (3aA is an extremal monomorphism, because the class of extremal
monomorphisms is closed under composition and
((3a)A := (aHA)(E(3A) = (F0A)(aGA),
by definition of horizontal composition (see Construction 2.3.5).
4.2 Meseguers Lemmas
Lemma 4.2.1. Suppose A is an SFcategory, {T,p,rf) is a monad on A, and F is a sub
functor ofT, such that (r]A)+1{A) C FA. There is a smallest subfunctor F of T such that
F C F C T and for all A Obj(yi)
(pA)(F2A) C FA.
If the equation
(fj,A)(G2A) C GA
holds for G, we say G is closed under p.
51
Proof. Let T denote the class of all subfunctors of T, that are larger than F and closed
under p. S' is nonempty because it contains T. By Construction 4.1.6, F = PT exists.
Since S (pA)(S) is an order preserving map Sub(T2v4) Sub(2L4), so for each A
(pA)(F2A) C FA.
The preceding gives an easy candidate for the functor part of the monad generated
by F. One needs a more complex argument if one wants detailed information about the
natural transformations related to rÂ¡ and /i which make F into a monad.
The proofs of the following lemmas require detailed computation. For clarity, the
goal of each paragraph in the proof is written in boldface.
Lemma 4.2.2. (Assume notation and hypotheses from Lemma If.2.1.) There are natural
transformations : id^ F and m : F2 > F which make (F, n, m) a monad.
Proof. One defines four sequences of natural transformations:
1. f\ : F\ > T the subfunctor generated at stage A,
2. n\ : id^ F\ a natural transformation obtained from rj by suitably modifying the
domain and codomain,
3. m\ : Ff > F\+1 a natural transformation obtained from p by suitably modifying
domain and codomain, and
4. c\ : F\ > F\+i the inclusion.
Define F0 = F; use the notation f0 : F0 T. m0 and Co are defined according to the same
pattern that defines later ms and cs, which is described below.
52
Defining fX'.Fx>T (A > 0), mA : Fx* * Fx+i, and cx : Fx > Fx+i. Assume
that fx has been previously defined. Consider the following diagram:
(M)+1(*?) U FXA
Given fx and Â¡iA, the square is obtained by (epi, extremal mono)factorization of the com
posite ([iA)(fxA). All maps in the triangle are extremal monomorphisms, obtained by
comparing subobjects of TA. Define
Fx+1A = (fxA)+\FÂ¡A)UFxA
and
mx : FXA > Fx+iA
to be the map shown on the left side of the diagram. It follows from Constructions 4.1.7 and
4.1.6 that /A+i : FA+1 T is a subfunctor. Note that
{fx+i)(mxA) = (nA)(flA)
and mx is natural. For bookkeeping purposes, let us call
c\: Fx > Fx+1
the subfunctor which exhibits Fx C Fx+i.
If At is a limit ordinal, define
FK = \J{Fx : A < /t};
53
fK'FK>T is a subfunctor by 4.1.6.
Defining the sequence of n* s: The following definition of nx is not recursive; nx
can be defined once we know fx, but the definition of f\ does not involve nx at all. Consider
the following commutative diagram of functors and natural transformations.
The natural transformations z and e are defined by condition that rj = ez is the (epi, extremal
mono)factorization of 77, as described in 4.1.8. By assumption and the construction of the
sequence (Fx),
(r)idA)+1 CF0C Fx.
Let i\ denote the natural transformation such that e = fxiX) ix exists because e C i. Define
nx := i\z. Evidently,
V = fXnx
and
^a+i = cxnx.
What happens when the sequence terminates: Since A is extremally wellpow
ered, for each A, the sequence (F\A)\ of subobjects of TA eventually terminates, say when
A = 1. Again using extremal wellpoweredness there is an ordinal, say k2, such that the
sequence (F\FK(A))x of subobjects of TFKlA terminates at k2. Define = nK2, f fK2
and m = mK2. These assignments give natural transformations; to check this, one considers
a map / : A B, and chooses k large enough that the subobject sequences (described
above) terminate for both A and B. It should be clear that each subobject sequence {FXA)x
terminates at FA.
54
Verifying that (F, n, m) is a monad. To prove the unit laws, it suffices that for
all A,
('mxA){FxnxA) = (mxA)(nxFxA) = cxA
(because, once the sequence terminates, the caAs become identity maps on F\A). Fix any
ordinal A. Naturality of fx and nx implies
(TnxA)(fxA) = (fxFxA)(FxnxA)
and
CnxTA){fxA) = (FxfxA)(nxFxA);
Note that 77 = fxnx and, by definition of horizontal composition,
flA = (TfxA)(fxFxA) = (fxTA)(FxfxA).
Calculating, one finds
(.flA){FxnxA) = (fxTA)(FxfxA)(FxnxA) def. noted above
= (fxTA)(Fxr]A) noted above
= (TrjA)(fxA). /a is natural
Thus,
UlA){FxnxA) (TVA)(fxA)
and (by a similar computation)
(fxA)(nxFxA) = (r,TA)(fxA).
To show (mxA)(FxnxA) = cxA, consider the following diagram:
F,AI1^U FlA
55
the upper square commutes by an identity proved above; the lower square is the definition
of my, the upper triangle is trivial; the lower triangle is the unit law for (T, rÂ¡, /i). Reading
the perimeter of the diagram, one finds
fx A = (fx+1A)(mxA)(FxnxA).
By uniqueness of maps manifesting inequalities between subobjects,
cxA = (mxA){FxnxA).
The diagram needed to prove
(mxA)(nxFxA) = cxA
is similar and omitted. This establishes the unit laws.
To prove the associative law for (F,,m), one may choose sufficently large A, then
draw a diagram comparing the associative squares for T and F, using appropriate powers of
fx to compare the corners. The comparison squares commute because of the definitions of
mx and horizontal composition.
Definition 4.2.3. Assume that A is cocomplete. In A, unions of chains are colimits, if
whenever D is a diagram, where Obj(D) consists of a chain of extremal subobjects of A and
Map(D) consists of all inclusions (in Sub(A)) that exist among elements of Obj(D), the map
c : colimD > A induced by colimit properties is an extremal monomorphism.
In this case, c : colimD > A is an extremal subobject which contains each element of
Obj(D); thus there is an extremal monomorphism colimD > (J Obj(D). Since (J Obj(D) is
the supremum of Obj(D) in the Sub(A), this map must be an isomorphism.
Remark 4.2.4. The condition that unions of chains are colimits is satisfied in many nat
urally occuring categories. For example, in relational categories like Top, P, and Set,
extremal subobjects are just subsets with the induced structure, so all unions are colimits.
Assume A is a category of finitary algebras. The (set theoretic)union of a chain
(Ai) of subobjects is a subobject. The (set theoretic)union A' of (A,) lies inside a smallest
56
extremal subobject, which by our notation is (J A. Evidently, the coproduct of the chain
(A) of subobjects is A'. So the issue about whether unions of chains are coproducts
amounts to checking whether A! = 1J A,. This depends on the delicate issue of whether
epimorphisms are surjective. If epis are not necessarily surjective, then 1J A, will be the
largest subobject in which A' is epi, which will generally be larger than A'.
Definition 4.2.5. As above, assume A is an SFcategory, is a monad, and F is
a subfunctor of T which contains r?+1 id^. A partial algebra (A, a) is an object A equipped
with a map a : FA > A such that a(noA) = id^.
Let pAlg(Fo,no) denote the category of all partial algebras, with maps / : (A, a) >
(B, b) such that / : A > B is an Amap, and fa b(Fof).
Lemma 4.2.6. Continue with situation and hypotheses from f.2.2. Suppose A is cocomplete
and unions of chains are colimits. Suppose (A, a) is a partial algebra. The partial algebra
structure a extends to an Falgebra structure map a : FA > A if and only if for each ordinal
A there is a unique ax+\ making the diagram below commute.
Fl(A)
m\
*FX (A)
FX(A)^ A
Proof. To prove a map a : FA > A exists, use transfinite induction. The hypothesis gives ax
at sucessor ordinals. To construct ax, when A is a limit ordinal use the coproduct property
of FXA Ua Fq(A).
Inductively one shows that
(ext) for each A, ax = aA+i(cAA) (each ax+i extends the preceding ax),
(unit) for each A, ax(nxA) = kU,
57
If (unit) holds for sucessor ordinals, then it holds for all ordinals. Suppose
k, is a limit ordinal. The map aK is epi, because it is a limit of epimorphisms. Note that aK
and aKnKaK are both compatible maps FKA > A; the definition of colimits implies that
because aK is epi, id^ = aKnK.
The algebra associative law holds for (^4, a). If a aA, then aA = aA+1, so by
the hypothesis regarding the existence of ax such that a\+i(rri\A) = a\(F\a\) proves the
associative law.
If (unit) holds for A, then (ext) holds for A. Note that the equation a\+\(rn\A)
a\{F\a\) defines a\+1. By induction hypothesis (unit),
{F\a\)(F\n\A) id/r^^;
therefore precomposing both sides of the defining equation for aA+i with (F\ri\A) shows
aA = ax+i(m\A)(FxnxA) = ax+i{cxA).
The second equality follows from (mxA){FxnxA) = (cxA), which was proved in verifying the
unit laws for F.
If (unit) holds at A, then it holds for A + 1. Consider the following diagram.
The left triangle commutes because (unit) holds for A; the bottom trapezoid commutes by
definition of ax+1; the top trapezoid commutes because
(nxA)(FxnxA)(mxA) = (nxA)(cxA) = nx+iA.
58
The first equality holds by proof of the unit law for F; the second equality holds because of
the compatibility between nAs. Thus,
aA+i(nA+i) = ax+i(mxA)(FxnxA)(nxA)
= ax(Fxax)(FxnxA)(nxA)
= ax(nxA)
= id a;
this establishes the unit law for (A, a).
Use the notation Alg(F0, no) to denote the full subcategory of pAlg(Fo, no) containing
objects satisfying the hypotheses of Lemma 4.2.6.
Lemma 4.2.7. Suppose the preceding lemmas apply. Also suppose each mx is epi. The map
(A, a) i> (A, a) induces an equivalence of categories; the functors involved are
E : Alg(F0, n0) * AT,
defined by extension of structure and
F:AF> Alg(F0,n0)
given by restriction.
Proof. E is legitmately defined. Lemma 4.2.6 defines E on objects. Maps in AF are
Amaps compatible with the structure. So it suffices to show that for any Alg(Fo,n0) map
f): A B,
FXA FXB
bx
B
59
commutes for each ordinal A. By definition of Alg(F0, n0), the square commutes for A = 0.
Suppose it commutes for A; note that
a\(Fxa,\)
def. a^+i
= bx(Fx)(Fxax)
ind. hyp.
= bx(Fxbx)(F&)
ind. hyp.
= bx+,{mxB){Fl4>)
def. bx+i
= bx+i{Fx+i4>){mxA)
naturality mx.
Because mA is epi, one concludes ). At limit ordinals, properties of
colimits insure that the diagram commutes. By transfinite induction, and the defintion of
the extended structure, any map which preserves (Fo, n0)structure preserves Fstructure.
One also notes that restriction respects maps, because Fo is a subfunctor of F.
E and R form an equivalence. Obviously,
V(A, a0) G Obj(Alg(Fo, n0)), RE(A,ao) (A,a0).
The construction of the extended structure shows the restriction of any Fstructure to Fo
uniquely determines the Fstructure; hence
ER(A, a) = (A, a).
4.3 A Partial Algebra Which Does Not Extend
The category Set is an SFcategory and unions are colimits, so Meseguers Lemma
applies to Set. We show the necessity of Lemma 4.2.6s hypothesis that a map a\+j such that
ax+\m\ = a\(F\a\) can be defined. The section discusses a subfunctor of the freemagma
monad, which has an algebra that cannot extend to a monad algebra. A magma is a set
with a binary operation, subject to no equations.
The free magma monad (T, rj, Â¡i) has the following parts:
1. Given a set X, TX consists of all words with variables in A. A magma word is any
expression formed by finitely many applications of the rules:
60
(i) If x G X, then (^is a magma word.
(ii) If s, t are magma words, then
is a magma word the product of s and t. For ease in reading, we use binary tree
notation for products.
2. T defines a functor: given / : X > Y, to compute Tf we apply / to all members of
X in a given word, leaving the tree and circle structure unchanged.
3. The insertion of variables map r]X : X > TX. (rjX)(x) = (^.
4. The semantic composition map [iX : T2X > TX sends a word s G T2X of words
to a word in TX, by removing the circles around each element of TX used in making
s. n is also a natural transformation.
The notation takes a little while to soak in; to expedite the process, we consider a
calculation with TN. Suppose s G TN is the word
and ti = ss G T2(N) is the word
61
then (/N)(tx) is
Now we define a subfunctor F of T. Define the depth dep(n) of a node n in a binary
tree inductively by: the depth of the dep(root) = 0; if a is immediately below b, then the
dep(a) = dep(b) + 1. Define a leaf to be a node that has nothing below it. Let F consist of
all rooted, labeled, binary trees (i.e., magma words) such that the depth of each leaf is the
same. For example s and (/rN)(ti) are in F(N), but
is not, because dep{57) = 1 and dep(l) = dep(4) = 2. Evidently, F is a subfunctor, and for
any X, (rÂ¡X)+1(X) C F(X). But, for any nonempty X, (pX)(F2(X)) is not contained in
F(X). One readily verifies that F = T.
Now we define an partial algebra structure on N that does not extend to a Talgebra
structure. As in Lemma 4.2.6, a structure map for a pair (F,rj : id^ > F) is a map
a : FA > A satisfying id = a(rjA). Define n : F( N) t Non a tree r as follows: if the
depth of each leaf of r is odd then n(r) is the leftmost label, if the depth of each leaf of r
is even (or zero), then n(r) is the rightmost label. For example, n(s) 1 (s defined above)
and n(/iN)(ti) = 4.
62
In order for n : F(N) > N to extend to n : Fi(N)
making
F2( N)
Fn
F{ N)
/xN
F( N)
n i
*N
commute. No such a: can exist, for (/zN)(ti) is
N there must be a function rii
which equals (/iN)(t2), where t2 is
Finally, note that (Fn)(ti) is
so n(Fn)(t1) = 1. Note that (Fn)(t2) = (/rN)(t2), so n(Fn)(t2) = 4. Thus we have elements
11,12 G F2N, that /xN identifies and n(Fn) does not identify, so there is no function n\ such
that n\(nN) = n(Fn).
CHAPTER 5
FREE ALGEBRAS
This chapter explores a generalization of complete distributivity for P^objects. In
several steps, the free complete distributive lattice monad is constructed, then Meseguers
Lemmas are used to construct an appropriate submonad, whose algebras generalize com
pletely distributive lattices.
Section 5.1 describes monads for complete semilattices; D is the monad for complete
join semilattices, 11 is the monad for complete meet semilattices.
The gist of section 5.2 is that there is a lifting of 'll over V (and a lifting of D over 11).
Therefore, UV and VU are monads over P. (See results of Beck [5] summarized in Section
3.4.) In Raney [27], it was shown that complete meet distributivity is the same as complete
join distributivity. Hence, the composite monads UV and Dll have the same category of
algebras. The objects in either category are complete lattices where meets distribute over
joins, and joins distribute over meets.
Completely distributive complete lattices have been thoroughly studied. The basic
structure is described in Raney [27], [28], and [29]. Free objects over Set were initially
described in Markowsky [24]. Tunnicliff [32] discusses properties of the free completely
distributive lattice over a poset. Free objects over P and the relationship between completely
distributive lattices and continuous lattices are described in Hoffman and Mislove [14]. The
approach here is apparently new, but yields obviously equivalent free objects.
Each pair (j,m) of subset systems gives rise to a subfunctor of It'D. Meseguers
Lemmas are applied to this subfunctor to produce a monad F. Any P^,objects which
63
64
is P^embeddable in a completely distributive complete lattice is an Falgebra. Any F
algebra has a natural structure. Because of computational difficulties, no exact algebraic
characterization of Falgebras is given here.
A word about notation: the (functor parts) of the monads described below are given
by families of sets. Thus, checking the unit and associative laws requires working with many
levels of the power set tower. Roman letters S,T, denote sets. A subscipt designates the
power set complexity: Si G CP(A), S2 G S2A S2 is a family of sets, S3 G 73A S3 is a
family of families of sets, etc.
5.1 Complete semilattices
In this section, we describe the free complete join (resp. meet) semilattice on a poset
A, using the monad (V, d, fi) (resp. ('U,i,/x)). The description of D is well known; for
example, Meseguer [25] uses it. The reader will have noticed that q is used as a name for
two different natural transformations; this would normally be horrible notation, but in this
case, the formula for y is the same. Thus, our notational econony should cause no confusion.
The functors 'll and T> act on a poset A by
11(A) = {SCA:x>y6S =Â£> x G 5}
 the set of increasing subsets of A ordered by reverse inclusion and
T>(A) = {SCA:x x G 5}
the set of decreasing subsets of A ordered by inclusion.
Given monotone f : A > B, we define
Vf : 'D(A) * V{B) : S {b Â£ B : 3s S, b < f(s)}
and
Uf : U(A) U(B) : S ^ {be B :3s e S,b> f{s)}
The following facts will be of later use.
65
1. t and D are functors: trivially they respect identity arrows. Observe that for any
S 6 A and montone functions / : A > B, g : B > C,
T> (gf)(S) = {cC:3seS,c
= {ceC:3f(s)e'Df(S),c
 (Â£
and, similarly, U(gf)(S) = (Ug)(Uf)(S).
2. For each poset A, D(A) is a complete lattice, with supremum operation given by set
theoretic union, and infimum given by intersection. tt(A) is a complete lattice, with
infimum given by union and supremum given by intersection.
3. For each S2 C D(A),
In particular, this holds if 6 is empty. So Vf preserves all suprema. Similarly, if
S2 C 'U(A), then
Uf(\j6)=\j{(Uf)(S):See}
so that Uf preserves all nfima.
Define
dA: A > T>(A) : a 1> {x G A : x < a},
iA : A 'U(A) :aH{iÂ£i:i>a}, and
fiA : T>2{A) V(A) : S2 K2(A) ^ (J S2.
It may be puzzling that T>(A) is ordered by subset inclusion and li(.4) is ordered by
reverse subset inclusion. The fact that
x < y dA(x) C dA(y) <=> iA(y) C iA(x)
66
motivates the choice, because one wants iA to be order preserving. Moreover, one wants 1if
to preserve all intima, which only happens if the infimum operation in UA is union.
The reader may check that i, d and p are natural transformations. The monad
assocative law
Vy
D3
D2
yV
D2
V
holds for both D and 'll, because if S3 e T3(^4), then
UUs3 = U{l>:S^ss}.
The monad unit laws
D D2 D
hold for V, because if S 6 T)(A), then
5 = (J {(dA)(x) : x 5] = (J jr T>(A) : T C 5}.
Similar computation shows the unit laws hold for U.
Now to describe algebras over these monads:
Lemma 5.1.1. Let T C 1>(A), such that for all x E A, (dA)(x) T. Then the following
are equivalent:
1. there is an order preserving map a : T > A such that id^ = a(dA);
2. each S G T has a supremum;
3.dA : A T>A has a right adjoint.
67
The analogous conditions involving T C 1LA are also equivalent.
Proof. We give the proof for T C 1>A, leaving the upsidedown argument for T C UA to the
reader.
(2 ==> 1) Define a(S) = \J S and compute.
(1 => 2) a(S) is an upper bound for S, because if x G S, then (dA)(x) C 5,
whence x = a(dA)(x) < a(S). Suppose for all x G S, x < u. Then (dA){x) C (dA)(u), so
S = U{(gL4)(x) : x G 5} C (dA)(u), therefore q(5) < u.
(2 3) By definition of V S, j x C S <*=> x <\J S.
Proposition 5.1.2. P is equivalent to the category of T)algebras. Poo is equivalent to the
category ofUalgebras.
Proof. The first assertion is proved, leaving the second to the reader. By Lemma 5.1.1, any
Dalgebra has a supremum for each decreasing set; the supremum of an arbitrary set S is
equal to the supremum of the decreasing set ]. S. By the definition Dalgebra homomor
phisms coincide with order preserving functions which preserve all suprema of decreasing
sets. It is easy to see that a map preserves suprema of all decreasing sets if and only if it
preserves suprema of all sets.
Conversely, if A G P it has structure map
a : T>(A) A : S >*\J S.
In Lemma 5.1.1 it was noted that a satisfies the unit law for structure maps, a satisfies the
associative law by the order theoretic fact proved in Lemma 5.1.3 below.
Lemma 5.1.3. For any S2 G D2(A),
V {x A : 3S1 S2, X < VSi} = VU^2'
Proof. To verify this equality, one notes that for all x G G S2 such that x 5!,
so x < V S\. Thus the left hand side dominates the right. If u is an upper bound for (J S2,
and 35i G S2, x < V Si, then x < u. So the left hand side is the least upper bound of [J S;
this establishes the lemma.
68
5.2 Completely Distributive Complete Lattices
In the following arguments, the forgetful functors are supressed from the notation.
The lemmas that follow establish that there is a lifting of 'll over V (and similar arguments
show that there is a lifting of V over It). Explicitly this means:
1. For any Ualgebra A, there is a ltstructure on V(A).
2. It maps Dalgebra maps to Dalgebra maps, so we may view 'll as a functor P > P.
3. Both natural transformations i and p : U2 > U preserve all joins; similarly, d and
/x : D2 > D preserve all meets.
Lemma 5.2.1. For any poset A, both 11(A) and V{A) are complete lattices. Thus, 11(A)
and D(A) are both It and T> algebras.
Lemma 5.2.2. For any poset A,
1. dA : A > D(A) prserves all existing nfima.
2. iA : A 11(A) prserves all existing suprema.
Proof. Suppose S C A and u = f\S exists. Since dA is order preserving ,
(dA)(u) C P jcL4(:r) : x G
For the reverse inequality, suppose i G f]{dA(x) : x E S}, i.e., i is a lower bound of S. Then
l < u, so Â£ E dA(u). The proof for uA is upsidedown, but otherwise identical.
Lemma 5.2.3. If f : A B preserves all suprema, then It/ : 11(^4) > U(B) also pre
serves all suprema. Similarly, if f preserves nfima, so does T>(/). Therefore, for any order
preserving map f, both UT>f and VIIf preserve all nfima and suprema.
Proof. A proof of the first fact is given, the second is very similar, but notationally easier. Use
the fact that lt(A) is a complete sublattice of T(A), and therefore completely distributive.
69
Let (S\)xeL be an indexed subset of li(A); for each S\, choose a family (xx,K
that
S\ =T (x\,K k K).
Recall that:
joins in IL4 are intersections;
iA preserves suprema, which reads
T \Jyx = yx) = P T y\,
A
for any indexed family (yx) C A.
One computes as follows:
 Mnib^.")
A K
= ^(LKn : c:L+k}')
= (j{u(\Jxx,cW) C.L+K]
'a
= U { t V /(**<*)) =
= U { Pi T /(^a,^a) : c.L^k}
1 A
= n uf^
Corollary 5.2.4. For any poset A,
1. p, : T>2(A) > T)(A) prserves all nfima.
2. p : li2(A) > U(A) prserves all suprema.
: k K) such
70
Proof. Suppose S2,a Â£ X>2(^), for A L : claim 1 amounts to
unsw=nusw
A A
Clearly, the left hand side is contained in the right. Suppose that x is a member of the right
hand side, that is, for each A there exists Si,a 6 2,a with x Si,a Then
Si = Pi Si,a
A
is in each S2,a because each S2,a is downward closed relative to the inclusion order in 2(.A).
Since a: Si, this proves that x is in the right hand quantity, and the desired equality of
sets holds.
The proof for the It works similarly.
The preceding lemmas show that there are liftings of 'll over (and vice versa), so
we have:
Corollary 5.2.5. It = (ItCD, 77, where
7] : idp It : a i> {S VA :aCS}
and
v : UDlt > It : S4 ^ un*.
is a monad over IP whose algebras are completely distributive complete lattices.
Proof. The only thing left to be proved is the formulas for the natural transformations. By
Becks reasoning as summarized in Section 3.4, g is the horizontal composition i d. To
compute v : ltlt 4 It one uses the fact that vA is the structure map for the free algebra
UVA. Consider the situation in light of the discussion following Theorem 3.4.5. (Here It
plays the role of 5, while plays the role of T.) The relation a = ar{Tas) applied to the
object UVA implies
(vA)(s,) = u(n)(*)
= U{reBA:3S2efS,,rcfS2}
= Ui>
71
The assertion about algebras follows from Theorem 3.4.5.
Remark 5.2.6. UV is a monad because there is a lifting of U over P. One could use
the correspondence between liftings and distributive laws outlined in 3.4 to find a natural
transformation VU > UT>. This distributive law is not needed for the calculations which
follow, and is somewhat cumbersome, so its explicit description is omitted.
5.3 Some categories of algebras
In this section, Meseguers Lemmas are applied to the monads described above. Note
that P the category of posets and order preserving maps is a cocomplete SFcategory in
which unions of chains are colimits. Note that extremal monomorphisms in P are inclusions
of subsets with the induced order: each order preserving map / : A > B factors as A >
f(A) y B, where f(A) is the settheoretic image of A, with the order induced from B.
Thus, the results of Section 4.2 do apply to P.
The reader is advised to review the definition of subset system, given in 1.2.1, if
necessary.
Assume all subset systems Z are nontrivial in the sense that for each A, and a A,
{a} Z(A); this does not reduce the generality of the argument, because given any subset
system, one may adjoin all singletons to it without changing which optimum bounds are
preserved.
Theorem 5.3.1. [25] P7 and Pm are monadic.
Proof. Given a subset system j
J0(A) = {iS:Sej(A)}
defines a subfunctor of V. This extends to a subfunctor j of T> such that (j, d, Jl) is a monad.
The natural transformations d and JL are obtained from d and /i by suitably modifying the
domains and codomains. The precise definition of these maps is contained in the proof of
Lemma 4.2.6.
72
If (A, a0) has a J0structure, Lemma 5.1.1 shows that
ao(S) = \/5
for all 5 e Jo(A). Thus, to define ax one must show some map JXA A makes the following
diagram commute.
JSA J(V) *J0A
U
JXA
Lemma 5.1.3 shows that \J S is defined for all S J\A and, moreover, that for S2 JqA,
VU S2 = V(JV )(>)
Identical arguments show show that for each A, if
a\(S) = \J S
that
aA+1(S) = V5
It follows that any poset, A, in which each member of JqA has a supremum, each member of
J(A) has a supremum. Moreover, each map preserving J0suprema also preserves Jsuprema.
Thus, one obtains an equivalence of categories between P and PJ.
An upsidedown version of this argument shows that Pm is a category of algebras.
Given subset systems m and j, F := M0J0 forms a subfunctor of
T ILD.
By Corollary 5.2.5, (T,u,rj) is a monad on P. By 4.2.1, there exists a smallest F above F
that is closed under v. This monad is used to discuss the categories defined below.
73
Definition and Remarks 5.3.2. Let DP^ denote the category of completely distributive
complete lattices with maps, which preserve order and all optimum bounds. DP^ denotes
the full subcategory of containing objects such that there is a PJm map <Â¡>: A > B, with
4> Pextremal mono, and B Obj(DP^).
denotes the category of Falgebras, where F is the monad described above.
It will be shown that
SpFm C mi c Ml.
To see the first inclusion, one notes that any family T of subsets of a set X which is also a
P^ object is in DP^; the inclusion T C 7(X) is a P^embedding and T(X) is a completely
distributive complete lattice. Proposition 5.3.7 shows the second inclusion.
Let A 6 BP^, and <Â¡> : A + B be given as in the definition of OP^; let b : TB > B
be the Tstructure for B. The verification of the hypothesis of 4.2.6 proceeds by comparing
the diagram to be completed with the lt!Dalgebra associativity diagram for B. Prom this
point on, the notation of Lemma 4.2.2 is adopted with slight modification; v : T2 T is
the monad multiplication; v\ : F% F\+i is defined inductively (and plays the role that
was played by m\ in the proof of 4.2.2); rÂ¡ : idp T is the unit; rÂ¡\ : idp > F\ are defined
inductively, and play the role that was played by n\ in 4.2.2. The natural transformations
fxF\*T and c\ : F\ > F\+1 play the same roles as in Lemma 4.2.2.
Remark 5.3.3. By Lemma 3.4.5, the structure map b : TB > B is given by:
= A {V5i: 5i M
Since is a P^map, the following diagram commutes.
Fd> f\B
FA *FB >TB
ao
B
B
74
The left square expresses that <Â¡> commutes with the Fstructure; the right square expresses
that the Fstructure on B is the restriction of the Tstructure.
Lemma 5.3.4. Define ao : FA > A by a0(S2) = AC^ VX^O this map satisfies a0T]o = id^.
Proof. Let us begin by more explicitly calculating ao. Using the definition of li on maps we
find that
a0(S2) = A(UV)(*)
= f\{xeA:3S1eS2, \/Si
= /\{\/S1:S1eS2}.
This definition makes sense; each Si G Jq(A) so each V Si exists because A G P^. (U V) is
an order preserving map, and S2 G mojo(A), which implies that
(U\J)(S2) e M0(A).
Thus, the meet defining ao(S2) exist.
Recall rj(x) = {S G DA x C 5}. Now one calculates
ao(?(*)) = /\{\/S:SeDA,lxQS}
= x
For later calculations, it is crucial to know that some of the maps are epi or mono.
Lemma 5.3.5. For each ordinal X, u\A is surjective. Therefore v\A is Pepi.
Proof. Because of the structure of P, unions of subposets actually are set theoretic unions.
Let 5 G F\+1 = F\ U (i/\A)+1(F%A). If S G (v\A)+1(FfA), then there is an S2 G F%A such
that (v\A)(S2) = S. Since c\A = (i'xA)(r]xFxA) so for any Si G FxA, T2 := (r]xFxA)(S) is a
member of FfA such that (uxA){T2) Si.
This lemma holds in any category where unions of subobjects actually are settheoretic
unions.
75
After noting that extremal monomorphisms in P are injections / : A > B, where
f(A) has the order induced as a subset of B, one easily verifies:
Lemma 5.3.6. The functors U, T>, T, and F\ all preserve extremal monomorphisms.
Now one can verify the hypothesis of 4.2.6 is satisfied.
Proposition 5.3.7. Let A Obj(DP^), B e Obj(BP^), : A > B as in the definition of
BP,. Then
(ex) for each A, there is a unique map
A+i F\+iA A,
such that
a\(F\a\) a+iI'a;
(com) for each A, fiax = b(f\B)(F\).
Thus, OP^ C
Proof. Remark 5.3.3 established (com) for A = 0. If (com) holds for sucessor ordinals,
properties of unions insure that (com) holds at all ordinals.
76
Suppose (com) holds at stage A; we will establish (com) and (ex) hold for
stage A + 1. Consider the following diagram.
T2B >TB
The diagram commutes: the outer square commutes because of the Talgebra associative law
for B\ the left outer trapezoid commutes because of the definition of v\B\ commutativity of
the right outer trapezoid is obvious; the left inner trapezoid expresses naturality of v\\ the
top and right inner trapezoids commute by (com).
Establishing commutativity of the top outer trapezoid requires a bit more. Since
fx : F\ T is natural, (Tb)(fxTB) (fxB)(Fxb). By definition of horizontal composition
flB = (fxTB)(FxfxB).
Thus,
CTb){flB) = (Tb)(fxTB)(FxfxB)
= (fxB)(Fxb)(FxfxB)
This proves that the top outer square commutes.
Consider the (epi,extremal mono)factorization of
b(fx+1B)(Fx+iy,
77
say it factors as
Fx+lA Ai'CB.
Since is epi, the (epi, extremal mono)factorization of
q := b(fx+1B)(Fx+1)(uxA)
is
FXA A! C B.
The proof of Lemma 4.2.6 shows that each existing ax is split epi, which implies both ax and
(Fax) are epi. Since
q = 4>ax(Fxax),
the (epi, extremal mono) factorization of q is also:
FXA a(4a) ACB.
Uniqueness of factorization implies there is an isomorphism i : A > A! compatible with the
factorizations. So a^+i = ia is a map making (com) hold for A + 1.
Reading off the diagram one finds that:
ax(Fax) b(Tb)U2xB)(F2x)
= b{vB){fxB)(Fx)
= b(fx+1B)(Fx+1)(vxA)
=
The last step used that (com) holds for A + 1. Since 4> is a monomorphism, one concludes
that (ex) is satisfied for A + 1.
ax+1 is the only map which makes the inner square commute because vxA is epi, by
Lemma 5.3.5.
78
Corollary 5.3.8. DP^ is a full subcategory ofMPm.
Proof. By Lemma 5.3.3, any P^map preserves the Fstructure of a OP^ object. Lemma
4.2.7 implies that any P^map extends to a map.
Corollary 5.3.9. The forgetful functor DP^ > P has a left adjoint.
Proof. First note that FA G Obj(DP^) for any A, because fA : FA > TA is an embedding
of FA in a completely distributive lattice. Because FA is the free Falgebra on A, if
/ : A B is an order preserving map from A G Obj(P) to Be Obj(DP^) C Obj(PF), then
there is a DP^unique map f* : FA > B. This proves that rj has the universal property
described in 2.5.2.3; thus, F is the left adjoint to the forgetful functor DP^ P.
Corollary 5.3.10. The forgetful functors DP^ > OP^ and DP^ > have left adjoints.
Proof. Apply Proposition 3.3.6 and Corollary 6.2.8. (Note: the proof of 6.2.8 does not
depend on the arguments in this section, so the result is listed here. There is no circular
argument.)
There is some, rather limited, information about members of M^.
Proposition 5.3.11. If A G Obj(M^), then A G Obj(P^).
Proof. Define JA, MA, a, /?, 7, and 5, by the requirement that the squares below are
pullbacks.
JA ^ FA
7 a
T>A UHA
MA
FA
fA
UA ^UDA
By assumption, j(A) and rn(A) contain singletons, therefore
(iT>A)+1{J0A) C Mq Jo A C FA,
79
and
('UJA)+1(M0A) C Mq Jo A C FA.
It follows that
J0A C JA = (iDA)_1FA
and
Mo A C MA = (ltdA)_1FA
Again using the fact that j and m contain singletons,
(JA)+1(A) C JA
and
(A)+1(A) C MA,
so corestrictions dA\ : A > JA and iA : A > MA satisfying a(dA)\ = dA and j(iA) = fA
exist.
If a : FA > A is an Falgebra, then a(3(dA\) id^ and a5(iA\) = id^. Lemma 5.1.1
shows that for any subset 3 C DA containing each principal downsegment [ x, an order
preserving map y : 7 A satisfies y(dA\) id.4 if, and only if, y{S) = VS for each 5 J
(and analogously for subsets of 11 A). Thus, we conclude a(3 = \J and a6 = /\.
Remark 5.3.12. If j = m = u> the subset system which selects all finite subsets, then
SpF^ = DP^. For general j and m, the author does not know if equality holds, but he
suspects that the equality does not always hold.
Remark 5.3.13. A fundamental difficulty working with M^n is that one does not imme
diately know any order theoretic formula for the Fstructure maps. One might hope that
each Fstructure map is S2 /\(U \f)S2, but the author cannot presently substantiate such
hopes.
80
Remark 5.3.14. The author knows no examples of objects in Obj(M^) \ Obj(DP^). In
light of Corollary 5.3.10, any BP^ object is freely embedded in a completely distributive
lattice. In fact, A Obj(M^) is in Obj(BPJm) if, and only if, the unit of the adjunction
mentioned in 5.3.10 is Pextremal mono.
Li[18] explicitly constructs a map u : P > ISF(P) (not necessarily mono), with the
following properties:
ISF(P) is a complete, completely distributive lattice.
u preserves designated meets and joins.
If / : P > A is a map preserving designated meets and joins, and A is complete,
completely distributive, then there is a unique map /* : ISF(P) A such that
/ = f*u.
To explain the relationship between Lis results and the results here, note:
Li uses families SP C SP and IP C SP, which are only required to contain singletons.
Joins of SPsets and meets o IP sets are required to be preserved. In this document,
the choice of distinguished optimum bounds is made for all posets at once, via a subset
system.
Comparing universal properties, one sees that Lis construction applied to P Obj(M^)
with SP = j(P) and IP m(P) yields the left adjoint mentioned in 5.3.10.
Unfortunately, it is very difficult to see when u is an embedding. Let [5] denote the
smallest downward closed family of A containing S which is closed under SPjoins. Let SP
denote the class of increasing sets, which are closed under IPmeets. Li defines IS(P) to
be the familly of all decreasing sets in SP, ordered by S < T if, and only if, there is an
indexed chain (5')6/ with I = Q fl [0,1] such that S0 = S, SÂ¡ = T and whenever i < j,
Si fl [P \ Sj] = 0. The map u : P > ISP is defined by u(x) = {U 6 UA : U < x}.
81
Nonetheless, BP^, it is a fairly nice category. In fact, it is complete.
Proposition 5.3.15. BP^ is complete.
Proof. Note that a poset A is a P^ object if and only if there is an order preserving map
a : J0A x M0A > A such that a(d0A x i0A) kU.
Suppose (Ai)ii is a family of P^ objects and for each i I,
(j)i ^ Pi
is an embedding of Ai into a completely distributive lattice The proof of Lemma 3.1.3
implies that the poset product At is a P^object. Moreover, the productinduced map
$: ne/ A* * n/ Bi is an embedding which preserves (j, m)optiinum bounds. (Essentially
what is going on is that meets and joins are computed coordinatebycoordinate.)
If /, g : A > B axe P, maps, then the proof of Lemma 3.1.3 implies eq(f,g) is
a (j, m)complete subset of A. If A Obj(BP^), then A can be (j,m)embedded into a
completely distributive lattice, so eq(/, g) may also be so embedded.
Thus, any set of objects in BP^ has a product, and any pair of BP^ maps has an
equalizer. By Borceux [6, Volume 1, 2.8.1], this implies BPfa is complete.
CHAPTER 6
COEQUALIZERS
6.1 Epis and Equalizers in P
The results of this section characterize epis and regular epis in the category P of posets.
The presentation and proofs (except Construction 6.1.2 which is discussed in Meseguer [25])
are the work of the author, but the author believes it likely that they are not new. In all
statements, A and B are arbitrary posets.
Lemma 6.1.1. Let f : A > B be a monotone map. f is epi if and only if f is onto.
Proof. Since the forgetful functor P > Set is faithful, if / is onto then / is epi. For the
converse, suppose b G B \ f(A). Define
Si =i (f(A)n  b)
and
S2 =1 b.
Evidently, both S\ and S2 have the same intersection with f(A), but b G S2 \ Si. Therefore
characteristic functions of B \ Si and B\ S2 are distinct maps (say
ci,c2 : B 2
are respectively the characteristic functions of B \ Si and B \ S2) such that Cif = c2f. Both
Q are order preserving, because Si and S2 are decreasing sets.
Definition and Remarks 6.1.2. The following construction is paraphrased, following Mese
guer [25] pp 7374.
If / : A > B is any order preserving map, we may define a preorder X on 4 by
ai 4 a2 <=> /(ax) < /(a2).
82
83
This relation is obviously reflexive and transitive, but because / may not be injective
may not be antisymmetric.
Suppose (A, <) is a poset and ^ is a preorder strengthening <, i.e.,
i < a2 => ai < 2
Then we have an order preserving map a : A > A/ <. (Recall that A/ < is the set of
equivalence classes
{x A : x < a and a < x}
partially ordered by ;<.)
Define maps /1; h : A Bt to be equivalent, if there is an isomorphism, i : Bx > B2
such that
B2
commutes.
One verifies that:
1. The maps (f : A * B) and (<)>+ a \ A A/ < are mutually inverse
correspondences between
the set (modulo equivalence) of surjective maps / with domain A.
the set of preorders on A which strengthen <.
2. For any / : A* B, f = ca, where a : A > A/ ^ and c(a(a)) = /(a).
3. Given surjections fi,h'.A* B, there is a c : B\ > B2 if and only if
/i h
al ^ a2 ==> al ^ a2
84
Remark 6.1.3. Note that any intersection of preorders is a preorder. Thus, the class of
preorders strengthening < is a complete lattice. We say a set S of ordered pairs generates
< if < is the smallest preordered containing S. Note that if 5 generates <, we can describe
< explicitly: a0 < an if
clq dm or
there is a finite sequence ai,a2, a_j, such that for all i with 0 < i < n 1,
() t+i) Â£ S.
(The first bulleted condition insures that < is reflexive. The second requirement insures
that < is transitive; generally, if S is a relation, the relation obtained by applying the second
bulleted item only is called the transitive closure of S.)
One may verify the following construction using 6.1.2.
Construction 6.1.4. Let f,g : A > B be order preserving maps. The coequalizer of
f and g is the quotient of (B, <) by the smallest preorder containing < and C(g, h)
{(gx,hx),(hx,gx) : x G A).
Lemma 6.1.5. Let f : A > B be a surjective order preserving map. The following are
equivalent.
1. f is the coequalizer of some pair g, h : Ao h A.
2. There exist a poset Ao, and maps g, h : Ao > A such that the preorder < on (A, <) is
generated by < and C(g,h).
3. The preorder ^ on (A, <) is generated by < and K(f) {(x,y) E BxB : f(x) = f(y)}
f. The preorder < on (A, <) is generated by < and some equivalence relation.
Proof. (1 <=> 2) is evident from 6.1.4. (3 => 4) is trivial, because K{f ) is an equivalence
relation. We have
C(g,h)CK(f) c4
85
because fg = fh and
(x,y) eK(f) 4=^ f(x) < f{y) and f{y) < f(x).
f f
Since C(g,h) and
For (4 => 1), let X be generated by < and the equivalence relation E C A x A
with the trivial order. Then the projection maps tt\, 1T2 : E > A, given by
TVi(xi,X2) = Xi
are order preserving. The coequalizer of 7Ti and 7r2 is /.
Example 6.1.6. Some readers may be surprised by the fact that surjections are not always
Pregular epi. An example of this situation is any map / : 2/a( 2 from the (trivially
ordered) two point set onto a twopoint chain. Any such map / fails condition 3 from
Lemma 6.1.5.
6.2 Factorization of Maps Using Preorders
This section modifies the preorder factorization, to give a first approximation to
coequalizers in FJm.
Definition 6.2.1. Let Z be a subset system. Z is said to admit congruences if for all posets
A,
Z(A x A) = {S A x A : Z(A),i= 1,2}.
Lemma 6.2.2. Suppose both j and m are subset systems which admit congruences. Let
(A, <) be a poset and X be a preorder strengthening <. Then a : A A/ X preserves
(j, m)optimum bounds if and only if X is a FJmsubobject of Ax A.
Proof. We show that q(Vx) = Va(r) for any (x : i G 1} j{A). Clearly a(Vx,) is an
upper bound of (a(x) : i I}. Since X is a P^subobject of A x A, {xÂ¡ : i 1} G j(A) and
for all i, Xi < a,
V{(xj, a) : i E 1} = (Vxi(a)
86
is a member of d', Vx < a. Thus, any upper bound of {a(x) : i Â£ 1} dominates a(Vx).
(The proof for meets is identical, so omitted.)
Conversely, if a preserves jjoins and mmeets, and (for example)
{xi'.i /}, {yt : i Â£ 1} Â£ j{A)
and for all i Â£ /, x d then Vx, d So ^ is a P^subobject of ,4 x A.
Definition and Remarks 6.2.3. Suppose j and m admit congruences. Let (A, <) Â£
Obj(P^). If ^ is a preorder strengthening < and simultaneously a P^subobject of A x A,
then we say < is a (j, m)pocongruence.
Pocongruences of P^objects behave very much like preorders of posets.
1. The maps (f : A B) >>4 and (^) a : A > A/ < are mutually inverse
correspondences between
the set (modulo equivalence) of surjective maps /, which preserve (j, m)optimum
bounds and have domain A.
the set of (j, m)pocongruences on A.
2. For any / : A > B, f = ca, where a : A > A/ < and c(a(a)) = /(a). If / preserves
(j, m)optimum bounds, so do c and a.
3. Given (j,m)optimum bound preserving surjections fi,f2'A>Bl, there is c : B\
B2 if and only if
A Si
ai di a2 i di a2
Consideration of Lemma 6.2.2 shows that if / : A > B is surjective and S Â£ j(A)
(resp. S Â£ rn(A)) then f(S) has a supremum (resp. infimum) in B. Moreover, sets of the
form f(S) for S Â£ j(A) (resp. S Â£ m(A)) are the only sets whose optimum bounds are guar
anteed to exist. Thus, we are motivated to offer the following conditions on subset systems
to guarantee surjective (resp. Pregular epi) images of P, objects are (j, m)complete.
87
Definition and Remarks 6.2.4. Let Z be a subset system. We say Z preserves surjections
if whenever / : A B is a poset surjection,
Z(B) = {f(S):SeZ(A)}.
Similarly, we say Z preserves regular epis if whenever / : A B is a poset regular
epimorphism,
Z(B) = {f(S) : 5 Z(A)}.
Since poset regular epimorphims are surjective, any Z that preserves surjections also
preserves regular epimorphisms.
Any cardinal k, the associated subset system k preserves surjections. For if f : A > B
is surjective, then / is split epi (in Set). Hence, there is s : B > A (not necessarily order
preserving) such that idg = fs. Therefore, any set S C B with cardinality less than n is the
image of some set s(S) C A. The following remark provides an example of a subset system
that does not preserve surjections or regular epis. The author currently does not know of
any subset system that preserves regular epis without preserving surjections, but it seems
likely such a subset system exists.
A more interesting question, which the author also cannot currently answer is: do
there exist subset systems other than cardinals which preserve regular epimorphisms?
Remark 6.2.5. Let j = dir be the subset system which selects all upward directed subsets
of a poset. Let N denote a countable disjoint union of twopoint chains. (See drawing below
for help visualizing, and to fix notation.)
to t\
h
bo
Evidently N is jjoin complete.
88
Consider the function / : N N, where N = {0,1,2, } with the usual order,
defined by /(fc) = i (for i > 0) and /(_i) = i (for i > 1). Speaking roughly, f stacks
the two point chains. / is regular epi as a map of posets. Moreover, ^ is a jsubobject of
N x N, because / preserves all existing joins. But N is not jjoin complete, because N is
upward directed but has no join!
Because of the pocongruence factorization for (j, m)bound preserving maps (outlined
in 6.2.3) we have the following first approximation to the coequalizer in P^. The only thing
stopping map a : B > B/ < (described below) from actually being a coequalizer in P^ is
that B/ < is generally not {j, m)complete.
Construction 6.2.6. Suppose f,g:A> (B, <) are Â¥Jm maps, and C(f,g) is defined as in
6.I.4 Let X be the smallest pocongruence containing < and C(/, g). Then a : B > B/ X
has the following universal property: if h : B > C preserves (j, m)optimum bounds and
hf = hg, then there is a unique map i : B/ < > C such that h ia.
Since the quotient map a : B > B/ ^ is a surjection we have 
Corollary 6.2.7. Continue with the notation from Construction 6.2.6. If j and m preserve
surjections, then B/ < is (j,m)complete. Thus, a : B B/ < is the coequalizer of f and
g in FJm. In particular, if j = k and m = A are cardinality subset systems, then a is the
coequalizer.
Corollary 6.2.8. DP has coequalizers.
Proof. The preceding shows that any pair f,g:A>Be P^ has a coequalizer. Suppose
B is completely distributive. Since a : B > coeqPoo (/, g) is a surjection which preserves all
meets and joins, coeqp* (/, g) is completely distributive. In particular if both A and B are
completely distributive, then coeqp(f,g) is also the coequalizer in DP.
6.3 Factorization of Meetsemilattice maps
Several simplifications occur describing coequalizers in P, if the objects have a meet
semilattice structure, i.e., m> uj.
89
Lemma 6.3.1. Let f : A > B be a meetsemilattice map. f is Fregular epi if and only if f
is surjective.
Proof. Since Pregular epimorphisms are always surjective, one implication is trivial. For
the other, suppose / : A B is a surjection preserving binary meets. Then
x k V <=> f(xAy) = f(x).
Thus we have the sequence x,x A y, y, where f(x) = f(x A y) and x Ay < y. By Lemma
6.1.5, this proves / is regular epi.
As noted in the proof, the relation < is completely described by K(f). If objects
have a meetsemilattice structure, one may use congruences (equivalence relations that are
simultaneously (j, m)subalgebras) rather than the more complex pocongruences.
Finally we have:
Theorem 6.3.2. Suppose m > to and j and m preserve regular epis. Let f : A > B be a
P]m map. The following are equivalent.
1. f is surjective,
2. f is a quotient by some congruence K,
3. f is P regular epi,
4. f P^regular epi,
5. f is Ff,extremal epi.
Proof. The basic facts about pocongruences show (1 2). The preceding lemma shows
(1 <=> 3). (4 => 5) holds in any category.
To show (2 => 4), assume / : A > A/K. Since K C A x A, we have projection
maps ir i, 7t2 : K > A. f is plainly the coequalizer of 7ri,7r2.
90
(5 => 2) holds because we may factor any map / through its associated pocogruence.
If / is extremal epi, then / must coincide with the quotient by its associated pocongruence.
Corollary 6.3.3. Let j and m be subset systems which preserve regular epis. Let m > w.
The forgetful functors Up : FJm > P and USet : P, ~1' Set preserve and reflect regular epis.
Lemma 6.3.4. The forgetful functor Um^ : DP^ > Set reflects kernel pairs.
Proof. One verifies that the kernel pair of a map / : A B (calculated in DPJ consists of
the projections (from Ax A) restricted to the set
{(x,y) A x A : f(x) = f(y)}.
Applying Theorem 3.3.5, one obtains the following.
Corollary 6.3.5. Let j and m be subset systems which preserve regular epis. Let m > uj.
The forgetful functor Uset MP, ~1 Set is monadic.
6.4 Coequalizers in DP,
Theorem 6.4.1. For any pair g,h : A > B IMP^, the coequalizer coeq(
Proof. The construction of the coequalizer of maps g and h is illustrated by the following
diagram. The maps on the diagram will be defined below.
~D{C)
91
Let a : B > B/ < be the map given by Construction 6.2.6. Let / : B > C be any
BP^ map such that fg = fh. The existence and uniqueness of / is given by 6.2.6.
To construct the leftmost square, we apply the free completely distributive complete
lattice functor D = lD; rÂ¡ is the natural transformation which injects a poset H into DH.
To construct the rightoutside square, we note that (oo, oo)pocongruences are closed
under intersection. So there is a smallest (in the sense that it makes the fewest possible
identifications) (oo, oo)quotient
k : D(B/ <) > E(B/ X)
such that k(r](B/ X))a preserves all (j, m)optimum bounds. Similarly, define
Â£ : D(C) > E(C)
to be the smallest quotient such that Â£(rÂ¡C)f preserves all (j, m)bounds. Since
Â£D(J)(V(B/ l))a = Â£(VC)fa = Â£(rjC)f,
preserves all (j, m)bounds we have the induced map E(f).
Define Eo to be the smallest BP^subobject of E(B/ <) through which k(rj(B/ ^))
factors. (By Proposition 5.3.15, the intersection defining Eo exists.) The maps j : B/ < >
Eq and E0 > E(B/ X) are obtained by factoring (rÂ¡(B/ <))a) through E0.
By construction, the map Â£(rÂ¡C) : C > E(C) has the universal property that any map
(j, m)optimum bound preserving : C B, with B Obj(BP^) factors as (f) = tÂ£(rÂ¡C) for
a uniquely determined map t. Since C Obj(BP^r[), there is such a (j) which is Pextremal
mono. Thus, Â£(r]C) is Pextremal mono. Therefore, the P(epi, extremal mono) factorization
of Â£(r)C) produces the factorization Â£(jgC) = qm. By construction, m is both Pepi and
Pextremal mono; so m is a Pisomorphism. Moreover, q : C > E(C) is necessarily the
smallest BP^subobject of E(C') through which Â£{gC) factors.
We claim that coeq(g,h) = Eo via j : B > Eo. It suffices to show the existence
of a unique compatible map : Bo C. The commutativity of the largest rectangle
92
in the diagram above implies that k(rj(B/ ;<)) factors uniquely through the KMP^object
E(f)~1(C) C E(B). Because E0 is the smallest BP^object through which k(r](B/ :<))
factors, Eq C E(f)~l(C). This insures the existence and uniqueness of *.
CHAPTER 7
(j,m)SPACES
This chapter studies spaces obeying a Tostyle separation axiom. Section 7.1
defines spaces and describes a functorial Galois connection, which specializes to Galois con
nections between DP^ and SJm. Section 7.2 develops a convenient description of epimorphisms
in SJm, which generalizes a known characterization of epimorphisms of Tospaces. Section 7.3
gives constructions of limits, similar to those for topological spaces. Section 7.4 describes
quotient maps, and characterizes extremal and regular epis as quotient maps.
The last two sections are related to the problem of finding reflections in SJm] Section 7.5
gives the flat spectrum (co)reflection on spatial objects which is equivalent to a reflection
on Sj. This reflection on spaces is a generalization of the Tireflection of topological spaces.
Last, but not least, Section 7.6 partially describes the epicomplete SJrn objects; the
description is complete for T0spaces. Epicomplete Tospaces are chains with the specializa
tion order. Products of epicomplete T0spaces are not epicomplete, so there is no functorial
epicompletion in the category of Tospaces.
As mentioned in the introductory chapter, the research leading to this dissertation
began in an attempt to find reflections and coreflections in categories of generalized topo
logical spaces. After proving the results of this chapter, and reading Meseguer [25], the
author realized that additional assumptions where required on subset systems to insure that
subspaces could be reasonably defined. This realization prompted much of the thought sum
marized in Chapter 6 in particular Section 6.3; the author wanted to find when the theory
of this chapter was valid. These considerations, and construction of free DP^objects, be
came the main focus of the dissertation. However, this state of affairs leaves many questions
concerning reflections and coreflections in SJm untouched.
93

Full Text 
PAGE 1
68%6(7 6<67(06 $1' *(1(5$/,=(' ',675,%87,9( /$77,&(6 %\ (5,& 5 =(1. $ ',66(57$7,21 35(6(17(' 72 7+( *5$'8$7( 6&+22/ 2) 7+( 81,9(56,7< 2) )/25,'$ ,1 3$57,$/ )8/),//0(17 2) 7+( 5(48,5(0(176 )25 7+( '(*5(( 2) '2&725 2) 3+,/2623+< 81,9(56,7< 2) )/25,'$
PAGE 2
$&.12:/('*0(176 ZRXOG OLNH WR WKDQN f RUJH 0DUWLQH] 6RPHWLPHV ZRUGV DUH LQDGDTXDWH +H KDV EHHQ NLQG JHQHURXV SDWLHQW DQG LQWHUHVWLQJ WR ZRUN ZLWK KDYH JURZQ XQGHU KLV JXLGDQFH f 7KH SHRSOH ZKR LQKDELWHG WKH 8QLYHUVLW\ RI )ORULGD PDWK GHSDUWPHQW IURP WR 7KH\ PDNH WKH GHSDUWPHQW WKH ZRQGHUIXO SODFH LW LV f 0\ IULHQG &LHOR HDUWK LV ZRQGHUIXO ZKHQ RQH FDQ VHH LQWR WKH VN\ f 0\ 0 / f 0\ PDWKHPDWLFDO VLEOLQJV RWKHU JUDGXDWH VWXGHQWV ZKR ZRUNHG ZLWK RUJH HVSHFLDOO\ 5LFDUGR &DUUHUD f 0\ IDPLO\ :D\QH 3K\OOLV 0DUJLH HII DQG 5RE f 7KRVH WHDFKHUV ZKR E\ FKDOOHQJLQJ PH FDXVHG PH WR LPSURYH f 3DUWLFLSDQWV LQ WKH fI5LQJV DQG 2UGHUHG $OJHEUDLF 6WUXFWXUHVf FRQIHUHQFHV LQ *DLQHVn YLOOH DQG 1DVKYLOOH GXULQJ P\ WLPH LQ JUDGXDWH VFKRRO %HFDXVH RI WKHVH FRQIHUHQFHV IHHO OLNH DP MRLQLQJ D IDPLO\ RI UHVHDUFKHUV UDWKHU WKDQ MXVW fJHWWLQJ D GHJUHHf ,W LV DQ KRQRU WR NQRZ WKHP f 7KRVH ZKR KHOSHG SURRIUHDG WKLV GRFXPHQW FRQVWUXFWLYH FRPPHQWV ZHUH PDGH E\ RUJH 0DUWLQH] 6FRWW 0F&XOORXJK DQG 3KDP 7LHS f $Q\RQH WDNLQJ WKH WLPH WR UHDG WKHVH ZRUGV D GLVVHUWDWLRQ OLNH DQ\ ERRN LV PHDQW WR EH UHDG X
PAGE 3
7$%/( 2) &217(176 $&.12:/('*(0(176 $%675$&7 Y &+$37(56 ,1752'8&7,21 'LVWULEXWLYH /DWWLFHV 6XEVHW 6\VWHPV 0HWKRGV DQG 5HVXOWV 35,0(5 21 &$7(*25,(6 $1' 326(76 'LVWLQJXLVKHG 0DSV %RXQGV 1DWXUDO 7UDQVIRUPDWLRQV &Rf/LPLWV $GMRLQW )XQFWRUV $/*(%5$6 2) $ 021$' &DWHJRULHV RI $OJHEUDV $GMRLQW &RQQHFWLRQV LQGXFH 0RQDGV 'HWHFWLQJ &DWHJRULHV RI $OJHEUDV 'LVWULEXWLYH /DZV *(1(5$7,1* 68%021$'6 6XEIXQFWRUV 0HVHJXHUfV /HPPDV $ 3DUWLDO $OJHEUD :KLFK 'RHV 1RW ([WHQG )5(( $/*(%5$6 &RPSOHWH VHPLODWWLFHV &RPSOHWHO\ 'LVWULEXWLYH &RPSOHWH /DWWLFHV 6RPH FDWHJRULHV RI DOJHEUDV &2(48$/,=(56 (SLV DQG (TXDOL]HUV LQ 3 LLL
PAGE 4
)DFWRUL]DWLRQ RI 0DSV 8VLQJ 3UHRUGHUV )DFWRUL]DWLRQ RI 0HHWVHPLODWWLFH PDSV &RHTXDOL]HUV LQ '8f MPf63$&(6 6SDWLDO6REHU )XQFWRULDO *DORLV &RQQHFWLRQ 7KH 6NXOD 7RSRORJ\ DQG ([WUHPDO 0RQRV &RPSXWLQJ /LPLWV 4XRWLHQWV ([WUHPDO DQG 5HJXODU (SLV )ODW 6SHFWUD (SLFRPSOHWH REMHFWV LQ 5()(5(1&(6 %,2*5$3+,&$/ 6.(7&+ QR
PAGE 5
$EVWUDFW RI 'LVVHUWDWLRQ 3UHVHQWHG WR WKH *UDGXDWH 6FKRRO RI WKH 8QLYHUVLW\ RI )ORULGD LQ 3DUWLDO )XOILOOPHQW RI WKH 5HTXLUHPHQWV IRU WKH 'HJUHH RI 'RFWRU RI 3KLORVRSK\ 68%6(7 6<67(06 $1' *(1(5$/,=(' ',675,%87,9( /$77,&(6 %\ (ULF 5 =HQN $XJXVW &KDLUPDQ RUJH 0DUWLQH] 0DMRU 'HSDUWPHQW 0DWKHPDWLFV 'LVWULEXWLYH ODWWLFHV DORQH RU ZLWK HQULFKHG VWUXFWXUH DUH PDWKHPDWLFDO REMHFWV RI IXQGDPHQWDO LPSRUWDQFH 7KLV WH[W VWXGLHV JHQHUDOL]HG GLVWULEXWLYH ODWWLFHV WKH JHQHUDOn L]DWLRQ LV WKDW FHUWDLQ LQILQLWH PHHWV DQG MRLQV DUH UHTXLUHG WR H[LVW 6XEVHW V\VWHPV QDWXUDO UXOHV ZKLFK VHOHFW D IDPLO\ RI VXEVHWV RI HDFK SRVHWf M DQG P ODEHO ZKLFK VHWV KDYH MRLQV DQG PHHWV UHVSHFWLYHO\ $ FDOFXOXV RI VXEIXQFWRUV LV GHYHORSHG XVLQJ WKLV FDOFXOXV LW LV VKRZQ WKDW DQ\ VXEIXQFWRU ) RI D PRQDG FRQWDLQLQJ WKH LPDJH RI WKH XQLWf JHQHUDWHV D VXEPRQDG ) 8QGHU VXLWDEOH FRQGLWLRQV DQ\ SDUWLDO )DOJHEUD H[WHQGV WR DQ )DOJHEUD 7KH PRQDG ) IRU WKH IUHH GLVWULEXWLYH M PfFRPSOHWH ODWWLFH LV WKH VXEPRQDG RI WKH FRPSOHWHO\ GLVWULEXWLYH FRPSOHWH ODWWLFH PRQDG JHQHUDWHG E\ D VXEIXQFWRU REWDLQHG IURP M DQG P 7KH FDWHJRU\ '3A RI M PfFRPSOHWH ODWWLFHV ZKLFK FDQ EH HPEHGGHG LQ D FRPSOHWHO\ GLVWULEXWLYH FRPSOHWH ODWWLFH LV D IXOO VXEFDWHJRU\ RI )DOJHEUDV '3Q LV FRPSOHWH DQG KDV FRHTXDOL]HUV M PfFRPSOHWH IDPLOLHV RI VXEVHWV RI D VHW JHQHUDOL]HG WRSRORJLFDO VSDFHVf DUH LQYHVn WLJDWHG LQ DQDORJ\ WR FODVVLFDO SRLQWVHW WRSRORJ\ $VVXPLQJ VXLWDEOH UHVWULFWLRQV RQ M DQG
PAGE 6
P VXEVSDFHV FDQ EH GHILQHG $VVXPLQJ WKHVH UHVWULFWLRQV WKHUH DUH ZHOOEHKDYHG FDWHJRULHV FRUUHVSRQGLQJ WR 7R DQG VREHU VSDFHV
PAGE 7
&+$37(5 ,1752'8&7,21 'LVWULEXWLYH /DWWLFHV $ VHPLODWWLFH LV D VHW ZLWK DQ RSHUDWLRQ $ VDWLVI\LQJ WKH IROORZLQJ XQLYHUVDOO\ TXDQWLILHG HTXDWLRQV f D $ D D f D $ E fÂ§ E $ D DQG f D $ Ef $ F D $ E $ Ff $ DQG 9 ZKLFK DUH UHODWHG E\ WKH FRQGLWLRQ D 9 E fÂ§ E E WR PHDQ D $ E D ZLWK WKLV RUGHU D $ E LV WKH ODUJHVW WKLQJ VPDOOHU WKDQ ERWK D DQG E DQG D 9 E LV WKH VPDOOHVW WKLQJ ODUJHU WKDQ ERWK D DQG E $ ODWWLFH LV GLVWULEXWLYH LI HLWKHU RI WKH IROORZLQJ HTXLYDOHQW XQLYHUVDOO\ TXDQWLILHG HTXDWLRQV KROG D $ E 9 Ff fÂ§ D $ Ef 9 D $ Ff D 9 E $ Ff D 9 Ef $ D 9 Ff /HW XV FRQVLGHU WKH FRQFHSW DQG LWV UHOHYDQFH $ GLVWULEXWLYH ODWWLFH EHDUV VRPH UHVHPEODQFH WR RUGLQDU\ DULWKHPHWLF ZKHUH $ DQG 9 FRUUHVSRQG WR DGGLWLRQ DQG PXOWLSOLFDWLRQ WKH SULQFLSDO GLIILFXOW\ ZLWK WKLV YLHZ LV WKDW D $ D D 9 D D ZKLFK GRHV QRW KROG LQ DULWKPHWLF 7KHUH LV PRUH V\PPHWU\ LQ WKH $ ODWWLFH LV D VHW ZLWK WZR VHPLODWWLFH RSHUDWLRQV WKDW D $ E D $ ODWWLFH PD\ EH SDUWLDOO\ RUGHUHG E\ GHILQLQJ D
PAGE 8
HTXDWLRQV GHILQLQJ GLVWULEXWLYH ODWWLFHV WKDQ LQ RUGLQDU\ DULWKPHWLF IRUPDOO\ f$f DQG f9f DUH LQWHUFKDQJDEOH DQG VZLWFKLQJ WKHP UHYHUVHV WKH RUGHU 'LVWULEXWLYH ODWWLFHV DUH LQWHUHVWLQJ DOJHEUDLF VWUXFWXUHV LQ WKH VDPH ULJKW DV ULQJV VWUXFWXUHV ZLWK DQG Â‘ VHQVLEO\ GHILQHG $QRWKHU SHUVSHFWLYH LV WKDW GLVWULEXWLYH ODWWLFHV DUH PRGHOV RI ORJLF ZLWK f$f UHSUHn VHQWLQJ fDQGf DQG f9f UHSUHVHQWLQJ fRUf 2EYLRXVO\ WKH FRQQHFWLYHV fDQGf DQG fRUf ERWK VDWLVI\ WKH VHPLODWWLFH UXOHV ,Q WKLV LQWHUSUHWDWLRQ WKH GLVWULEXWLYH ODZV DUH WDXWRORJLHV DQG D E PHDQV WKH SURSRVLWLRQ D LPSOLHV E &RQYHQWLRQDO ORJLFV DUH RIWHQ GHVFULEHG E\ GLVn WULEXWLYH ODWWLFHV REH\LQJ H[WUD HTXDWLRQV ZKLFK FRUUHVSRQG WR DGGLWLRQDO WDXWRORJLHV WR EH PRGHOHG $ SHUVSHFWLYH SDUWLFXODUO\ UHYHODQW WR WKH DXWKRU LV WKDW GLVWULEXWLYH ODWWLFHV ZLWK VRPH DGGLWLRQDO VWUXFWXUHf GHVFULEH WRSRORJLFDO VLWXDWLRQV ,QWXLWLYHO\ D WRSRORJLFDO VSDFH LV DQ DPRUSKRXV EORE IURP ZKLFK FHUWDLQ SLHFHV FDQ EH FOHDQO\ UHPRYHG 7KH UHPRYDEOH SLHFHV DUH FDOOHG FORVHG SDUWV DQG WKH FRPSOHPHQWV LH WKLQJV OHIW RYHU DIWHU D FORVHG SDUW KDV EHHQ UHPRYHGf DUH FDOOHG RSHQ SDUWV 7KH XVXDO GHILQLWLRQ RI D WRSRORJLFDO VSDFH LV D VHW ; WRJHWKHU ZLWK D GHVLJQDWHG IDPLO\ RI RSHQ VXEVHWV VXFK WKDW f ; DQG WKH HPSW\ VXEVHW DUH RSHQ f LI 8 DQG 9 DUH RSHQ VR LV WKH LQWHUVHFWLRQ 8 IW 9 DQG f LI 8Lf LV D IDPLO\ RI RSHQ VXEVHWV WKHQ WKH XQLRQ _W IÂ LV DOVR RSHQ 7KH ODWWLFH RI RSHQ VHWV HQFRGHV KRZ D VSDFH LV ZRYHQ WRJHWKHU 5HDO DQDO\VLV SURYLGHV VRPH MXVWLILFDWLRQ IRU WKH XVXDO GHILQLWLRQ RI D WRSRORJLFDO VSDFH +RZHYHU WKH DXWKRU ZRQGHUHG KRZ WKH FRQFHSW RI D WRSRORJ\ FKDQJHV LI RQH YDULHV WKH GHILQLWLRQ E\ UHTXLULQJ HLWKHU IHZHU XQLRQV RI RSHQ VHWV EH RSHQ RU PRUH LQWHUVHFWLRQV RI RSHQ VHWV EH RSHQ 6XEVHW 6\VWHPV $ FDWHJRU\ LV DQ DEVWUDFW FODVV RI REMHFWV ZLWK VWUXFWXUH DQG PDSV RU KRPRPRU SKLVPVf ZKLFK SUHVHUYH WKH UHOHYDQW VWUXFWXUH 7KH PDSV LQ D FDWHJRU\ DOORZ FRPSDULVRQV
PAGE 9
EHWZHHQ REMHFWV &DWHJRU\ WKHRU\ DOORZV IRUPDO FRPSDULVRQV EHWZHHQ YDULRXV WKHRULHV RI DOJHEUDLF REMHFWV HJ RQH FDQ FRPSDUH WKH FDWHJRU\ RI DOO ULQJV ZLWK WKH FDWHJRU\ RI DOO GLVWULEXWLYH ODWWLFHV 0DQ\ FDWHJRULHV RI SDUWLDOO\ RUGHUHG VHWV ZLWK DGGLWLRQDO VWUXFWXUHf ILW LQWR D VLPSOH JHQHUDO SDWWHUQ 7KH REMHFWV DUH SDUWLDOO\ RUGHUHG VHWV LQ ZKLFK FHUWDLQ LQWHQWLRQDOO\ GLVWLQn JXLVKHG VXEVHWV KDYH VXSUHPD RU QILPD ZKLOH WKH PDSV DUH RUGHU SUHVHUYLQJ IXQFWLRQV ZKLFK SUHVHUYH VDLG ÂQILPD DQG VXSUHPD 3 GHQRWHV WKH FDWHJRU\ RI DOO SRVHWV DQG RUGHU SUHVHUYLQJ PDSV 7KH FKDOOHQJH LV KRZ RQH VHOHFWV VXEVHWV ZKLFK KDYH ÂQILPD DQG VXSUHPD :H IROORZ 7KDWFKHU :ULJKW DQG :DJQHU >@ ZKR LQWURGXFHG WKH XVHIXO EXW EODQGO\ QDPHGf FRQFHSW RI D VXEVHW V\VWHP 'HILQLWLRQ DQG 5HPDUNV $ VXEVHW V\VWHP = LV D UXOH ZKLFK DVVLJQV D IDPLO\ =$f RI VXEVHWV WR DQ\ SRVHW $ VXFK WKDW IRU DQ\ RUGHU SUHVHUYLQJ PDS $ fÂ§! % ^I6f =$f` & =%f =FRPSOHWH SRVHWV DUH SRVHWV $ LQ ZKLFK HDFK VHW 6 =$f KDV D VXSUHPXP 6HH (UQH >@ IRU PRUH H[WHQVLYH ELEOLRJUDSK\ UHJDUGLQJ VXEVHW V\VWHPVf 7KH IROORZLQJ H[DPSOHV RI VXEVHW V\VWHPV PD\ FRQYH\ WKH JHQHUDOLW\ DQG XVHIXOQHVV RI WKH FRQFHSW ,I N LV D FDUGLQDO ZH XVH Q WR GHQRWH WKH VXEVHW V\VWHP ZKLFK VHOHFWV DOO VXEVHWV ZLWK FDUGLQDOLW\ OHVV WKDQ N :H XVH RR IRU WKH IXQFWRU ZKLFK SODFHV QR UHVWULFWLRQ RQ FDUGLQDOLW\ 7KLV LV D VXEVHW V\VWHP EHFDXVH IRU DQ\ IXQFWLRQ $ fÂ§! % 6 & $ ZLWK _$_ N LPSOLHV _$f_ Q $ VXEVHW 6 & $ LV XSZDUGf GLUHFWHG LI IRU HDFK [ \ 6 WKHUH H[LVWV X 6 VXFK WKDW [ X DQG \ LW 7KH UXOH GLU ZKLFK VHOHFWV DOO GLUHFWHG VXEVHWV RI D SRVHW LV D VXEVHW V\VWHP LI I^[fI\f 6nf DQG 6 LV GLUHFWHG WKHQ WKHUH LV X 6 VXFK WKDW [ X DQG \ X 7KXV IXf LV D FRPPRQ XSSHU ERXQG IRU I[fI\f
PAGE 10
$ VXEVHW 6 & $ LV XSZDUGf FRPSDWLEOH LI IRU DQ\ [\ 6 WKHUH LV X f $ VXFK WKDW [ X DQG \ X f&RPSDWLEOHf GLIIHUV IURP fGLUHFWHGf EHFDXVH IRU WKH IRUPHU X 6 ZKLOH IRU WKH ODWHU ZH RQO\ UHTXLUH X $ 6LPLODU DUJXPHQWV VKRZ WKDW FRPSDW ZKLFK VHOHFWV FRPSDWLEOH VXEVHWV LV D VXEVHW V\VWHP $ VXEVHW RI 6 & $ LV D FKDLQ LI [ \ f 6 LPSOLHV [ \ RU \ [ 7KH UXOH FK ZKLFK VHOHFWV DOO FKDLQV LQ D SRVHW LV D VXEVHW V\VWHP :H VD\ D VXEVHW & $ LV XSZDUGf VHOIERXQGHG LI WKHUH LV V 6 VXFK WKDW IRU DOO [ f 6 [ V D VHOIERXQGHG VHW FRQWDLQV D PD[LPXP HOHPHQW 7KH UXOH VE ZKLFK VHOHFWV DOO VHOIERXQGHG VHWV LV D VXEVHW V\VWHP 1RWH WKDW DQ\ RUGHU SUHVHUYLQJ PDS SUHVHUYHV MRLQV RI XSZDUG VHOIERXQGHG VHWV $ QRQH[DPSOH $Q DQWLFKDLQ LV D VHW RI SDLUZLVH LQFRPSDUDEOH HOHPHQWV 7KH UXOH DF ZKLFK VHOHFWV DQWLFKDLQV LV QRW D VXEVHW V\VWHP EHFDXVH WKHUH LV DQ RUGHU SUHVHUYLQJ VXUMHFWLRQ fÂ§! ZKHUH LV WKH WZRSRLQW DQWLFKDLQ DQG LV WKH WZR SRLQW FKDLQ *HQHUDWLQJ H[DPSOHV /HW 4 EH DQ\ FODVV RI SRVHWV FORVHG XQGHU RUGHU SUHVHUYLQJ VXUMHFWLRQV 2QH PD\ GHILQH D VXEVHW V\VWHP =T E\ =D$f ^6 & $ 6 H 4` 7KLV FRQVWUXFWLRQ VKRZV WKHUH LV D JUHDW PXOWLWXGH RI VXEVHW V\VWHPV 7KH VXEVHW V\Vn WHP FRPSDW GHVFULEHG DERYH LV QRW JHQHUDWHG WKLV ZD\ EHFDXVH RQH FDQQRW GHWHUPLQH LI 6 & $ LV FRPSDWLEOH PHUHO\ E\ ORRNLQJ DW WKH SRVHW 6 ZLWK LWV LQGXFHG RUGHU :H XVH VXEVHW V\VWHPV ZKLFK ZH JHQHULFDOO\ FDOO M DQG P WR VHOHFW ZKLFK VXEVHWV KDYH MRLQV DQG PHHWV UHVSHFWLYHO\ 1RZ ZH HQXPHUDWH VRPH FDWHJRULHV RI LQWHUHVW LQ WKLV GLVFXVVLRQ WKH FDWHJRU\ RI DOO M PfFRPSOHWH SRVHWV WKDW LV SRVHWV LQ ZKLFK MVXSUHPD DQG 72LQILPD H[LVW DQG DUH SUHVHUYHG E\ DOO PDSV
PAGE 11
'3A WKH FDWHJRU\ RI DOO FRPSOHWHO\ GLVWULEXWLYH FRPSOHWH ODWWLFHV '3A WKH IXOO VXEFDWHJRU\ RI 3A FRQWDLQLQJ REMHFWV ZKLFK FDQ EH 3AHPEHGGHG LQ D FRPSOHWHO\ GLVWULEXWLYH FRPSOHWH ODWWLFH '3A LV GLVFXVVHG LQ 6HFWLRQ 0Â WKH IXOO VXEFDWHJRU\ RI 3A FRQWDLQLQJ SRVHWV ZLWK DQ ) VWUXFWXUH ZKHUH ) LV WKH PRQDG GHILQHG LQ 6HFWLRQ 6S)P WKH IXOO VXEFDWHJRU\ RI 3Â FRQWDLQLQJ VSDWLDO SRVHWV 6HH &KDSWHU WKH FDWHJRU\ RI JHQHUDOL]HG VSDFHV ZLWK M PfFRPSOHWH IDPLOLHV RI GLVWLQJXLVKHG VXEVHWV VHH &KDSWHU )e WKH IXOO VXEFDWHJRU\ RI 6S+"P FRQWDLQLQJ VSDWLDO SRVHWV ZLWK IODW VSHFWUXP $ LV GHILQHG WR KDYH IODW VSHFWUXP LI WKH PDSV $ fÂ§ DUH WULYLDOO\ RUGHUHG VHH 6HFWLRQ 0HWKRGV DQG 5HVXOWV 7KH GLVFXVVLRQ RI '3A DQG 0MQ ZKLFK RIIHU JHQHUDOL]DWLRQV RI GLVWULEXWLYH ODWWLFHV XVHV WKH ODQJXDJH RI FDWHJRU\ WKHRU\ $ TXLFN LQWURGXFWLRQ RFFXUV LQ &KDSWHU 7KH FUXFLDO QRWLRQ RI D IUHH REMHFW LV IRUPDOL]HG E\ PRQDGV ZKLFK DUH LQWURGXFHG LQ &KDSWHU &KDSWHU GHVFULEHV D WKHRU\ RI VXEIXQFWRUV 7KH FODVV RI VXEIXQFWRUV RI D IXQFWRU ) EHDUV D VWURQJ VLPLODULW\ WR WKH SRZHU VHW ODWWLFH RI D VHW ; *LYHQ D PRQDG ZLWK IXQFWRU SDUW 7 0HVHJXHUfV /HPPDV DQG VKRZ WKDW DQ\ VXEIXQFWRU ) RI 7 FRQWDLQLQJ DOO FRQVWDQWV KDV D fPRQDGLF FORVXUHf LH D VPDOOHVW VXEPRQDG ) RI 7 ZKLFK H[FHHGV ) 0HVHJXHUfV /HPPDV ZHUH IRUPXODWHG DQG SURYHG E\ WKH DXWKRU EXW WKH WHFKQLTXH LV VLPLODU WR RQH LQ 0HVHJXHU >@f ,QWXLWLYHO\ 7; LV WKH IXOO VHW RI SRO\QRPLDOV LQ WKH VHQVH RI XQLYHUVDO DOJHEUDf ZLWK YDULDEOHV LQ ; ); LV D QDWXUDO VXEVHW RI SRO\QRPLDOV DQG ); LV WKH VPDOOHVW QDWXUDO VXEVHW RI SRO\QRPLDOV ZKLFK LV FORVHG XQGHU FRPSRVLWLRQ DQG FRQWDLQV ); $Q DOJHEUD VWUXFWXUH IRU 7 LV D ZD\ RI HYDOXDWLQJ fDOO SRO\QRPLDOVf D SDUWLDO DOJHEUD VWUXFWXUH IRU ) LV D ZD\ RI fHYDOXDWLQJ SRO\QRPLDOV LQ ) 8QGHU VXLWDEOH FRQGLWLRQV SDUWLDO DOJHEUDV H[WHQG WR )DOJHEUDV
PAGE 12
,Q &KDSWHU 0HVHJXHUfV /HPPDV DUH EURXJKW WR EHDU XSRQ PRQDGV IRU IUHH FRPSOHWH VHPLODWWLFHV DQG IUHH FRPSOHWHO\ GLVWULEXWLYH FRPSOHWH ODWWLFHV *LYHQ DQ\ VXEVHW V\VWHPV M DQG P WKHUH LV D VXEPRQDG ) RI WKH IUHH FRPSOHWHO\ GLVWULEXWLYH FRPSOHWH ODWWLFH PRQDG 7KH FDWHJRU\ 0e RI )DOJHEUDV RIIHUV D VRPHZKDW P\VWHULRXVf JHQHUDOL]DWLRQ RI WKH FDWHJRU\ RI GLVWULEXWLYH ODWWLFHV 7KH VXEFDWHJRU\ %3A FRQWDLQLQJ DOO M PfFRPSOHWH SRVHWV ZKLFK PD\ EH HPEHGGHG LQ D FRPSOHWHO\ GLVWULEXWLYH ODWWLFH LV VRPHZKDW HDVLHU WR XQGHUVWDQG DQG VWLOO ZHOO EHKDYHG 7KH H[LVWHQFH RI IUHH REMHFWV LQ 23A FRQWUDVWV ZLWK WKH QRQH[LVWHQFH RI IUHH REMHFWV LQ 3ee >@ DQG WKH FDWHJRU\ RI FRPSOHWH %RROHDQ DOJHEUDV > @ $ IXQGDPHQWDO GLIIHUHQFH EHWZHHQ WKHVH FDWHJRULHV DQG '3Â LV WKH UHTXLUHPHQW WKDW MRLQV DQG PHHWV REH\ D GLVWULEXWLYH ODZ 7KH SRZHU RI FDWHJRU\ WKHRU\ FRPHV DV PXFK IURP ZKDW LW LJQRUHV DV ZKDW LW H[DPLQHV 6LJQLILFDQW FRQFOXVLRQV DUH RIWHQ REWDLQHG ZLWKRXW H[DPLQLQJ WKH fJUXEE\ GHWDLOVf RI ZKDW LV JRLQJ RQ %XW WKLV LQQRFHQFH RI fJUXEE\ GHWDLOVf OLPLWV WKH VFRSH RI LQYHVWLJDWLRQ ,Q WKH FDVH RI WKLV GRFXPHQW VHYHUDO QLFHO\ EHKDYHG FDWHJRULHV 0"P 3,f DQG 6SgP DUH LQWURGXFHG )RU JHQHUDO VXEVHW V\VWHPV M DQG P WKH DXWKRU GRHV QRW HYHQ NQRZ LI WKHVH FDWHJRULHV GLIIHU 7KH HQG RI 6HFWLRQ IURP &RUROODU\ RQZDUGV GHVFULEHV PRVW RI WKH DXWKRUfV NQRZOHGJH RQ WKH UHODWLRQVKLS EHWZHHQ WKHVH FDWHJRULHV &KDSWHU H[SORUHV FRQJUXHQFHV TXRWLHQWV DQG FRHTXDOL]HUV LQ 3 3P DQG 2+f 0XFK FODVVLFDO DOJHEUD ULQJ WKHRU\ ODWWLFH WKHRU\ JURXS WKHRU\ HWFf LV VLPSOLILHG E\ WKH IDFW WKDW DQ\ VXUMHFWLRQ LV D UHJXODU HSLPRUSKLVP )RU WKH FDWHJRULHV LQWURGXFHG KHUH WKH VLWXDWLRQ LV QRW VR VLPSOH ([DPSOH VKRZV WKDW D MMRLQ SUHVHUYLQJ VXUMHFWLYH LPDJH RI D MFRPSOHWH SRVHW QHHG QRW EH MFRPSOHWH 7KH UHVXOWV RI &KDSWHU SUHGDWH WKH RWKHU UHVXOWV SUHVHQWHG KHUH +HUUOLFK >@ FRQWDLQV D GHWDLOHG H[DPLQDWLRQ RI UHIOHFWLRQV DQG FRUHIOHFWLRQVf LQ FDWHJRULHV RI WRSRORJLFDO VSDFHV 7KLV GLVVHUWDWLRQ DLPHG WR JHQHUDOL]H UHVXOWV VXPPDUL]HG LQ +HUUOLFK >@ IRU MPf VSDFHV $ M PfVSDFH FRQVLVWV RI DQ XQGHUO\LQJ VHW DQG D IDPLO\ RI fRSHQf VXEVHWV ZKLFK DUH
PAGE 13
FORVHG XQGHU MXQLRQV DQG PLQWHUVHFWLRQV &RQWLQXRXV PDSV RI M PfVSDFHV DUH IXQFWLRQV VXFK WKDW SUHLPDJHV RI RSHQ VHWV DUH RSHQ 7KH LQLWLDO DLP ZDV WR ILQG UHIOHFWLRQV DQG FRUHIOHFWLRQV RI WKH FDWHJRU\ RI M PfVSDFHV REH\LQJ D 7RVW\OH VHSDUDWLRQ D[LRPf DQG VWXG\ KRZ WKH H[LVWHQFH DQG SURSHUWLHV RI UHIOHFWLRQV DQG FRUHIOHFWLRQV YDULHG GHSHQGLQJ XSRQ WKH VXEVHW V\VWHPV M DQG P $Q REYLRXV SUHUHTXLVLWH WR VXFK D SURMHFW LV NQRZOHGJH RI IDFWRUL]DWLRQV RI FRQWLQXRXV PDSV 7KH FKDSWHU FRQWDLQV D GHVFULSWLRQ RI M PfVXEVSDFHV DQG M PfTXRWLHQWV ,Q DGGLn WLRQ 6HFWLRQ GHVFULEHV D UHIOHFWLRQ RI M PfVSDFHV WKDW FRUUHVSRQGV WR WKH f UHIOHFWLRQ RI WRSRORJLFDO VSDFHV /DVWO\ 6HFWLRQ GHVFULEHV HSLFRPSOHWH M PfVSDFHV
PAGE 14
&+$37(5 35,0(5 21 &$7(*25,(6 $1' 326(76 7KH WH[W DVVXPHV D IDPLOLDULW\ ZLWK WKH WKHRU\ RI VHWV W\SLFDOO\ XVHG LQ PDWKHPDWLFDO DUJXPHQWV 6R IDPLOLDU FRQVWUXFWLRQV XQLRQV LQWHUVHFWLRQV FDUWHVLDQ SURGXFWV TXRWLHQWV E\ HTXLYDOHQFH FODVVHV IXQFWLRQV =RUQfV /HPPD DQG WUDQVILQLWH LQGXFWLRQ DUH XVHG ZLWKRXW IXUWKHU FRPPHQW 6HH +DOPRV >@ LI WKLV EDFNJURXQG LV QHHGHGf $ EDVLF IDPLODULW\ ZLWK JHQHUDO WRSRORJ\ LV KHOSIXO $OVR VRPH FRPIRUW ZLWK FDWHJRU\ WKHRU\ LV DVVXPHG 1DPHO\ WKH UHDGHU FDQ ILOO LQ WKH EODQNV LQ WKH IROORZLQJ LQIRUPDO GHILQLWLRQV f $ FDWHJRU\ $ FRQVLVWV RI D FODVV RI REMHFWV 2EM$f DQG PDSV 0DS$f VXFK WKDW HDFK REMHFW KDV DQ LGHQWLW\ PDS DQG WKHUH LV DQ DVVRFLDWLYH QRWLRQ RI FRPSRVLWLRQ RI PDSV f 7KH VHW RI $PDSV IURP $? WR $ LV GHQRWHG $$L $f f $ IXQFWRU ) $ fÂ§! % DVVLJQV HDFK $ 2EM$f DQ REMHFW )$f f 2EM%f DQG HDFK $PDS $M fÂ§} $ D %PDS )I )$f fÂ§! )$f 7KH DVVLJQPHQW UHVSHFWV FRPSRVLWLRQ DQG LGHQWLW\ DUURZV f ,I $ LV D FDWHJRU\ $RS LV WKH FDWHJRU\ ZLWK WKH VDPH REMHFWV DV $ EXW DOO DUURZV UHYHUVHG )RU D FDWHJRU\ WKHRUHWLF FRQFHSW & WKH GXDO LV REWDLQHG E\ DSSO\LQJ & WR $rS f $ FRQWUDYDULDQW IXQFWRU $ fÂ§ % LV D IXQFWRU $ fÂ§ %RS f 'LDJUDPV DUH XVHG WR GLVSOD\ WKH EHKDYLRU RI D FROOHFWLRQ RI PDSV D GLDJUDP FRPPXWHV LI DQ\ FRPSRVLWHV ZLWK WKH VDPH GRPDLQ DQG FRGRPDLQ DUH HTXDO )RU H[DPSOH WKH
PAGE 15
GLDJUDP : I K < FRPPXWHV LI DQG RQO\ LI JI fÂ§ LK L 5HFDOO WKH IROORZLQJ SURSHUWLHV RI IXQFWRUV 'HILQLWLRQ /HW ) $ fÂ§! % EH D IXQFWRU )RU HDFK $?$ ( $ ) JLYHV D IXQFWLRQ IURP WKH KRPVHW $$L $"f LQWR U%)$L )$f E\ $W $f a )I $[ $f ,I IRU HDFK $L DQG $ WKLV PDS LV RQWR WKHQ ) LV VDLG WR EH IXOO ,I IRU HDFK $? DQG $ WKLV PDS LV RQHWRRQH WKHQ ) LV VDLG WR EH IDLWKIXO $ IXOO VXEFDWHJRU\ RI D FDWHJRU\ $ LV D FDWHJRU\ % VXFK WKDW 2EM%f & 2EM$f DQG DOO $ fÂ§! % ZLWK $ % 2EM%f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fLQMHFWLYHf DQG fVXUMHFWLYHf LQ WKH FDWHJRU\ 6HW $ PDS $? fÂ§! $ LV HSL DND HSLF LQ QRXQ IRUP DQ HSLPRUSKLVPf LI ZKHQHYHU J DQG K DUH PDSV $ fÂ§ $ VXFK WKDW JI KI WKHQ J K
PAGE 16
$ PDS $ fÂ§! $ LV PRQR DND PRQLF LQ QRXQ IRUP D PRQRPRUSKLVPf LI ZKHQHYHU J DQG K DUH PDSV $? fÂ§! $ DQG VXFK WKDW IJ IK WKHQ J fÂ§ K 2QH PD\ YHULI\ WKDW D FRPSRVLWLRQ RI HSLPRUSKLVPV UHVS PRQRPRUSKLVPVf LV HSL UHVS PRQRf 0RUHRYHU LI I DE LV HSL UHVS PRQRf WKHQ D LV DOVR HSL UHVS E LV DOVR PRQRf 'HILQLWLRQ $ PDS $ fÂ§! % LV DQ LVRPRUSKLVP LI WKHUH LV J % fÂ§r $ VXFK WKDW LGA JI DQG LG% IJ ,Q PRVW FDWHJRULHV RI fVHWV ZLWK VWUXFWXUHf D PDS LV PRQR LI DQG RQO\ LI LW LV LQMHFWLYH VXUMHFWLYH PDSV DUH HSL EXW HSLPRUSKLVPV PD\ QRW EH VXUMHFWLYH ([DPSOH &RQVLGHU WI$E WKH FDWHJRU\ RI WRUVLRQIUHH DEHOLDQ JURXSV LH DEHOLDQ JURXSV VXFK WKDW QD D IRU DQ\ QDWXUDO QXPEHU Q DQG JURXS HOHPHQW D WRJHWKHU ZLWK JURXS KRPRPRUSKLVPV 7KH LQFOXVLRQ L = fÂ§} 4 RI WKH LQWHJHUV LQ WKH UDWLRQDO QXPEHUV LV HSL EXW QRW RQWR ,Q FDWHJRULHV RI fVHWV ZLWK UHODWLRQDO VWUXFWXUHf ELMHFWLYH PDSV DUH QRW QHFHVVDULO\ LVRPRUSKLVPV ([DPSOH &RQVLGHU 7RS WKH FDWHJRU\ RI WRSRORJLFDO VSDFHV DQG FRQWLQXRXV PDSV 7KH LGHQWLW\ IXQFWLRQ L fÂ§! 5 IURP WKH UHDOV ZLWK GLVFUHWH WRSRORJ\f WR WKH UHDOV ZLWK WKH XVXDO WRSRORJ\f LV D FRQWLQXRXV ELMHFWLRQ +RZHYHU WKH LQYHUVH IXQFWLRQ LB LV QRW FRQWLQXRXV ([DPSOH &RQVLGHU 3 WKH FDWHJRU\ RI SDUWLDOO\ RUGHUHG VHWV DQG RUGHU SUHVHUYLQJ PDSV (LWKHU ELMHFWLRQ If IURP WKH WULYLDOO\ RUGHUHG VHW ZLWK WZR HOHPHQWV WR WKH FKDLQ ZLWK WZR HOHPHQWV LV RUGHU SUHVHUYLQJ %XW WKH LQYHUVH IXQFWLRQ !B LV QRW RUGHU SUHVHUYLQJ )RU IXUWKHU GLVFXVVLRQ DQG PRUH H[DPSOHV RI HSLPRUSKLVPV DQG PRQRPRUSKLVPV VHH +HUUOLFK DQG 6WUHFNHU > 6HFWLRQ @ DQG %RUFHX[ > 9ROXPH 6HFWLRQV DQG @
PAGE 17
'HILQLWLRQ DQG 5HPDUNV &RQVLGHU D SDLU RI PDSV $L fÂ§! $ $ PDS L $T fÂ§Â‘! $? ULJKWLGHQWLILHV I DQG J LI IL JL $ PDS L $T fÂ§! $? LV FDOOHG WKH HTXDOL]HU RI DQG J LI HTOf L ULJKWLGHQWLILHV DQG J DQG HTf L KDV WKH IHDWXUH WKDW ZKHQHYHU M % fÂ§! $? ULJKWLGHQWLILHV DQG J WKHUH LV D XQLTXH PDS H %R fÂ§! $T VXFK WKDW M fÂ§ LH 7KH GHILQLWH DUWLFOH LV XVHG IRU HTXDOL]HUV EHFDXVH HTf LPSOLHV WKDW LI L $R fÂ§ $ DQG Ln $n fÂ§! $ D[H HTXDOL]HUV IRU DQG J WKHQ WKHUH LV DQ LVRPRUSKLVP M $T fÂ§A $T VXFK WKDW Ln fÂ§ LM 1RWDWLRQ L HTIJf ,I WKHUH DUH DQG J VXFK WKDW L $ fÂ§! $? LV WKH HTXDOL]HU RI DQG J WKHQ L LV UHJXODU PRQR 5HJXODU PRQRPRUSKLVPV DUH PRQRPRUSKLVPV ,I LV HSL DQG UHJXODU PRQR WKHQ LV DQ LVRPRUSKLVP )RU SURRIV RI WKH DVVHUWLRQV LQ WKLV SDUDJUDSK DQG D GLVFXVVLRQ RI H[DPSOHV VHH %RUFHX[ > 9ROXPH 6HFWLRQ @ RU +HUUOLFK DQG 6WUHFNHU > 6HFWLRQ @ 7KH GHILQLWLRQV RI fFRHTXDOL]HUf DQG fUHJXODU HSLf DUH GXDO WR fHTXDOL]HUf DQG fUHJXODU PRQRf EXW DUH UHSHDWHG IRU HPSKDVLV H $ fÂ§! $ OHIWLGHQWLILHV I DQG J LI HI fÂ§ HJ $ PDS H $ fÂ§} $ LV WKH FRHTXDOL]HU RI DQG J LI FRHTOf H OHIWLGHQWLILHV DQG J DQG FRHTf H KDV WKH IHDWXUH WKDW ZKHQHYHU G $ fÂ§}f % OHIWLGHQWLILHV DQG J WKHUH LV D XQLTXH PDS F $ fÂ§} % VXFK WKDW H FG 1RWDWLRQ H FRHT f ,I H LV WKH FRHTXDOL]HU RI VRPH SDLU RI PDSV WKHQ H LV FDOOHG UHJXODU HSL 7KH GXDOV RI DOO EDVLF SURSHUWLHV RI UHJXODU PRQRPRUSKLVPV KROG IRU UHJXODU HSLPRUSKLVPV 'HILQLWLRQ $Q HSLPRUSKLVP LV H[WUHPDO LI ZKHQHYHU I JK DQG J LV PRQR WKHQ J LV DQ LVRPRUSKLVP 'XDOO\ D PRQRPRUSKLVP LV H[WUHPDO LI ZKHQHYHU JK DQG K LV HSL WKHQ K LV DQ LVRPRUSKLVP )RU PRUH GHWDLOHG GLVFXVVLRQV VHH %RUFHX[ > 9ROXPH 6HFWLRQ @ DQG +HUUOLFK DQG 6WUHFNHU > 6HFWLRQ @
PAGE 18
'HILQLWLRQ 7KH PDS $ fÂ§ % LV VSOLW PRQR LI WKHUH H[LVWV J % fÂ§r $ VXFK WKDW LGÂ JI 7KH PDS J % fÂ§! $ LV VSOLW HSL LI WKHUH H[LVWV $ fÂ§! % VXFK WKDW LGA JI )RU PRUH LQIRUPDWLRQ VHH +HUUOLFK DQG 6WUHFNHU > 6HFWLRQ @ 1RWH WKDW RXU WHUPLQRORJ\ GLIIHUV VOLJKWO\ IURP WKH UHIHUHQFH fVHFWLRQf DQG fVSOLW PRQRf DUH V\QRQ\PV DV DUH fUHWUDFWLRQf DQG fVSOLW HSLf /HPPD )RU $ HLWKHU fPRQRf RU fHSLf FRQVLGHU WKH IROORZLQJ VWDWHPHQWV I LV VSOLW$ I LV UHJXODU$ I LV H[WUHPDO$ I LV $ 7KH LPSOLFDWLRQV DOZD\V KROG 1RQH RI WKH FRQYHUVHV JHQHUDOO\ KROG )RU SURRI VHH f +HUUOLFK DQG 6WUHFNHU > @ f +HUUOLFK DQG 6WUHFNHU > @ RU %RUFHX[ > 9ROXPH f@ f KROGV E\ GHILQLWLRQ %RXQGV 5HFDOO WKDW D SDUWLDO RUGHU RQ D VHW $ LV D UHODWLRQ VDWLVI\LQJ SROf )RU DOO D $ D D SRf :KHQHYHU D E DQG E F D F SRf :KHQHYHU D E DQG E D D E $ SUHRUGHU LV D UHODWLRQ WKDW VDWLVILHV SROf DQG SRf ,I A LV D SUHRUGHU RQ $ GHILQH DQ HTXLYDOHQFH UHODWLRQ a RQ $ E\ D a E D AE DQG E Â‘ D 7KH UHODWLRQ LV D SDUWLDO RUGHU RQ $ a $ VHW $ ZLWK D SDUWLDO RUGHU UHVS SUHRUGHUf LV FDOOHG D SDUWLDOO\ RUGHUHG VHW UHVS SUHRUGHUHG VHWf
PAGE 19
'HILQLWLRQ DQG 5HPDUNV /HW $ EH D SUHRUGHUHG VHW 'HILQH WKH XS DQG GRZQ FORVXUHV RI [ $ E\ [ ^D f $ D Â‘ [` DQG f? [ ^D f $ [ Â‘ D` 0RUH JHQHUDOO\ LI & $ GHILQH fÂ§ 8^_ LL*6` ^Df\OV*6D;V` DQG 8^I LL*6` ^D$Lf6_6A D` ,I & $ DQ XSSHU UHVS ORZHUf ERXQG IRU LV D e $ VXFK WKDW IRU DOO V V Â‘ D UHVS D Â‘ Vf 8VH WKH IROORZLQJ QRWDWLRQ IRU WKH VHW RI XSSHU ERXQGV RI f Q^I [ [ f ` ^D $ 9V V Â‘ D` DQG VLPLODU QRWDWLRQ IRU WKH VHW RI ORZHU ERXQGV f Q^M [ [ f ` D $ 9V D Â‘ V` ,I D $ 6 & $ DQG D f WKHQ D LV D OHDVW XSSHU ERXQG DND MRLQ DND VXSUHPXP RI $Q HTXLYDOHQW ZD\ WR VD\ WKLV LV D [ 9 V f V [ ,I ERWK D DQG Dn DUH VXSUHPD RI WKHQ DDn f f 7KXV D a Dn ,Q D SDUWLDOO\ RUGHUHG VHW WKH XQLTXHf VXSUHPXP RI LV ZULWWHQ ? ,I D f $ 6 & $ DQG M D fÂ§ f WKHQ D LV D JUHDWHVW ORZHU ERXQG DND PHHW DND LQILPXP RI $Q HTXLYDOHQW ZD\ WR VD\ D LV D OHDVW XSSHU ERXQG LV [ D 9 V f [ V
PAGE 20
,Q D SDUWLDOO\ RUGHUHG VHW WKH XQLTXHf LQILPXP RI LV GHQRWHG ? ,Q XVLQJ WKLV QRWLRQ ZLWK D SUHRUGHUHG VHW $ f RQH UHIHUV WR WKH DVVRFLDWHG SRVHW $ )RU H[DPSOH ?6 WKHQ GHQRWHV WKH HTXLYDOHQFH FODVV FRQWDLQLQJ DOO VXSUHPD RI f ,I D 9 RU D I? D LV DQ RSWLPXP ERXQG IRU 7KH IROORZLQJ DUH HOHPHQWDU\ SURSHUWLHV RI WKH ERXQGRSHUDWRUV DQG RSWLPXP ERXQGV /HPPD )RU DQ\ SUHRUGHUHG VHW $ LQ SDUWLFXODU DQ\ SRVHWf WKH IROORZLQJ SURSHUWLHV KROG 8VH ,% WR GHQRWH HLWKHU RSHUDWRU p RU % )RU [\ f $ [ A \ > [ & M \ @ \ &@ [ )RU 67 & $ 6 &7 %7f & f )RU 6& $ f ,_ f f 7 f )RU MfM D IDPLO\ RI VXEVHWV RI $ %XÂf Q%Âf ,II $A% LV RUGHU SUHVHUYLQJ WKHQ %ff & %Vff 'HILQLWLRQ DQG 5HPDUNV /HW $ DQG % EH SUHRUGHUHG VHWV $ *DORLV FRQQHFWLRQ EHWZHHQ $ DQG % LV D SDLU RI IXQFWLRQV $ fÂ§! % DQG J % fÂ§! $ VXFK WKDW JHOf DQG J DUH RUGHU UHYHUVLQJ JFf )RU DOO D f $ DQG E % D A JI^Dff DQG E Â‘ IJEff %HORZ EDVLF SURSHUWLHV RI *DORLV FRQQHFWLRQV DUH OLVWHG 6\PPHWU\ LQ WKH GHILQLWLRQ DOORZV V\PPHWU\ LQ SURRIV )RU DQ\ WUXH VWDWHPHQW DERXW *DORLV FRQQHFWLRQV WKHQ WKH VWDWHPHQW REWDLQHG E\ VZLWFKLQJ WKH UROHV RI DQG J DORQJ ZLWK $ DQG % LV DOVR D WUXH VWDWHPHQW DERXW *DORLV FRQQHFWLRQV 1RWH WKDW DQG J UHVSHFW a 7KH QRWDWLRQ a J PHDQV fIRU DOO D f $ IDf a JEf Â‘ JIJÂ‘ 6XSSRVH D e $ %\ JFf D U IJ^Dff DQG JDf Â‘ JIJDfff 8VLQJ D A IJ^Dff DQG JHOf JIJDfff A JDf
PAGE 21
I$ af LV GXDOO\ RUGHU LVRPRUSKLF WR ^ ( % E a IJEf` 7KLV IROORZV IURP WKH SUHFHGLQJ VWDWHPHQW 6LQFH a IJI DQ\ PHPEHU IDf ( I$ af LV HTXLYDOHQW WR IJIDfff 0RUHRYHU LI E ( % DQG E a IJEff WKHQ E ( I$ af EHFDXVH E a IJEff D J^Ef E Â‘ Df LI D A JEf WKHQ E Â‘ IJEff I^Df 7KH FRQYHUVH LV SURYHG VLPLODUO\ )RU DQ\ 6 & $ %I6f AB%f &DOFXODWH [ ( %I6ff 9VJL6Vf A 9V f 6V A J[f r J[fH%6f [ f SBpf ,I $ DQG % DUH SRVHWV WKH SUHYLRXV LWHP LPSOLHV ?f I? 6f ,W LV KHOSIXO WR UHSKUDVH FRQGLWLRQ IRU SRVHWV XJEf LV WKH ODUJHVW D ZLWK E Dff ,Q V\PEROV Ef ?^D ( $E Df` %DVLF LQIRUPDWLRQ DERXW *DORLV FRQQHFWLRQV KDV EHHQ ZHOONQRZQ VLQFHWKH V 5DQH\ >@ FRQWDLQV D ELEOLRJUDSK\ RI WKLV HDUO\ OLWHUDWXUH 7KH EDVLF SURSHUWLHV DQG GHIn LQLWLRQ DUH OLVWHG LQ +HUUOLFK DQG 6WUHFNHU > ([HUFLVH 4@ 7KH SDUWLFXODU VXPPDU\ KHUH LV E\ WKH DXWKRU 7KH FRQFHSW RI D *DORLV FRQQHFWLRQ LV V\PPHWULF DQG DOORZV RQH WR WUDQVIHU D JUHDW GHDO RI LQIRUPDWLRQ EHWZHHQ SUHRUGHUHG VHWV +RZHYHU WKH IDFW WKDW WKH IXQFWLRQV LQYROYHG DUH RUGHU UHYHUVLQJ LV VRPHWLPHV LQFRQYHQLHQW 7KH FRQFHSW RI DQ DGMRLQW FRQQHFWLRQ EHWZHHQ SRVHWV LV REWDLQHG E\ IRUPDOO\ UHYHUVLQJ RQH RI WKH SRVHWV LQYROYHG
PAGE 22
'HILQLWLRQ DQG 5HPDUNV 6XSSRVH $ DQG % DUH SUHRUGHUHG VHWV $ SDLU $ fÂ§! % J % fÂ§! $ RI RUGHU SUHVHUYLQJ IXQFWLRQV LV DQ DGMRLQW FRQQHFWLRQ EHWZHHQ $ DQG % LI DGOf )RU DOO D $ D Â‘ JIDff DGf )RU DOO E H % IJEff E (DFK EDVLF SURSHUW\ IRU *DORLV FRQQHFWLRQV FRUUHVSRQGV WR D EDVLF SURSHUW\ RI DGMRLQW FRQQHFWLRQV 7KH EDVLF SURSHUWLHV RI DGMRLQW FRQQHFWLRQV DUH OLVWHG EHORZ SURRIV DUH RPLWWHG f IJI DQG J a JIJ I$ af ^E % E a JI^Eff` DQG J^% af ^D $ D a IJDff` Df A E Â‘ !f D ; JEf )RU DQ\ 6& $ %J6f fÂ§ B%f DQG %6f Ja06f ,I $ DQG % DUH SRVHWV I? 6f I? I6f DQG J?6f ? J6f %HFDXVH RI WKH DV\PPHWU\ EHWZHHQ DQG J DQG SURSHUW\ LV FDOOHG WKH OHIW DGMRLQW DQG J WKH ULJKW DGMRLQW $JDLQ WKHUH LV DQ LQWHUSUHWDWLRQ RI LQ ZRUGV fDf LV WKH VPDOOHVW E VXFK WKDW D Â‘ J^Ef? JEf LV WKH ODUJHVW D VXFK WKDW Df A f 7KH GHILQLWLRQ DQG EDVLF SURSHUWLHV RI DGMRLQW FRQQHFWLRQV DUH fIRONORUHf )RU DQRWKHU GLVFXVVLRQ RI WKHP VHH RKQVWRQH > &KDSWHU 3DUDJUDSK @ 'HILQLWLRQ $ SRVHW $ LV FRPSOHWH LI HDFK VXEVHW 6& $ LQFOXGLQJ WKH HPSW\ VHWf KDV D VXSUHPXP 6LQFH WKH VXSUHPXP RI VHW RI ORZHU ERXQGV IRU 6 LV D ORZHU ERXQG IRU 6 $ LV FRPSOHWH LI DQG RQO\ LI HDFK VXEVHW 6 & $ LQFOXGLQJ WKH HPSW\ VHWf KDV DQ LQILPXP 7KHUH LV D FULWHULRQ IRU WHVWLQJ ZKHQ D JLYHQ RUGHU SUHVHUYLQJ UHVS UHYHUVLQJf PDS EHWZHHQ SRVHWV LV SDUW RI DQ DGMRLQW UHVS *DORLVf FRQQHFWLRQ 7KHRUHP $GMRLQW ([LVWHQFH 6XSSRVH $ DQG % DUH SRVHWV
PAGE 23
6XSSRVH I $ fÂ§! % LV RUGHU SUHVHUYLQJ DQG $ LV FRPSOHWH 7KHQ I LV D OHIW DGMRLQW LI DQG RQO\ LI I?6f ?I I6f IRU DOO 6 & $ 6XSSRVH J $ fÂ§! % LV RUGHU SUHVHUYLQJ DQG $ LV FRPSOHWH 7KHQ J LV D ULJKW DGMRLQW LI DQG RQO\ LI JI? 6f ? J6f IRU DOO 6 & $ 6XSSRVH I $ fÂ§! % LV RUGHU UHYHUVLQJ DQG $ LV FRPSOHWH 7KHQ I LV SDUW RI D *DORLV FRQQHFWLRQ LI DQG RQO\ LI I?6f I? I6f IRU DOO 6& $ 7KLV WKHRUHP LV WKH SRVHW YHUVLRQ RI WKH DGMRLQW IXQFWRU WKHRUHP )RONORUH VHH RKQVWRQH > &KDSWHU 6HFWLRQ 3DUDJUDSK @ RU R\DO DQG 7LHUQH\ > &KDSWHU 6HFWLRQ @ 3URRI $ SURRI IRU IROORZV WKH RWKHU LWHPV DUH VLPLODU 'HILQH J % fÂ§! $ E\ Â3f ?^DH$ IDf E` 6LQFH SUHVHUYHV VXSUHPD IJEff E 7KXV JEf LV WKH ODUJHVW D VXFK WKDW Df E Â’ ,Q DGMRLQW UHVS *DORLVf FRQQHFWLRQV Df DQG JEf FDQ EH GHILQHG DV VXSUHPD RU LQWLPD 7KHUH LV D VRUW RI FRQYHUVH WR WKLV RQH FDQ YLHZ WKH VXSUHPXP DV DQ DGMRLQW WR D SDUWLFXODU PDS 6HH /HPPD 1DWXUDO 7UDQVIRUPDWLRQV 0DSV FRPSDUH REMHFWV LQ D FDWHJRU\ IXQFWRUV FRPSDUH FDWHJRULHV DQG QDWXUDO WUDQVn IRUPDWLRQV FRPSDUH IXQFWRUV 7KLV VHFWLRQ FRQWDLQV QR QHZ UHVXOWV UHVXOWV DQG VRPH H[SRn VLWLRQ DUH SDUDSKUDVHG IURP +HUUOLFK DQG 6WUHFNHU > 6HFWLRQ @ 'HILQLWLRQ /HW ) $ fÂ§! EH IXQFWRUV $ QDWXUDO WUDQVIRUPDWLRQ D ) fÂ§! LV D UXOH ZKLFK DVVLJQV D PDS D$ )$ fÂ§! *$ WR HDFK $ f $ VXFK WKDW LI $ fÂ§! % LV DQ $PDS WKH IROORZLQJ GLDJUDP FRPPXWHV )I )$ fÂ§!)% U?$ U@% *$ *I *%
PAGE 24
&RQVWUXFWLRQ /HW ) + ; fÂ§! $ EH IXQFWRUV D ) fÂ§! DQG I fÂ§! + QDWXUDO WUDQVIRUPDWLRQV 7KH VLWXDWLRQ LV GUDZQ LQ WKH GLDJUDP EHORZf ) 4 + n 7KHQ WKH DVVLJQPHQW Df$ $fD$f LV D QDWXUDO WUDQVIRUPDWLRQ 3URRI /HW $ fÂ§! % EH D PDS 6LQFH D DQG DUH QDWXUDO HDFK VTXDUH EHORZ FRPPXWHV )$ )I Â‘r)% *$ *I *% +$ +I +% 7KHUHIRUH WKH RXWVLGH UHFWDQJOH FRPPXWHV KHQFH D LV QDWXUDO Â’ &DOO D WKH YHUWLFDO FRPSRVLWLRQ RI D DQG &RQVWUXFWLRQ ,I ) $ fÂ§! % DQG + fÂ§r DUH IXQFWRUV DQG D fÂ§! + LV D QDWXUDO WUDQVIRUPDWLRQ WKHQ D)f$ D)$f LV D QDWXUDO WUDQVIRUPDWLRQ D) *) fÂ§! +) &RQVWUXFWLRQ ,I ) + fÂ§! & DQG + $ fÂ§r + DUH IXQFWRUV DQG D fÂ§! + LV D QDWXUDO WUDQVIRUPDWLRQ WKHQ )Df$ )D$f LV D QDWXUDO WUDQVIRUPDWLRQ )D )* fÂ§! )+ &RQVWUXFWLRQ 6XSSRVH ) $ fÂ§! % DQG +7!fÂ§r* DUH IXQFWRUV DQG D ) fÂ§rÂ‘ + fÂ§! DUH QDWXUDO WUDQVIRUPDWLRQV 7KH VLWXDWLRQ LV GUDZQ EHORZf ) + H
PAGE 25
7KHQ IRU HDFK REMHFW $ f 2EM$f WKH IROORZLQJ VTXDUH FRPPXWHV )+$ )3$ )$ D+$ D$ *+$fÂ§*$ !*$ 7KH DVVLJQPHQW Â‘ Df$ D$f)$f fÂ§ *I$fD+$f LV D QDWXUDO WUDQVIRUPDWLRQ 3URRI 7KH VTXDUH FRPPXWHV EHFDXVH D LV D QDWXUDO WUDQVIRUPDWLRQ WR VHH WKLV RQH DSSOLHV WKH QDWXUDO SURSHUW\ DW WKH PDS $ +$ fÂ§ $ 7R SURYH WKH QDWXUDO SURSHUW\ RI D ZH XVH WKH VTXDUHV ZKLFK GHILQH D $ DQG D% &RPSDUH WKH FRUQHUV RI WKH VTXDUHV XVLQJ PDSV REWDLQHG IURP E\ DSSOLFDWLRQ RI WKH IXQFWRUV )+ *+ )DQG *7KH UHVXOWLQJ FRPPXWDWLYH FXEH VKRZV D LV D QDWXUDO WUDQVIRUPDWLRQ Â’ &DOO Â‘ D WKH KRUL]RQWDO FRPSRVLWLRQ RI D DQG 'HILQLWLRQ $ QDWXUDO HTXLYDOHQFH D ) fÂ§r LV D QDWXUDO WUDQVIRUPDWLRQ VXFK WKDW HDFK FRPSRQHQW D$ LV DQ LVRPRUSKLVP *LYHQ IXQFWRUV ) DQG DUH QDWXUDOO\ HTXLYDOHQW LI WKHUH LV D QDWXUDO HTXLYDOHQFH D ) fÂ§ &DWHJRULHV $ DQG % DUH HTXLYDOHQW LI WKHUH DUH IXQFWRUV ) $ fÂ§! 7 DQG 7! fÂ§! $ VXFK WKDW )* LV QDWXUDOO\ HTXLYDOHQW WR LGV DQG *) LV QDWXUDOO\ HTXLYDOHQW WR LGA $ DQG p DUH GXDO LI WKHUH H[LVW FRQWUDYDULDQW IXQFWRUV ) $ fÂ§ % DQG % fÂ§! $ VXFK WKDW )* LV QDWXUDOO\ HTXLYDOHQW WR LGV DQG *) LV QDWXUDOO\ HTXLYDOHQW WR LGA &Rf/LPLWV )RU LQWXLWLRQ LW LV XVHIXO WR YLHZ WKH REMHFWV RI D FDWHJRU\ DV D SUHRUGHUHG FODVV ZLWK WKH SUHRUGHU $ % LI DQG RQO\ LI WKHUH LV D PDS I $ fÂ§r % &DWHJRULHV DUH PRUH FRPSOH[ WKDQ SUHRUGHUHG FODVVHV EHFDXVH WKHUH PD\ EH PDQ\ PDSV ZKLFK PDQLIHVW $ % 'HILQLWLRQ DQG 5HPDUNV /HW $ EH D FDWHJRU\ $ GLDJUDP LQ $ DND D VPDOO VXEFDWHJRU\ RI $f LV D VHW RI REMHFWV 2EM'f & 2EM$f DQG PDSV 0DS'f & 0DS$f
PAGE 26
EHWZHHQ WKHP IRU WHFKQLFDO UHDVRQV RQH UHTXLUHV WKDW LI IJ 0DS'f DQG IJ LV GHILQHG WKHQ IJ 0DS'f DQG WKDW IRU DOO $ 2EM'f LG$ f 0DS'f 7KLV GHILQLWLRQ RI D GLDJUDP LV HTXLYDOHQW WR EXW GLIIHUV IURP WKH RQH LQ WKH PDMRULW\ RI WKH OLWHUDWXUH VHH 0DF/DQH >@ %RUFHX[ >@ RU +HUUOLFK DQG 6WUHFNHU >@ IRU WKH VWDQGDUG GHILQLWLRQ 6XSSRVH LV D GLDJUDP $ VRXUFH 6 ^VA 6 fÂ§} $ $ 2EMeff`f IRU FRQVLVWV RI 6 2EM$f DQG PDSV V$ 6 $ VXFK WKDW LI $ $n LV D PDS LQ IV$ V$! $ VRXUFH IRU LV D fORZHU ERXQGf FRPSDWLEOH ZLWK DOO PDSV LQ $ VRXUFH / ^e$ $ 2EM=ff`f IRU LV WKH OLPLW RI LI ZKHQHYHU $ 2EM'f`f LV D VRXUFH WKHUH LV D XQLTXH PDS F 6 fÂ§ / VXFK WKDW IRU HDFK $ 2EM'f V$ fÂ§ L$F 'XDOO\ GHILQH D VLQN RU FRVRXUFH IRU WR EH DQ REMHFW 6 WRJHWKHU ZLWK PDSV L$ $ fÂ§! 6 IRU $ 2EM'ff VXFK WKDW IRU HDFK I $fÂ§! $n LQ 0DS'f L$ L$nI $ FROLPLW & ^cMOD $ 2EM'f`f IRU LV D VLQN IRU VXFK WKDW LI ^P $ 2EM'f`f LV DQ\ VLQN WKHUH LV D XQLTXH PDS F & fÂ§! 6 VXFK WKDW IRU HDFK $ 2EM'f +D FL$ 1RWH WKDW OLPLWV DQG FROLPLWV RI DUH XQLTXH XS WR FRPSDWLEOH LVRPRUSKLVP 7R SURYH WKLV IRU OLPLWVf VXSSRVH /e$f DQG /nen$f DUH ERWK OLPLWV IRU 7KH OLPLW SURSHUW\ JXDUDQWHHV WKDW WKHUH DUH XQLTXH FRPSDWLEOH PDSV F / fÂ§! /n DQG G /n fÂ§! / %XW FG /n fÂ§! /n DQG GF / / DUH ERWK FRPSDWLEOH PDSV %\ XQLTXHQHVV FG LGne DQG GF LG/ )URP QRZ RQ ZH VKDOO XVH WKH GHILQLWH DUWLFOH ZKHQ ZULWLQJ DERXW OLPLWV DQG FROLPLWV ([DPSOH /HW XV FRQVLGHU 6HW ,I ; LV D VHW RI VHWV IRUP D GLDJUDP FRQWDLQLQJ DOO PHPEHUV RI ; DQG QR IXQFWLRQV 7KH OLPLW IRU WKLV GLDJUDP LV WKH &DUWHVLDQ SURGXFW Q[U[f 5HFDOO WKDW c4 ; FRQWDLQV DOO IXQFWLRQV ; fÂ§! _;ZLWK WKH IHDWXUH WKDW I;f ; IRU DOO ; ; 6XFK DUH FDOOHG FKRLFH IXQFWLRQV EHFDXVH WKH\ FKRRVH RQH PHPEHU RI HDFK ; ; 7KH IXQFWLRQV Q[ n:; ; DUH GHILQHG E\ Q[If I;f
PAGE 27
,I WKHUH LV D VRXUFH 6 ^Vr 6 ; ; f ;`f WKHUH LV D XQLTXH IXQFWLRQ F 6 I@ A PDNLQJ FRPPXWH IRU HDFK ; ,W LV GHILQHG E\ FDf;f fÂ§ V[Df 0RWLYDWHG E\ WKLV H[DPSOH RQH GHILQHV WKH FDWHJRU\WKHRUHWLF SURGXFW RI D VHW ; RI REMHFWV LQ DQ\ FDWHJRU\f DV WKH OLPLW RI WKH GLDJUDP FRQWDLQLQJ DOO PHPEHUV RI ; DQG QR PDSV ,Q PRVW FDWHJRULHV RI fVHWV ZLWK VWUXFWXUHf SURGXFWV H[LVW DQGf ORRN OLNH SURGXFWV LQ 6HW ZLWK VXLWDEOH VWUXFWXUH DGGHG ,Q 6HW WKH FROLPLW RI WKH GLDJUDP FRQWDLQLQJ DOO PHPEHUV RI ; DQG QR IXQFWLRQV LV WKH GLVMRLQW XQLRQ @; ZLWK LQFOXVLRQ PDSV S[ nÂ‘ ; fÂ§L ; 7KH FROLPLW SURSHUW\ LV VDWLVILHG E\ ; S[f EHFDXVH LI WKHUH LV D VLQN & L[f WKH IXQFWLRQ F @; fÂ§ & GHILQHG E\ F[f L[[f IRU WKH XQLTXH ; f ; FRQWDLQLQJ [ LV WKH XQLTXH FRPSDWLEOH PDS ,Q FDWHJRULHV RWKHU WKDQ 6HW FRSURGXFWV DUH GHILQHG LGHQWLFDOO\ 8VXDOO\ WKH FRSURGn XFW RI D VHW ; RI AREMHFWV LV fWKH r$REMHFW IUHHO\ JHQHUDWHG E\ ;f ([DPSOH 7KH >FR@HTXDOL]HU RI J $ fÂ§! % LV WKH >FR@OLPLW RI WKH GLDJUDP FRQWDLQLQJ REMHFWV $ DQG % DORQJ ZLWK PDSV DQG J 7KH QRWDWLRQ IRU HTXDOL]HUV LV FXVWRPDULO\ VLPSOLILHG E\ RPLWWLQJ WKH VRXUFH PDS WR % 3UHYLRXVO\ L ( fÂ§! $ ZDV GHILQHG WR EH WKH HTXDOL]HU LI IL JL DQG L IDFWRUV WKURXJK DQ\ RWKHU PDS ZKLFK ULJKW LGHQWLILHV DQG J L LV WKH VRXUFH PDS WR $ 7KH VRXUFH PDS WR % LV UHGXQGDQW LW PXVW EH IL JL ( fÂ§r % 6LPLODU QRWDWLRQDO HFRQRP\ LV DSSOLHG WR FRHTXDOL]HUV
PAGE 28
([DPSOH /HW XV FRQVLGHU WKH GLDJUDP $ I % r& 7KH GLDJUDP DOVR FRQWDLQV LGHQWLWLHV IRU DOO REMHFWV EXW IRU EUHYLW\ WKHVH DUH RPLWWHGf 7KH OLPLW % ;F $ Q$ WWEf RI WKLV GLDJUDP LV FDOOHG WKH SXOOEDFN RI DORQJ J RU WKH SXOOEDFN RI J DORQJ 7R EH H[SOLFLW IQ$ Q%J DQG LI 4T$ 4 fÂ§r $T% 4 fÂ§! %f VDWLVILHV IT$ T%J WKHQ WKHUH LV D XQLTXH PDS L 4 fÂ§! % ;S $ ,Q 6HW % [F $ ^E Ff H % [ $ JEf Ff` DQG WKH SURMHFWLRQ PDSV Q DUH WKH UHVWULFWLRQV RI WKH SURMHFWLRQV IURP WKH FDUWHVLDQ SURGXFW 3XOOEDFNV LQ DQ\ FRQFUHWH FDWHJRU\ $ HTXLSSHG ZLWK D OLPLW SUHVHUYLQJ IDLWKIXO IXQFWRU 8 $ fÂ§r 6HW DUH FRPSXWHG LGHQWLFDOO\ 7ZR SDUWLFXODU LQVWDQFHV RI SXOOEDFNV GHVHUYH VSHFLDO DWWHQWLRQ )LUVW OHW $ fÂ§ & EH DQ\ PDS DQG J 6 fÂ§! & EH D VXEVHW LQFOXVLRQ RI 6 & & 1RWH WKDW V Df f 6 [F $ LI DQG RQO\ LI JVf Df VXSSUHVVLQJ PHQWLRQ RI J WKLV UHDGV 6 [F $ fÂ§ ^Vf IEf V` 7KXV WKLV SXOOEDFN LV FDQRQLFDOO\ LVRPRUSKLF WR WKH SUHLPDJH RI 6 XQGHU 7KLV H[DPSOH SDUWLDOO\ PRWLYDWHV WKH QDPH fSXOOEDFNf 6HFRQG OHW $ fÂ§! % EH DQ\ PDS DQG FRQVLGHU WKH SXOOEDFN RI WKH GLDJUDP $ I $ r% LH WKH SXOOEDFN RI DORQJ LWVHOI 8VLQJ WKH FRPSXWDWLRQ IRU SXOOEDFNV LQ 6HW JLYHQ DERYH $ [% $ fÂ§ ^D Dnf f $ [ $ IDf Dnf`
PAGE 29
7KLV UHODWLRQ RQ $ LV RIWHQ FDOOHG WKH NHUQHO RI 7KXV RQH FDOOV $ [ $QOHIWQULJKWf WKH NHUQHO SDLU RI 0RUH LQGHSWK GLVFXVVLRQ RI SXOOEDFNV LV JLYHQ LQ %RUFHX[ > 9ROXPH 6HFWLRQ @ DQG +HUUOLFK DQG 6WUHFNHU > 6HFWLRQ @ 'HILQLWLRQ DQG 5HPDUNV $ FDWHJRU\ LV VDLG WR EH FRPSOHWH LI HDFK GLDJUDP KDV D OLPLW LW LV VDLG WR EH FRFRPSOHWH LI HDFK GLDJUDP KDV D FROLPLW 8QOLNH WKH VLWXDWLRQ IRU SRVHWV D FDWHJRU\ PD\ EH FRPSOHWH ZLWKRXW EHLQJ FRFRPSOHWH DQG YLFH YHUVD 6HH +HUUOLFK DQG 6WUHFKHU > 6HFWLRQ SOII@ IRU D GHWDLOHG GLVFXVVLRQ RI WKLV DQG UHODWHG LVVXHVf &RQVLGHU D IXQFWRU ) $ fÂ§! % ,I LV D GLDJUDP LQ $ WKHUH LV D GLDJUDP )' ZLWK 2EM )'f ^)$ $ H 2EM'f` DQG 0DS )'f ^)I I 0DS'f` /HW EH D GLDJUDP LQ $ $ IXQFWRU ) $ fÂ§! f% SUHVHUYHV WKH OLPLW RI LI ZKHQHYHU WKH OLPLW /eDf H[LVWV LQ $ '/ 'L$ff LV WKH OLPLW RI )' ) SUHVHUYHV OLPLWV LI IRU DQ\ GLDJUDP ) SUHVHUYHV WKH OLPLW RI ,I $ fÂ§} % LV FRQWUDYDULDQW LV D GLDJUDP LQ $ DQG / ^e$ $ 2EM'f`f LV WKH OLPLW RI WKHQ */ ^*L$ *$ fÂ§! */`f LV D VLQN IRU *' ,I */ ^*e$ *$ fÂ§ */`f LV WKH FROLPLW IRU *' WKHQ WDNHV OLPLWV WR FROLPLWV $QDORJRXVO\ PD\ WDNH FROLPLWV WR OLPLWV $GMRLQW )XQFWRUV 7KHUH DUH VHYHUDO XVHIXO FRQFHSWV RI DGMRLQW FRQQHFWLRQV EHWZHHQ FDWHJRULHV 7KHUH LV D VWURQJ DQDORJ\ EHWZHHQ SUHRUGHUHG VHWV DQG FDWHJRULHV DQ\ FDWHJRU\ PD\ EH SUHRUGHUHG E\ $A% $ % &DWHJRULHV DUH PRUH FRPSOH[ EHFDXVH PDQ\ PDSV FRXOG PDQLIHVW $ $ % :H EHJLQ ZLWK WKH FRQFHSW DQDORJRXV WR 'HILQLWLRQ DQG 5HPDUNV $ IXQFWRULDOf *DORLV FRQQHFWLRQ EHWZHHQ FDWHJRULHV $ DQG n% FRQVLVWV RI FRQWUDYDULDQW IXQFWRUV 7 $ fÂ§! p % fÂ§! $ WRJHWKHU ZLWK QDWXUDO
PAGE 30
WUDQVIRUPDWLRQV LGr fÂ§} *) DQG H: fÂ§! )* VXFK WKDW )U@$fH)$f LGA DQG *H%fU@*%f LGHV IRU HDFK $ 2EM$f DQG % f 2EMf WKHVH HTXDWLRQV DUH WKH VRFDOOHG WULDQJOH LGHQWLWLHV $OWHUQDWH WHUPLQRORJ\ LI )*U@Hf LV D IXQFWRULDO *DORLV FRQQHFWLRQ WKHQ ) DQG DUH DGMRLQW RQ WKH ULJKWf 7KH FDWHJRULHV IL[Ucf DQG IL[Hf FRQWDLQLQJ DOO REMHFWV VXFK WKDW U@$ UHVS H%f LV DQ LVRPRUSKLVP DUH GXDO > 6HFWLRQ /HPPD @ 7KHUH LV D QDWXUDO ELMHFWLRQ $$ *%f fÂ§r p )$f JLYHQ E\ I $ *%f +! )IfH%f % fÂ§! )$ 7KH LQYHUVH PDS LV J % fÂ§! )$f !! *IfY$f $ *% 7KHUH DUH WZR LGHQWLFDOf FDOFXODWLRQV UHTXLUHG WR FKHFN WKDW WKH IXQFWLRQV DUH PXWXDOO\ LQYHUVH 2QH LV VXPPDUL]HG E\ WKH GLDJUDP EHORZ *)$ fÂ§*)If *)*% r *% 7KH VTXDUH FRPPXWHV EHFDXVH LV QDWXUDO 7KH WULDQJOH FRPPXWHV EHFDXVH RI WKH LGHQWLW\ LGHV *H%fUc*%f 7KH UHDGHU PD\ IRUPXODWH DQG FKHFN ZKDW LV PHDQW E\ fQDWXUDOLW\f RI WKH ELMHFWLRQ ) DQG ERWK WDNH FROLPLWV WR OLPLWV
PAGE 31
(DFK UM$ KDV WKH IROORZLQJ XQLYHUVDO SURSHUW\ LI $ fÂ§ *% WKHUH LV D XQLTXH PDS )IfH%f % fÂ§ (0 ZKLFK PDNHV WKH IROORZLQJ GLDJUDP FRPPXWH $ fÂ§fÂ§rÂ‘ *)$ (DFK H% KDV WKH DQDORJRXV XQLYHUVDO SURSHUW\ 0XFK RI WKH OLWHUDWXUH MXVW GHDOV ZLWK DGMRLQW FRQQHFWLRQV ZKHUH ERWK IXQFWRUV DUH FRYDULDQW %\ GXDOLW\ DQ\ VXFK UHVXOW FDQ EH WUDQVODWHG LQ WHUPV RI *DORLV FRQQHFWLRQV %DQDVFKHZVNL DQG %UXQV >@ LQFOXGHV D UHDVRQDEO\ WKRURXJK H[SRVLWRU\ VHFWLRQ RQ IXQFWRUV ZKLFK DUH DGMRLQW RQ WKH ULJKW 7KH fIXQFWRULDOf *DORLV FRQQHFWLRQf FRQFHSW LV V\PPHWULF EXW WKH FRQWUDYDULDQFH RI WKH IXQFWRUV LQYROYHG LV VRPHWLPHV DZNZDUG $Q DQDORJRXV DV\PPHWULF FRQFHSW ZLWK WKH IXQFWRUV ERWK FRYDULDQW LV GHVFULEHG EHORZ 'HILQLWLRQ DQG 5HPDUNV $ IXQFWRULDOf DGMRLQW FRQQHFWLRQ EHWZHHQ FDWHJRULHV $ DQG FRQVLVWV RI IXQFWRUV ) $ fÂ§! DQG fÂ§ $ DQG QDWXUDO WUDQVIRUPDWLRQV U@ LG\L fÂ§ *) DQG H )* fÂ§! LGp VXFK WKDW *H%fUc*%f LGHV DQG H)$f)Uc$f LGA WKHVH HTXDWLRQV DUH WKH VRFDOOHG WULDQJOH LGHQWLWLHV ,Q WKLV VLWXDWLRQ RQH DOVR VD\V f) DQG DUH DGMRLQW IXQFWRUVf f) LV WKH OHIW DGMRLQWf DQG f* LV WKH ULJKW DGMRLQWf 7KH EDVLF SURSHUWLHV RI IXQFWRULDO DGMRLQW FRQQHFWLRQV FORVHO\ FRUUHVSRQG WR WKH EDVLF SURSHUWLHV RI IXQFWRULDO *DORLV FRQQHFWLRQV 7KH FDWHJRULHV IL[Uf DQG IL[Hf FRQWDLQLQJ DOO REMHFWV VXFK WKDW UHVS H%f LV DQ LVRPRUSKLVP DUH HTXLYDOHQW 7KHUH LV D QDWXUDO ELMHFWLRQ $$ *%f fÂ§ )$ %f JLYHQ E\ I $fÂ§r *%f W! H%f)If )$ fÂ§! %
PAGE 32
7KH LQYHUVH PDS LV J % )$f } *fIRLf $A*% ) SUHVHUYHV FROLPLWV SUHVHUYHV OLPLWV %RUFHX[ > 9ROXPH @ (DFK UM$ KDV WKH IROORZLQJ XQLYHUVDO SURSHUW\ LI $ fÂ§! *% WKHUH LV D XQLTXH PDS H%f)If )$ fÂ§A % ZKLFK PDNHV WKH IROORZLQJ GLDJUDP FRPPXWH $ fÂ§fÂ§!*)$ (DFK H% KDV WKH DQDORJRXV XQLYHUVDO SURSHUW\ LI )$ fÂ§! % WKHUH LV D XQLTXH *IfUM$f $ fÂ§! *% VXFK WKDW FRPPXWHV $Q\ IXQFWRU IRU ZKLFK WKHUH LV D QDWXUDO WUDQVIRUPDWLRQ UM ZLWK WKH DERYH XQLYHUVDO SURSHUW\ LV SDUW RI DQ DGMRLQW FRQQHFWLRQ )RU GLVFXVVLRQV RI DGMRLQW IXQFWRUV VHH 0DF/DQH > &KDSWHU ,9@ +HUUOLFK DQG 6WUHFNHU > 6HFWLRQV @ RU %RUFHX[ > 9ROXPH &KDSWHU @ 7KHUH LV D FULWHULRQ IRU GHWHUPLQLQJ ZKHQ IXQFWRUV DUH DGMRLQWV ZKLFK FRUUHVSRQGV WR 7KHRUHP 7KHRUHP $GMRLQW )XQFWRU 7KHRUHP > 9@f /HW $ EH D FRPSOHWH FDWHJRU\ $ IXQFWRU $ fÂ§! % KDV D OHIW DGMRLQW LI DQG RQO\ LI SUHVHUYHV OLPLWV DQG
PAGE 33
VROXWLRQ VHW FRQGLWLRQf IRU HDFK % f WKHUH LV D VHW DQG DQ ,LQGH[HG IDPLO\ RI PDSV ILn%fÂ§! *$L VXFK WKDW DQ\ PDS I % fÂ§! *$ FDQ EH ZULWWHQ DV K *WfIL IRU VRPH L DQG W $c fÂ§! $
PAGE 34
&+$37(5 $/*(%5$6 2) $ 021$' 0RQDGV DND WULSOHV DND VWDQGDUG FRQVWUXFWLRQVf DQG WKHLU (LOHQEHUJ0RRUHf DOJHEUDV SURYLGH D FRQFLVH IRUPXODWLRQ RI PDQ\ LPSRUWDQW FDWHJRULFDO DVSHFWV RI XQLYHUVDO DOJHEUD 7KH FDWHJRU\ RI DOJHEUDV IRU D PRQDG KDV VSHFLDO SURSHUWLHV ZKLFK DUH VXPPDUL]HG LQ 6HFWLRQ D FDWHJRU\ RI DOJHEUDV LV RIWHQ FRPSOHWH DQG FRFRPSOHWH DQG DOZD\V KDV fIUHH REMHFWVf (DFK IXQFWRULDOf DGMRLQW FRQQHFWLRQ LQGXFHV D PRQDG WKH FRUUHVSRQGHQFH EHWZHHQ DGMXQFWLRQV DQG PRQDGV LV GLVFXVVHG LQ 6HFWLRQ ,Q 6HFWLRQ WKH TXHVWLRQ RI ZKHQ DQ DGMRLQW FRQQHFWLRQ FRQQHFWV $ WR D FDWHJRU\ RI DOJHEUDV LV DGGUHVVHG ,Q 6HFWLRQ WKH TXHVWLRQ fZKHQ GRHV WKH FRPSRVLWH RI WZR PRQDGLF DGMXQFWLRQV \LHOG D PRQDGLF DGMXQFWLRQ"f LV GLVFXVVHG 7KH UHVXOWV LQ WKLV FKDSWHU DUH UHDVRQDEO\ ZHOO NQRZQ 0DF/DQH > &KDSWHU 9,@ %RUFHX[ > 9ROXPH &KDSWHU @ %DUU DQG :HOOV > &KDSWHUV DQG @ 0DQHV >@ DQG WKH LQWURGXFWLRQ WR WKH VHPLQDU QRWHV >@ FRQWDLQ JRRG H[SRVLWLRQV RI PRQDGV IURP YDULRXV SHUVSHFWLYHV &DWHJRULHV RI $OJHEUDV 'HILQLWLRQ $ PRQDG 7 7 UM Lf RQ $ FRQVLVWV RI D IXQFWRU 7$A$ D QDWXUDO WUDQVIRUPDWLRQ 9 Â‘ L8 7 DQG D QDWXUDO WUDQVIRUPDWLRQ L 7 fÂ§! 7
PAGE 35
VXFK WKDW WKH IROORZLQJ LGHQWLWLHV H[SUHVVHG E\ FRPPXWDWLYH GLDJUDPVf KROG WKH XQLW ODZV 7 f ,Q WKHVH GLDJUDPV 7Q GHQRWHV WKH QIROG FRPSRVLWH RI 7 ZLWK LWVHOIf ,QWXLWLYHO\ 7$ LV WKH IUHH REMHFW RQ $ U@$ LV WKH fLQVHUWLRQ RI YDULDEOHVf PDS IM$ LV WKH fVHPDQWLF FRPSRVLWLRQf LH D PDS ZKLFK DOORZV RQH WR YLHZ D SRO\QRPLDO ZLWK SRO\QRPLDO YDULDEOHV DV D SRO\QRPLDO 6HH ([DPSOH IRU D FRQFUHWH LOOXVWUDWLRQ RI WKH UROHV RI 7 DQG c 'HILQLWLRQ DQG 5HPDUNV /HW 7 EH D PRQDG RQ $ $ 7DOJHEUD $ Df FRQVLVWV RI $ f 2EM$f DQG D 7$ fÂ§! $ WKH VRFDOOHG VWUXFWXUH PDSf VXFK WKDW DUf$f fÂ§ N8 XQLW ODZf DQG 7$ fÂ§ r7$ 0 D 7$
PAGE 36
DVVRFLDWLYH ODZf FRPPXWHV $ KRPRPRUSKLVP $ Df fÂ§! % Ef RI 7DOJHEUDV LV DQ $PDS VXFK WKDW 7I 7$ fÂ§r7% D E FRPPXWHV 7KH FDWHJRU\ RI DOO 7DOJHEUDV DQG 7DOJHEUD KRPRPRUSKLVPV LV GHQRWHG $7 7KHUH LV D IRUJHWIXO IXQFWRU 87 $7 fÂ§ $ LW LV GHILQHG E\ 87$Df $ DQG 87If I 87 KDV D OHIW DGMRLQW )7 $ fÂ§ $7 GHILQHG E\ )W$f 7$S$f DQG )7If 7I 7KH DVVRFLDWHG QDWXUDO WUDQVIRUPDWLRQV DUH U@7 U LGA 8W)W 7 DQG H7 )W8W fÂ§r LGAU H7$ Df D > 9,7KHRUHP @ /LPLWV LQ $U DUH fFRPSXWHG LQ $f LQ WKH IROORZLQJ VHQVH 3URSRVLWLRQ > @ > 9ROXPH @ 6XSSRVH 7U[f LV D PRQDG RQ $ ,I LV D GLDJUDP LQ $U VXFK WKDW 87'f KDV D OLPLW /SX$ff WKHQ WKHUH LV D XQLTXH VWUXFWXUH PDS 7/ fÂ§r / VXFK WKDW HDFK SX$f D 7DOJHEUD KRPRPRUSKLVP 3URRI /HW EH D GLDJUDP LQ $U VXFK WKDW 87' KDV D OLPLW / ,$f )RU HDFK $ 2EM7f QDPH WKH VWUXFWXUH PDS V$ 7$ $ 7KH JRDO LV WR SURGXFH D VWUXFWXUH PDS V 7/ fÂ§! / VXFK WKDW HDFK W$ / fÂ§ $ LV D 7DOJHEUD PDS
PAGE 37
7KH UHTXLUHPHQW WKDW HDFK LV 7DOJHEUD PDS DPRXQWV WR IRU HDFK $ WKH GLDJUDP EHORZ FRPPXWHV )/ 6$ 6LQFH /eDf LV D VRXUFH IRU 8' )/ VD),Dff LV D VRXUFH IRU 8' 7KXV WKHUH LV D XQLTXH $ PDS V )/ fÂ§! / PDNLQJ HDFK GLDJUDP DERYH FRPPXWH &KHFNLQJ WKH XQLW ODZ VL/f LGÂ WKH GLDJUDP FRPPXWHV XVLQJ WKH GHILQLWLRQ RI V DQG WKDW L LV D QDWXUDO WUDQVIRUPDWLRQ 7KXV IRU HDFK $ ÂD fÂ§ ÂDV/f 3URP WKH XQLTXHQHVV RI WKH PDS IURP WKH OLPLW RI WR DQ\ RWKHU VRXUFH IRU LW IROORZV WKDW LGÂ VL/f $ VLPLODU fFRPSDULVLRQ RI VTXDUHV GLDJUDPf FDQ EH XVHG WR YHULI\ WKDW WKH DOJHEUD DVVRFLDWLYH ODZ KROGV Â’ &RUROODU\ ,I $ LV FRPSOHWH DQG 7 LV D PRQDG RQ $ WKHQ $7 LV FRPSOHWH ,I $ LV FRFRPSOHWH $7 LV RIWHQ DOVR FRFRPSOHWH 7KH IROORZLQJ WKHRUHP ZDV RULJLQDOO\ SURYHG ZLWK IHZHU K\SRWKHVHV LQ /LQWRQ >@ 2WKHU H[SRVLWLRQV DUH JLYHQ LQ %RUFHX[ > 9ROXPH @ DQG %DUU DQG :HOOV > 6HFWLRQ @ 7KHRUHP /HW $ EH FRFRPSOHWH DQG 7 EH D PRQDG RQ 7 $7 LV FRFRPSOHWH LI DQG RQO\ LI $KDV FRHTXDOL]HUV
PAGE 38
$GMRLQW &RQQHFWLRQV LQGXFH 0RQDGV 7KH IROORZLQJ SURSRVLWLRQ JLYHV D FRUUHVSRQGHQFH EHWZHHQ PRQDGV DQG DGMRLQW FRQn QHFWLRQV ,W VKRXOG EH HPSKDVL]HG WKDW WKH FRUUHVSRQGHQFH LV QRW ELMHFWLYH (DFK PRQDG JLYHV ULVH WR D XQLTXH DGMRLQW FRQQHFWLRQ EXW LQ JHQHUDO PDQ\ DGMXQFWLRQV LQGXFH WKH VDPH PRQDG 3URSRVLWLRQ &RUUHVSRQGHQFH EHWZHHQ PRQDGV DQG DGMRLQW FRQQHFWLRQV /HW ) $ fÂ§! &% fÂ§! $ Uc N8 fÂ§! *) H )* fÂ§! LGp EH DQ DGMRLQW FRQQHFWLRQ 7KHQ *)Uc*H)f LV D PRQDG RQ $ ,I 7 7 7@ Lf LV D PRQDG WKHQ *W)W 7 7 7 K DQG *7H7)7 >L 7KH SURRI RI WKH SUHFHGLQJ 3URSRVLWLRQ FRQVLVWV RI YHULI\LQJ LGHQWLWLHV WKH WULDQJOH LGHQWLWLHV IRU DGMRLQW FRQQHFWLRQV LPSO\ WKH XQLW ODZV WKH DVVRFLDWLYH ODZ KROGV EHFDXVH WKH VTXDUH GHILQLQJ KRUL]RQWDO FRPSRVLWLRQV FRPPXWHV 'HWDLOV DUH JLYHQ LQ 0DF/DQH > 9, 7KHRUHP @ 7KLV FRUUHVSRQGHQFH DOORZV FRQVWUXFWLRQ RI PDQ\ H[DPSOHV RI PRQDGV 2IWHQ EXW QRW DOZD\V fD QDWXUDOO\ RFFXUULQJf DGMRLQW FRQQHFWLRQ FRUUHVSRQGV WR WKH FDWHJRU\ RI DOJHEUDV RYHU WKH LQGXFHG PRQDGV ([DPSOH &RQVLGHU WKH FDWHJRU\ *US RI JURXSV 7KH IRUJHWIXO IXQFWRU 8FUS Â‘ *US fÂ§} 6HW DQG WKH IUHH JURXS IXQFWRU )JW3 *US fÂ§} 6HW IRUP DQ DGMRLQW FRQQHFWLRQ EHWZHHQ *US DQG 6HW 5HFDOO WKDW )RUSL;f LV WKH VHW RI DOO UHGXFHG ZRUGV
PAGE 39
ZKHUH [Â ; VÂ = D ZRUG LV UHGXFHG LI IRU DOO L ZLWK L Q [BL [Â 7KH RSHUDWLRQ RQ 7f*US$nf LV FRQFDWHQDWLRQ RI ZRUGV IROORZHG E\ UHGXFWLRQ 7 A*USA*US LV WKH IXQFWRU SDUW RI WKH LQGXFHG PRQDG LGVHW fÂ§! 7 LV WKH QDWXUDO WUDQVIRUPDWLRQ ZKRVH FRPSRQHQW DW ; VHQGV [ f ; WR ZRUG [ LI Zr f f Â‘ X!fQ 7;f DQG IRU HDFK ] [A f f f WKHQ PDf.n Â‘f r\ar.VUmL BLPP f [ QPQf?D QPQf (DFK JURXS LV DQ 7DOJHEUD WKH JURXS PXOWLSOLFDWLRQ DQG LQYHUVLRQ JLYH D PDS IURP WKH IUHH JURXS RQ WKH XQGHUO\LQJ VHW RI WR &RQYHUVHO\ HDFK 7DOJHEUD VWUXFWXUH JLYHV D JURXS PXOWLSOLFDWLRQ DQG LQYHUVLRQ *US LV HTXLYDOHQW WR WKH FDWHJRU\ RI 6HWU 7KLV H[DPSOH LV W\SLFDO LQ WKH VHQVH WKDW DQ\ FDWHJRU\ RI ILQLWDU\ DOJHEUDV LQ WKH VHQVH RI XQLYHUVDO DOJHEUD LV DOVR D FDWHJRU\ RI PRQDG DOJHEUDV 7KH SDUWV 7 UM DQG c RI D PRQDG JHQHUDOO\ KDYH WKH VDPH UROHV DV LQ WKLV H[DPSOHf ([DPSOH 6RPH QRQWULYLDOf DGMXQFWLRQV LQYROYLQJ 6HW LQGXFH WKH WULYLDO PRQDG )RU H[DPSOH FRQVLGHU WKH FDWHJRU\ 7RS RI WRSRORJLFDO VSDFHV DQG FRQWLQXRXV PDSV 7KH IRUJHWIXO IXQFWRU LUES 7RS fÂ§! 6HW KDV D OHIW DGMRLQW )WRS 6HW 7RS ZKLFK VHQGV D VHW ; WR WKH GLVFUHWH WRSRORJLFDO VSDFH ZLWK XQGHUO\LQJ VHW ; 2QH HDVLO\ FKHFNV WKDW 7 *7RS)7RS LV WKH LGHQWLW\ IXQFWRU RQ 6HW DQG WKDW DOO DVVRFLDWHG QDWXUDO WUDQVIRUPDWLRQV DUH LGHQWLWLHV 'HILQLWLRQ 6XSSRVH ) $ fÂ§! p *%fÂ§r $ UM LGA *) DQG H )* fÂ§! LGp IRUPV DQ DGMRLQW FRQQHFWLRQ %\ 3URSRVLWLRQ WKH DGMRLQW FRQQHFWLRQ LQGXFHV D PRQDG 7 *) *IL)f ,I % LV HTXLYDOHQW WR $7 WKHQ ZH VD\ WKH DGMRLQW FRQQHFWLRQ VRPHWLPHV MXVW WKH ULJKW DGMRLQW *f LV PRQDGLF 2IWHQ D FRQFUHWH FDWHJRU\ % KDV D fFDQRQLFDOf IRUJHWIXO IXQFWRU % fÂ§! $? LQ WKLV FDVH RQH PD\ HYHQ VD\ % LV PRQDGLF RYHU $ RPLWWLQJ PHQWLRQ RI *
PAGE 40
7KHVH H[DPSOHV LOOXVWUDWH WKH TXDOLWDWLYHO\ GLIIHUHQW EHKDYLRUV RI PRQDGLF DQG QRQ PRQDGLF DGMRLQW FRQQHFWLRQV 0RQDGLF FDWHJRULHV DUH GHWHUPLQHG E\ WKH FRPELQDWRULDO VWUXFn WXUH RI WKH IUHH DOJHEUD IXQFWRU )7 0RVW fUHODWLRQDOf FDWHJRULHV OLNH 7RS 3 WKH FDWHJRU\ RI JUDSKV HWF KDYH IUHH IXQFWRUV EXW WKHVH IUHH IXQFWRUV GR QRW DGG DQ\ VWUXFWXUH WR WKH XQGHUO\LQJ VHW WKH\ PHUHO\ DWWDFK WKH fPRVW GLVFUHWHf SRVVLEOH UHODWLRQ WR WKH JLYHQ VHW 'HWHFWLQJ &DWHJRULHV RI $OJHEUDV %HFDXVH RI WKH VSHFLDO SURSHUWLHV RI $7 WKH TXHVWLRQ RI ZKHQ DQ DGMRLQW FRQQHFWLRQ LV PRQDGLF KDV JUHDW SUDFWLFDO LPSRUWDQFH 5HPDUN /HW $ Df EH D 7DOJHEUD %HFDXVH RI WKH DVVRFLDWLYH DQG XQLW ODZV DQG EHFDXVH LV QDWXUDO WKH IROORZLQJ HTXDWLRQV KROG QDWf 7DfU>7$f fÂ§ USfD XQLWOf M$fWc7$f LG7$ XQLWf DUf$f LGÂ DQG DVVRFf D7Df DIL$f 7KHVH HTXDWLRQV LPSO\ WKDW D fÂ§ FRHTcL$7Df? IRU LI E 7$ fÂ§ % ULJKWLGHQWLILHV c$ DQG 7D WKHQ EU@$f $ fÂ§ % DQG EUc$fD E7Df^7@7$f E\ QDWf EIL$fU@7$f VLQFH E ULJKW LGHQWLILHV fÂ§ E E\ XQLWOf 2QH YHULILHV WKDW EUc$f WKH XQLTXH PDS $ fÂ§! % ZLWK ID IRU LI ID E WKHQ IDUM$f E^Uc$f 7KHVH HTXDWLRQV JLYH D JUHDW GHDO RI LQIRUPDWLRQ DERXW D 7KXV WKH NH\ K\SRWKHVLV LQ %HFNfV WKHRUHP WKH FULWHULRQ IRU ZKHQ DQ DGMXQFWLRQ LV PRQDGLF LV WKH SUHVHUYDWLRQ DQG UHIOHFWLRQ FRHTXDOL]HUV REH\LQJ WKH HTXDWLRQV GHVFULEHG DERYH 8VLQJ WKH DERYH GHVFULEHG HTXDWLRQV IRU PRWLYDWLRQ ZH RIIHU WKH IROORZLQJ GHILQLWLRQV
PAGE 41
'HILQLWLRQ DQG 5HPDUNV &RQVLGHU PDSV IJ$fÂ§!% ,I H % fÂ§ & KDV WKH IHDWXUH WKDW IRU DQ\ IXQFWRU ) )H FRHT)I )Jf WKHQ ZH VD\ H LV DQ DEVROXWH FRHTXDOL]HU ,I WKHUH DUH PDSV H % fÂ§! & VF nÂ‘ & fÂ§! % DQG VE % fÂ§! $ VXFK WKDW IV% N8 JV% LGF HVF HI HJ WKHQ ZH VD\ H LV D VSOLW FRHTXDOL]HU 1RWH WKDW IRU HYHU\ 7DOJHEUD ^$ Df D LV D VSOLW FRHTXDOL]HU RI 7D DQG S$ ,W LV DOVR HDV\ WR VHH WKDW HYHU\ VSOLW FRHTXDOL]HU LV DEVROXWH 7KH IROORZLQJ WKHRUHP LV GXH WR %HFN XQSXEOLVKHGf /LQWRQ >@ FRQWDLQV D GHWDLOHG GLVFXVVLRQ RI YDULDWLRQV RQ WKH WKHRUHP 7KH WKHRUHP ZDV RULJLQDOO\ SKUDVHG LQ WHUPV RI VSOLW FRHTXDOL]HUV RQO\ 3DUH >@ UHILQHG WKH WKHRUHP WR LQFOXGH WKH fDEVROXWH FRHTXDOL]HUf FRQGLWLRQ 0DQ\ YDULDWLRQV RQ WKH K\SRWKHVHV H[LVW WKH YHUVLRQ VWDWHG KHUH LV IRXQG LQ 0DF/DQH > 9,@ 2WKHU H[SRVLWLRQV RI WKH WKHRUHP PD\ EH IRXQG LQ %DUU DQG :HOOV > 6HFWLRQ @ DQG %RUFHX[ > 9ROXPH 6HFWLRQ @ 7KHRUHP /HW ) $ fÂ§! fÂ§! $ U" Hf EH DQ DGMRLQW FRQQHFWLRQ DQG 7 *) UM J DQG J *H) EH WKH DVVRFLDWHG PRQDG 7KH IROORZLQJ FRQGLWLRQV DUH HTXLYDOHQW 7KH DGMXQFWLRQ )*U@Hf LV PRQDGLF ,I I J $ fÂ§r % f 0DS%f DQG WKH SDLU *I*Jf KDV DQ DEVROXWH FRHTXDOL]HU H WKHQ H FRHTIJf H[LVWV DQG *H fÂ§ Hn
PAGE 42
,I IJ $ fÂ§! % H 0DS%f DQG WKH SDLU *I*Jf KDV D VSOLW FRHTXDOL]HU Hn WKHQ H FRHT Jf H[LVWV DQG *H Hn 5HFRJQL]LQJ PRQDGLF DGMXQFWLRQV LV FRPSOH[ EHFDXVH FRPSRVLWLRQV RI PRQDGLF IXQFn WRUV DUH QRW JHQHUDOO\ PRQDGLF ([DPSOH 7KH FDWHJRU\ $E RI DEHOLDQ JURXSV LV PRQDGLF RYHU 6HW DQG WI$E VHH ([DPSOH f LV PRQDGLF RYHU $E EHFDXVH DQ\ UHIOHFWLRQ LV PRQDGLF (DFK IUHH DEHOLDQ JURXS LV WRUVLRQ IUHH VR WKH PRQDG RQ 6HW LQGXFHG E\ WKH IUHH DEHOLDQ JURXS IXQFWRU LV WKH VDPH DV WKH PRQDG LQGXFHG E\ WKH fIUHH WRUVLRQ IUHH DEHOLDQ JURXS IXQFWRUf 7KXV LI WI$E ZHUH PRQDGLF RYHU VHW WKHQ WI$E DQG $E ZRXOG EH HTXLYDOHQW FDWHJRULHV ERWK HTXLYDOHQW WR D VXLWDEOH FDWHJRU\ RI PRQDG DOJHEUDV 7KH FDWHJRULHV WI$E DQG $E DUH REYLRXVO\ QRW HTXLYDOHQW 7KH SUHFHGLQJ H[DPSOH LV VXPPDUL]HG IURP %RUFHX[ > 9ROXPH ([DPSOH @f ,I $ fÂ§r 6HW WKH K\SRWKHVHV RI %HFNfV WKHRUHP FDQ EH UHIRUPXODWHG WR PDNH LW HDVLHU WR FKHFN ZKHWKHU LV PRQDGLF 7KHRUHP > 7KHRUHP @ /HW $ fÂ§! 6HW EH D IXQFWRU ZLWK D OHIW DGMRLQW 6XSSRVH WKDW $ LV FRPSOHWH DQG KDV FRHTXDOL]HUV 7KH IROORZLQJ DUH HTXLYDOHQW LV PRQDGLF VDWLVILHV WKH IROORZLQJ FRQGLWLRQV f SUHVHUYHV DQG UHIOHFWV UHJXODU HSLPRUSKLVPV f LI I *$ fÂ§! ; LV DQ LVRPRUSKLVP WKHQ WKHUH LV D XQLTXH PDS J $ fÂ§! % VXFK WKDW *J fÂ§ I f UHIOHFWV NHUQHO SDLUV 7KH IROORZLQJ OHPPD SURYLGHV XVHIXO LQIRUPDWLRQ FRQFHUQLQJ ZKHQ D IXQFWRU EHWZHHQ WZR FDWHJRULHV RI DOJHEUDV LV PRQDGLF
PAGE 43
/HPPD > 9ROXPH &RUROODU\ @ 6XSSRVH 8 fÂ§! $ 9 fÂ§! $ DQG 4 % fÂ§! & DUH IXQFWRUV ,I8 fÂ§ 94 8 DQG 9 DUH PRQDGLF DQG KDV FRHTXDOL]HUV WKHQ 4 LV PRQDGLF ,Q SDUWLFXODU 4 KDV D OHIW DGMRLQW 'LVWULEXWLYH /DZV &RPSRVLWLRQV RI PRQDGLF DGMXQFWLRQV DUH QRW JHQHUDOO\ PRQDGLF ZKHQ D FRPSRVLWLRQ RI PRQDGLF DGMXQFWLRQV LV PRQDGLF LW LQGLFDWHV D GLVWULEXWLYH ODZ EHWZHHQ WKH WZR VWUXFWXUHV 7KH IROORZLQJ VHFWLRQ VXPPDUL]HV WKH UHVXOWV ODWHU QHHGHG IURP %HFNfV >@ WKLV H[SRVLWLRQ IROORZV %HFNfV QRWDWLRQ H[FHSW WKDW FRPSRVLWLRQ RI IXQFWLRQV KHUH ZLOO UHDG ULJKWWROHIW )RU WKH GXUDWLRQ RI WKH VHFWLRQ DVVXPH WKHUH DUH WZR PRQDGV 7 7 W@W S7f DQG 6 6 U@V SVf RYHU VRPH EDVH FDWHJRU\ $ )RU DQ\ PRQDG ZH XVH ) DQG 8 ZLWK VXSHUVFULSWV IRU WKH QDPH RI WKH PRQDG WR GHQRWH WKH IUHH DOJHEUD IXQFWRU DQG IRUJHWIXO IXQFWRU DVVRFLDWHG ZLWK D PRQDG UHVSHFWLYHO\ 6HH IRU WKH GHILQLWLRQV RI ) DQG 8 7KLV PDWHULDO LV UDWKHU DEVWUDFW VR LW KHOSV WR KDYH DQ H[DPSOH LQ PLQG DIWHU HDFK GHILQLWLRQ DQG WKHRUHP ZH ZLOO LOOXVWUDWH ZKDW LW PHDQV XVLQJ 6 WKH IUHH PRQRLG PRQDG RYHU 6HW DQG 7 WKH IUHH DEHOLDQ JURXS PRQDG RYHU 6HW WKH FRPSRVLWH PRQRLG 76 JLYHV WKH IUHH ULQJ )RU D VHW ; 6; FRQVLVWV RI DOO fVWULQJVf IURP ; ZLWK FRQFDWHQDWLRQ DV WKH RSHUDWLRQ DFWV RQ IXQFWLRQV E\ 6I[L[ Â‘fÂ‘[Qf I[fI[f Â‘ f f I[Qf 7; FRQVLVWV RI DOO IRUPDO ILQLWHf OLQHDU FRPELQDWLRQV RI HOHPHQWV RI ; ZLWK LQWHJHU FRHIILn FLHQWV 7 DFWV RQ IXQFWLRQV LQ WKH H[SHFWHG ZD\ 'HVSLWH WKH FRQFUHWH LOOXVWUDWLRQV LQ WHUPV RI WKHVH PRQDGV DOO WKHRUHPV DQG GHILQWLRQV ZLOO DSSO\ WR DQ\ PRQDGV 6 DQG 7 'HILQLWLRQ DQG 5HPDUNV $ GLVWULEXWLYH ODZ RI 6 RYHU 7 LV D QDWXUDO WUDQVIRUPDWLRQ e 67!76
PAGE 44
VDWLVI\LQJ WKH FRPSDWLELOLW\ FRQGLWLRQV 7 6 ,I WKHUH LV D GLVWULEXWLYH ODZ e 67 fÂ§ 76 RI 6 RYHU 7 WKHQ WKH FRPSRVLWH PRQDG LV 76 76U@76 U@77@VS76 S7 SV7e6ff ,Q WKH GHILQLWLRQV RI Uf76 DQG S76 MX[WDSRVLWLRQ GHQRWHV KRUL]RQWDO FRPSRVLWLRQ RI QDWXUDO WUDQVIRUPDWLRQV 7KH YHULILFDWLRQ WKDW 76 DFWXDOO\ GHILQHV D PRQDG LV RPLWWHG IRU PRUH LQIRUPDWLRQ VHH %HFN >@ ,Q WKH FDVH ZKHQ f6 IUHH PRQRLGf DQG f7 IUHH DEHOLDQ JURXSf WKH QDWXUDO WUDQVn IRUPDWLRQ e 67 fÂ§! 76 H[SUHVVHV D SURGXFWRIVXPV DV D VXPRISURGXFWV LQ WKH XVXDO ZD\ Q A DLN Q DFNfN NH. LHL F NH. ZKHUH F UDQJHV RYHU DOO FKRLFH IXQFWLRQV F fÂ§! ,W LV D VRPHZKDW HQOLJKWHQLQJ H[HUFLVH WR FKHFN WKDW Â GHILQHG WKLV ZD\ LV D QDWXUDO WUDQVIRUPDWLRQ 67 fÂ§ 76 $ GLVWULEXWLYH ODZ PD\ RI 7 RYHU 6 DOVR EH YLHZHG DV D fZD\ RI OLIWLQJ 7 WR D PRQDG RYHU 6
PAGE 45
'HILQLWLRQ 7 KDV D OLIWLQJ LQWR ,V LI WKHUH LV D PRQDG 7 VXFK WKDW 786 867 8VUM7 fÂ§ U@78V DQG 8VSA S78V 7R VNHWFK WKH VLWXDWLRQ 7 LV D OLIWLQJ RI 7 RQWR LI FRPPXWHV 7KHRUHP 7KHUH LV D ELMHFWLYH FRUUHVSRQGHQFH EHWZHHQ OLIWLQJV RI 7 WR ,V DQG GLVn WULEXWLYH ODZV RI 6 RYHU 7 7KH SURRI LV RXWOLQHG LQ %HFN >@ 7KH FRUUHVSRQGHQFH LV GHILQHG DV IROORZV LI L LV D GLVWULEXWLYH ODZ DQG $ Df LV DQ DOJHEUD WKHQ 7 LV GHILQHG DV IROORZV 7$Df 7$ 7Df^L$ff Uf7^$ Df W@7$ $ Df 7$ Df DQG IL7$ Df S7$ 77$ Df } 7$ Df &RQVLGHU WKH IROORZLQJ GLDJUDP WKH VTXDUH FRPPXWHV EHFDXVH Uc7 LV D QDWXUDO WUDQVIRUPDWLRQ WKH WULDQJOH FRPPXWHV EHFDXVH O LV FRPSDWLEOH ZLWK UI 6LQFH WKH SHULPHWHU RI WKH GLDJUDP FRPPXWHV UI $ LV DQ DOJHEUD
PAGE 46
PDS $ VLPLODU GLDJUDP ZKLFK XVHV WKH FRPSDWLELOLW\ EHWZHHQ DQG S7 VKRZV S) LV DQ DOJHEUD PDS 6LQFH WKH XQGHUO\LQJ PDSV DUH GHILQHG LQ $ DQG REH\ WKH PRQDG ODZV WMW DQG S) DOVR REH\ WKH PRQDG ODZV ,I 7 LV D OLIWLQJ RI 7 RYHU $V LV GHILQHG WR EH WKH IROORZLQJ FRPSRVLWLRQ 67 676 678V)V 86)6867)6 X6I)IV 867)6 67 ,Q WKH DERYH HTXDWLRQ HV LV WKH FRXQLW RI WKH DGMXQFWLRQ )V 8V U HVf HV$ Df Df $IWHU VRPH GHWDLOHG FRPSXWDWLRQ RQH YHUILHV WKDW WKLV LV D GLVWULEXWLYH ODZ DQG WKDW WKH FRUUHVSRQGHQFHV GHVFULEHG DUH PXWXDOO\ LQYHUVH &RUROODU\ ,I 7 KDV D OLIWLQJ WR $V WKHQ WKHUH LV D FRPSRVLWH PRQDG 76 7KHRUHP 6XSSRVH 67 fÂ§! 76 LV D GLVWULEXWLYH ODZ 7KH FDWHJRULHV $76 DQG $6f7 DUH HTXLYDOHQW /HW XV FRQVLGHU WKLV LQ PRUH GHWDLO $Q REMHFW LQ $6f7 FRQVLVWV RI DQ $REMHFW $ DORQJ ZLWK DQ VWUXFWXUH DV 6$ fÂ§! $ 7VWUXFWXUH D[ 7$ fÂ§ $ DQG VWUXFWXUH W 67 $ fÂ§! $ IRU 7$ VXFK WKDW WKH IROORZLQJ GLDJUDP FRPPXWHV 6FL7 67 $ fÂ§ 6$ W D6 7$ fÂ§ !$ *LYHQ D 7DOJEUD ^$ Df $ KDV D 7VWUXFWXUH D7 PDS GHILQHG E\ D7 D7U@6$f 7$!$ DQG DQ VWUXFWXUH PDS DV GHILQHG E\ DV DU@76$f 6$ fÂ§! $
PAGE 47
7KH IROORZLQJ GLDJUDP ZKLFK H[SUHVVHV WKH GLVWULEXWLYLW\ RI WKH VWUXFWXUH PDSV FRPPXWHV 7KXV ZH KDYH D PDS $76 fÂ§! $6f7 JLYHQ E\ $ Df H $76 L r $ DU DV7DVe$ff ZKLFK LV IXQFWRULDO EHFDXVH ERWK UMn DQG UI DUH QDWXUDO 7KH LQYHUVH IXQFWRU ^$6f7 fÂ§! $76 PDSV $DUDVWf WR DV6DUf fÂ§ DUW 7KH YHULILn FDWLRQV IRU WKHVH DVVHUWLRQV LV JLYHQ LQ %HFN >@
PAGE 48
&+$37(5 *(1(5$7,1* 68%021$'6 ,Q WKLV FKDSWHU D WHFKQLTXH IRU FUHDWLQJ PRQDGV LV ODLG RXW EHIRUH SURFHHGLQJ IRUPDOO\ OHW XV FRQVLGHU D URXJK RXWOLQH RI WKH WHFKQLTXH 6XSSRVH 7 U@ Uf LV D PRQDG RQ D UHDVRQDEOH FDWHJRU\ DQG WKDW ) fÂ§r 7 LV D VXEIXQFWRU RI 7 7KH JRDO LV WR H[WHQG ) WR D VXEPRQDG RI 7 ,QWXLWLYHO\ VSHDNLQJ ) LV D QDWXUDO FROOHFWLRQ RI SRO\QRPLDOV 2QH VKRXOG WKHUHIRUH UHTXLUH WKDW IRU DOO $ U$f$f & )$f LH WKDW IRU HDFK $ DQG D $ )$ FRQWDLQV WKH fFRQVWDQW SRO\QRPLDO ZLWK YDOXH Df ) LV QRW JHQHUDOO\ fFORVHG XQGHU FRPSRVLWLRQf LH LW LV QRW JHQHUDOO\ WUXH WKDW A$f)$f & $ 7R FRUUHFW WKLV SUREOHP ZH VWDUW ZLWK ) DQG LWHUDWLYHO\ DGG SRO\QRPLDOV REWDLQHG E\ FRPSRVLQJ PHPEHUV RI )$ 7KH SXUSRVH RI WKH FKDSWHU LV WR IRUPDOL]H WKH SUHFHGLQJ YDJXH RXWOLQH 7KH DXWKRU DEVWUDFWHG DQG FODULILHG > 3URSRVLWLRQ @ ZKLFK PD\ EH YLHZHG DV D VSHFLDO FDVH RI WKLV UHVXOW 7KXV WKH PDLQ UHVXOWV RI WKLV FKDSWHU DUH FDOOHG 0HVHJXHUfV /HPPDV DQG WR DFNQRZOHGJH WKH DQDORJ\ ZKLFK SURPSWHG WKH WHFKQLTXH 7KH DXWKRU EHOLHYHV WKH JHQHUDO IRUPXODWLRQ RI WKH WHFKQLTXH LV QHZ 6HFWLRQ GHWDLOV UHTXLUHPHQWV RQ D fUHDVRQDEOH FDWHJRU\f H[SODLQV SUHFLVHO\ ZKDW LV PHDQW E\ fVXEIXQFWRUf DQG JLYHV PHWKRGV IRU FRQVWUXFWLQJ VXEIXQFWRUV 6HFWLRQ JLYHV D SURRI RI 0HVHJXHUfV /HPPDV 6HFWLRQ GHVFULEHV DQ H[DPSOH VKRZLQJ WKH QHFHVVLW\ RI D WHFKQLFDO K\SRWKHVLV RI WKH OHPPDV 6XEIXQFWRUV 7KLV VHFWLRQ GHVFULEHV D JHQHUDO WKHRU\ RI VXEIXQFWRUV $ VXEIXQFWRU RI ) ; fÂ§! 6HW LV D QDWXUDO WUDQVIRUPDWLRQ U@ ( fÂ§ ) VXFK WKDW HDFK FRPSRQHQW LV D VXEVHW LQFOXVLRQ 7KH
PAGE 49
FRQFHSW fVXEIXQFWRUf LV TXLWH XVHIXO EXW LQ VRPH FDWHJRULHV VXFK DV 3 7RS DQG /RFDOH H[SODLQHG LQ ([DPSOH f WKHUH LV HLWKHU f QR REYLRXV PHDQLQJ IRU VXEVHW LQFOXVLRQ LQ /RFDOHf RU f PRUH WKDQ RQH SRVVLEOH VWUXFWXUH RQ HDFK VXEVHW LQ 3 DQG 7RSf 7KH IROORZLQJ VHFWLRQ H[SRXQGV D WKHRU\ LQ ZKLFK fVXEIXQFWRUf PHDQV fQDWXUDO WUDQVIRUPDn WLRQ ZKRVH FRPSRQHQWV DUH DOO H[WUHPDO PRQRf $[LRPV 7KURXJKRXW WKH EDVH FDWHJRU\ $ LV DVVXPHG WR KDYH WKH IROORZLQJ IHDWXUHV $ LV FRPSOHWH $ LV DQ HSL H[WUHPDO PRQRf FDWHJRU\ 5HFDOO WKDW DQ HSL H[WUHPDO PRQRf FDWHJRU\ LV D FDWHJRU\ ZLWK WKH IHDWXUHV WKDW f )RU HDFK PDS LW LV SRVVLEOH WR IDFWRU I DV I fÂ§ PH ZKHUH H LV HSL DQG P LV H[WUHPDO PRQR 7KLV IDFWRUL]DWLRQ LV XQLTXH LQ WKH VHQVH WKDW LI PnHn LV DQRWKHU ZD\ RI ZULWLQJ DV DQ HSL IROORZHG E\ DQ H[WUHPDO PRQR WKHUH LV DQ LVRPRUSKLVP L VXFK WKDW H FRPPXWHV f LI D DQG E DUH H[WUHPDO PRQR DQG DE LV GHILQHG WKHQ DE LV H[WUHPDO PRQR $ LV H[WUHPDOO\ZHOOSRZHUHG 7KLV PHDQV IRU HDFK $ WKHUH LV D VHW ^D nÂ‘ 6?! $` RI H[WUHPDO PRQRPRUSKLVPV VXFK WKDW LI P 7 fÂ§ $ LV DQ\ H[WUHPDO PRQRPRUSKLVP WKHUH LV DQ LQGH[ $ DQG LVRPRUSKLVP L VXFK WKDW I? fÂ§ PL
PAGE 50
,I $ KDV WKHVH IHDWXUHV WKHQ LW LV DQ 6)FDWHJRU\ f6)f VWDQGV IRU fVXEIXQFWRUff ([DPSOH %\ +HUUOLFK DQG 6WUHFNHU > @ DQ\ ZHOOSRZHUHG FDWHJRU\ ZLWK LQn WHUVHFWLRQV DQG HTXDOL]HUV LV DQ HSL H[WUHPDO PRQRf FDWHJRU\ 6R LI $ LV FRPSOHWH DQG ZHOOSRZHUHG WKHQ $ LV DQ 6)FDWHJRU\ ,W IROORZV WKDW 3 6HW 7RS DQG SUDFWLFDOO\ DQ\ UHDVRQDEOH FDWHJRU\ RI WRSRORJLFDO VSDFHV LV DQ 6)FDWHJRU\ ([DPSOH 7KH FDWHJRU\ RI IUDPHV LH FRPSOHWH ODWWLFHV LQ ZKLFK D D?6 ?^D $ V V ( 6` KROGV IRU DOO HOHPHQWV D DQG VXEVHWV 6 LV FRPSOHWH DQG FRFRPSOHWH 7KLV FDWHJRU\ DOVR KDV UHJXODU HSL PRQRf IDFWRUL]DWLRQV 7KXV /RFDOH WKH GXDO FDWHJRU\ WR WKH FDWHJRU\ RI IUDPHV LV DQ 6)FDWHJRU\ /RFDOH LV H[WUHPDOO\ZHOOSRZHUHG EXW QRW ZHOOSRZHUHG 7KH DXWKRUfV LQWHUHVW LQ ORFDOHV PRWLYDWHG KLP WR XVH WKH JLYHQ GHILQLWLRQ ZKLFK RQO\ UHTXLUHV H[WUHPDOn ZHOOSRZHUHGQHVVf IRU 6)FDWHJRU\ UDWKHU WKDQ GHILQLQJ 6)FDWHJRU\ WR PHDQ fFRPSOHWH DQG ZHOOSRZHUHGf VR WKDW WKH WKHRU\ RI VXEIXQFWRUV ZRXOG DSSO\ WR ORFDOLF VXEIXQFWRUV LQ DGGLWLRQ WR WKH SUHYLRXV H[DPSOHV )RU EDFNJURXQG RQ IUDPHV DQG WKHLU UHODWLRQ WR SRLQWVHW WRSRORJ\ VHH ,VEHOO >@ RKQVWRQH >@ R\DO DQG 7LHUQH\>@ DQG 0DGGHQ >@f /HPPD $VVXPH WKDW $ LV DQ HSL H[WUHPDO PRQRf FDWHJRU\ 7KHQ GLDJRQDOL]DWLRQf > @ ,I JH PI ZKHUH H LV HSL DQG P LV H[WUHPDO PRQR WKHQ WKHUH H[LVWV N VXFK WKDW PN J DQG NH I > f@ $Q\ LQWHUVHFWLRQ RI H[WUHPDO VXEREMHFWV LV H[WUHPDO > f@ 3XOOEDFNV RI H[WUHPDO PRQRPRUSKLVPV DUH H[WUHPDO PRQR 'HILQLWLRQ DQG 5HPDUNV 'HILQLQJ 6XE$f WKH ODWWLFH RI H[WUHPDO VXEREn MHFWV ,Q DQ 6)FDWHJRU\ $ WKH ODWWLFH RI HTXLYDOHQFH FODVVHVf RI H[WUHPDO VXEREMHFWV KDV SDUWLFXODUO\ QLFH VWUXFWXUDO IHDWXUHV
PAGE 51
$Q H[WUHPDO VXEREMHFW RI $ LV DQ H[WUHPDO PRQRPRUSKLVP V 6 fÂ§! $ 'HILQH D SUHRUGHU RQ WKH FODVV RI H[WUHPDO VXEREMHFWV E\ V 6 fÂ§ $f & W 7 fÂ§ $f } F 6 fÂ§r 7 V fÂ§ WF )RU EUHYLW\ RQH RIWHQ RQO\ PHQWLRQV WKH PDS RU REMHFW SDUW RI DQ H[WUHPDO VXEREMHFW 7R DYRLG FRQIXVLRQ WKH VDPH OHWWHU ZLOO EH XVHG WR GHQRWH ERWK SDUWV ZLWK WKH ORZHU FDVH OHWWHU XVHG IRU WKH IXQFWLRQ 6LQFH W LV PRQR WKHUH LV DW PRVW RQH PDS F ZKLFK PDQLIHVWV V & W ,I V & W DQG W & V WKHQ WKHUH DUH F[ DQG F VXFK WKDW V fÂ§ WF? DQG W fÂ§ VF 6R W IFF DQG V VFFM 6LQFH FFL VKRZV V & V F&L LGJ 6LPLODUO\ FLF LG\ 2QH LGHQWLILHV H[WUHPDO VXEREMHFWV V DQG W LI V & W DQG I & V RU HTXLYDOHQWO\ ZKHQ WKHUH LV DQ LVRPRUSKLVP F VXFK WKDW V fÂ§ WF 7KH VHW RI HTXL YDODQFH FODVVHV XQGHU WKLV UHODWLRQ LV GHQRWHG 6XE $f $Q\ PDS F ZKLFK H[KLELWV V & W LV H[WUHPDO PRQR )RU LI FD fÂ§ FE WKHQ VD WHD WHE VE VR VLQFH V LV PRQRf D VR F LV PRQR ,I F PH ZKHUH H LV HSL WKHQ V WF WPH 6LQFH V LV H[WUHPDO PRQR H PXVW EH DQ LVRPRUSKLVP 7KXV F LV H[WUHPDO PRQR 5HFDOO WKDW WKH LQWHUVHFWLRQ RI D VHW 0 RI PRQRPRUSKLVPV ZLWK D FRPPRQ FRGRPDLQ LV WKH OLPLW RI WKH GLDJUDP JHQHUDWHG E\ 0 6LQFH $ LV FRPSOHWH DQG LQWHUVHFWLRQV RI H[WUHPDO PRQRPRUSKLVPV DUH H[WUHPDO 6XE$f LV D FRPSOHWH ODWWLFH ZLWK PHHW RSHUDWLRQ If DQG MRLQ RSHUDWLRQ _,Q JHQHUDO MRLQV LQ 6XE$f DUH QRW GLVMRLQW XQLRQV LQ JHQHUDO $Â LV FRPSXWHG XVLQJ M$L If^$9L $&$n` 6HH %RUFHX[ > 9ROXPH @f $ PDS I $ fÂ§r % LQGXFHV DQ DGMRLQW FRQQHFWLRQ EHWZHHQ 6XE$f DQG 6XE%f 6HH %RUFHX[ > 9ROXPH @ %RUFHX[fV UHVXOWV DUH SKUDVHG LQ WHUPV RI fVWURQJ PRQRPRUn SKLVPVf 8QGHU RXU DVVXPSWLRQV D PDS LV VWURQJ PRQR LI DQG RQO\ LI LW LV H[WUHPDO PRQRf ,I V 6 fÂ§A $ f 6XE$f XVH WKH HSL H[WUHPDO PRQRf IDFWRUL]DWLRQ WR REWDLQ D XQLTXH
PAGE 52
Vn 6XEe"f VXFK WKDW $ I % 6 6 rI6f FRPPXWHV 7KH QRWDWLRQ 6f LV XVHG IRU WKH LPDJH RI 6 WR UHPLQG WKH UHDGHU RI WKH DQDORJ\ WR RUGLQDU\ VHW WKHRUHWLF LPDJHV RI VXEVHWV XQGHU PDSV 7KH PDS V 6 fÂ§! $f LfÂ§ Vn 6n I6f fÂ§! %f LV WKH OHIW DGMRLQW 7KH ULJKW DGMRLQW IaO 6XE6f fÂ§! 6XE$f LV REWDLQHG E\ WDNLQJ WKH SXOOEDFN RI V 6 fÂ§! % DORQJ 7KXV I^6f &7 A 6&I?7f O8Vf OEZV!n DQG UI@VLf I@UVLf 'HILQLWLRQ DQG 5HPDUNV /HW ) ; fÂ§A $ EH D IXQFWRU $ VXEIXQFWRU ( RI ) LV D UXOH WKDW VHOHFWV DQ H[WUHPDO VXEREMHFW H; (; fÂ§! ); IRU HDFK ; 2EM;f VXFK WKDW LI I ; fÂ§!< WKHQ )If(;f & (< $Q\ VXFK DVVLJQPHQW ( RI ) JLYHV ULVH WR D IXQFWRU ( ; fÂ§! $ (I (; fÂ§! (< LV GHILQHG WR EH WKH FRPSRVLWLRQ (; fÂ§ )If(;f & (< 6LQFH H< (< fÂ§! )< LV PRQR (I LV XQLTXHO\ GHWHUPLQHG E\ WKH FRQGLWLRQ WKDW )I ); fÂ§!)< H; H< (; fÂ§fÂ§r(< FRPPXWHV 7KXV )If(;f & (< LPSOLHV WKDW (I FDQ EH GHILQHG WR PDNH WKH DERYH VTXDUH FRPPXWH
PAGE 53
7KH FRQYHUVH DOVR KROGV LI WKHUH LV D PDS (I VXFK WKDW WKH VTXDUH FRPPXWHV WKHQ )If(;f & (< 7R SURYH WKLV VXSSRVH WKHUH LV D PDS (I ZKLFK PDNHV WKH VTXDUH FRPPXWH 8VH WKH XQLTXH IDFWRUL]DWLRQ (I DE ZKHUH D $ fÂ§ (< LV H[WUHPDO PRQR DQG E LV HSL 1RWH WKDW )IfH;f H ^H)If(;f r )7@ DQG )If(;f & (< 7KXV D VXEIXQFWRU LV H[DFWO\ fD QDWXUDO WUDQVIRUPDWLRQ ZKRVH FRPSRQHQWV DUH DOO H[WUHPDO PRQRPRUSKLVPVf /HW 6XEIXQ)f GHQRWH WKH FODVV RI VXEIXFWRUV RI ) ; fÂ§! $ 'HILQH D SUHRUGHU RQ 6XEIXQ)f E\ (? & ) : f 2EM$f (L^$f & ($f ,I (L & ( WKHQ IRU HDFK $ WKHUH LV D XQLTXH H[WUHPDO PRQRPRUSKLVP F$ VXFK WKDW H?$ fÂ§ H$fF$f 2QH HDVLO\ YHULILHV WKDW F$ DUH WKH FRPSRQHQWV RI D QDWXUDO WUDQVIRUPDWLRQ F (? fÂ§! ( LQ IDFW F (? fÂ§! ( LV D VXEIXQFWRU 7KH IROORZLQJ FRQVWUXFWLRQV VKRZ WKDW WKH FODVV 6XEIXQ)f EHKDYHV YHU\ PXFK OLNH 6XEMf &RQVWUXFWLRQ 6XEIXQ)f LV FRPSOHWH ,I ^(L L f ` LV DQ\ FODVV RI VXE IXQFWRUV WKHUH LV D VXSUHPXP )Â DQG LQILPXP 3f (W 7KH VXSUHPXP LV JLYHQ E\ ,(Lf$f fÂ§ 8)MMf 7KH LQILPXP LV JLYHQ E\ 3)Âf$f 3_ (L$f 3URRI 6LQFH $ LV H[WUHPDOO\ ZHOOSRZHUHG IRU HDFK $ WKH FODVV ^(I$f L f ` KDV D UHSUHVHQWDWLYH VHW 7KXV WKH REMHFWE\REMHFW GHILQLWLRQV IRU _) DQG IM ([ PDNH VHQVH ,W LV REYLRXV WKDW LI 8 )Â DQG Sf )Â DUH VXEIXQFWRUV WKDW WKH\ DUH WKH RSWLPXP ERXQGV LQ WKH VXEIXQFWRU ODWWLFH /HW I ; fÂ§r< EH DQ\ PDS $ K! 3_Â (;$f LV D VXEIXQFWRU EHFDXVH IRU HDFK L f )f3_ ($ff & )If?(L$ff & (L%f
PAGE 54
VR )QDLAF D ee"f !fÂ§ (L$f LV D VXEIXQFWRU EHFDXVH )f/IÂ$ff 8)IfO(IL$ff & 8Af L L D &RQVWUXFWLRQ 6XSSRVH D ) fÂ§ LV D QDWXUDO WUDQVIRUPDWLRQ 7KHQ WKHUH LV DQ RUGHU WKHRUHWLFf DGMRLQW FRQHFWLRQ D 6XEIXQ)f m! 6XEIXQ*f D 7KH OHIW DGMRLQW D LV GHILQHG E\ D(f$f D$f($f 7KH ULJKW DGMRLQW D LV GHILQHG E\ Da(f$f Da($f 3URRI 7KH RUGHU RQ 6XEIXQ)f LV GHILQHG fREMHFWE\REMHFWf DQG ZKHQHYHU I $ % I DQG IaO IRUP DQ DGMRLQW FRQQHFWLRQ EHWZHHQ 6XEf DQG 6XE)f 6R LW VXIILFHV WR FKHFN WKDW D( DQG DaO( GHILQH VXEIXQFWRUV
PAGE 55
/HW ( f 6XEIXQ)f ; fÂ§! < EH D PDS DQG FRQVLGHU WKH IROORZLQJ GLDJUDP Â‘D$f($f D%f(%f 7KH WUDSH]RLGV Lf DQG LLLf DUH REWDLQHG E\ IDFWRULQJ D%fH%f DQG D$fH$f UHVSHFWLYHO\ LQ HDFK FDVH LV WKH HSL SDUW DQG Hn LV WKH H[WUHPDO PRQR SDUW 7KH VTXDUH LYf H[SUHVVHV WKH QDWXUDOLW\ RI D 7KH WUDSH]RLG LLf H[SUHVVHV WKDW H ( fÂ§r ) LV D VXEIXQFWRU 7R VKRZ D( LV D VXEIXQFWRU RI LW VXIILFHV WR VKRZ WKHUH LV D PDS N D$f?($f r D%f(%f WKDW PDNHV WKH WRS WUDSH]RLG FRPPXWH )RU WKLV XVH WKH GLDJRQDOL]DWLRQ SURSHUW\ IURP /HPPD 'HILQH ]%(If J *IfHn$ H ]$ DQG P Hn% QRWH WKDW JH fÂ§ PI ZLWK H HSL DQG P H[WUHPDO PRQR 7KXV WKH GLDJRQDOL]DWLRQ SURSHUW\ JXDUDQWHHV WKH GHVLUHG N H[LVWV $VVXPH $ fÂ§ % $ VLPLODU GLDJUDP LV XVHG WR YHULI\ WKDW DB( LV D VXEIXQFWRU RI ) ZKHQHYHU ( H 6XEIXQ*f 7KH PLVVLQJ PDS D$f ^($f fÂ§! D%f ;^(%f LV REWDLQHG IURP WKH SXOOEDFN SURSHUW\ RI D%fa(%f Â’
PAGE 56
&RUROODU\ /HW D ) fÂ§! EH D QDWXUDO WUDQVIRUPDWLRQ DQG H ( fÂ§! ) EH D VXEIXQFn WRU RI ) /HW Hn D( fÂ§! ) EH WKH VXEIXQFWRU GHVFULEHG LQ &RQVWUXFWLRQ 7KHUH LV D QDWXUDO WUDQVIRUPDWLRQ ] D( fÂ§! ) VXFK WKDW IRU HDFK REMHFW $ D$f fÂ§ Hn$f]$f LV WKH HSL H[WUHPDO PRQRfIDFWRUL]DWLRQ RI D$ 3URRI 2QH GHILQHV Hn DQG DV LQ WKH SUHFHGLQJ SURRI ([DPLQLQJ WKH FRPSDULVRQ RI VTXDUHV GLDJUDP XVHG WR SURGXFH D( VKRZV WKDW DQG H DUH QDWXUDO Â’ &RQVWUXFWLRQ $VVXPH ()*+ DUH IXQFWRUV $ fÂ§! $ 6XSSRVH WKDW D ( fÂ§! ) DQG fÂ§! + DUH VXE IXQFWRUV (LWKHU ( RU ) SUHVHUYHV H[WUHPDO PRQRPRUSKLVPV WKHQ WKH KRUL]RQWDO FRPSRVLWLRQ D (* fÂ§r )+ LV D VXEIXQFWRU 3URRI 7KH KRUL]RQWDO FRPSRVLWLRQ RI QDWXUDO WUDQVIRUPDWLRQV LV DOZD\V QDWXUDO +\SRWKHVLV LPSOLHV WKDW IRU DQ\ $ D$ LV DQ H[WUHPDO PRQRPRUSKLVP EHFDXVH WKH FODVV RI H[WUHPDO PRQRPRUSKLVPV LV FORVHG XQGHU FRPSRVLWLRQ DQG Df$ D+$f($f )$fD*$f E\ GHILQLWLRQ RI KRUL]RQWDO FRPSRVLWLRQ VHH &RQVWUXFWLRQ f Â’ 0HVHJXHUfV /HPPDV /HPPD 6XSSRVH $ LV DQ 6)FDWHJRU\ ^7SUIf LV D PRQDG RQ $ DQG ) LV D VXEn IXQFWRU RI7 VXFK WKDW U@$f^$f & )$ 7KHUH LV D VPDOOHVW VXEIXQFWRU ) RI 7 VXFK WKDW ) & ) & 7 DQG IRU DOO $ f 2EM\Lf S$f)$f & )$ ,I WKH HTXDWLRQ IM$f*$f & *$ KROGV IRU ZH VD\ LV FORVHG XQGHU S
PAGE 57
3URRI /HW 7 GHQRWH WKH FODVV RI DOO VXEIXQFWRUV RI 7 WKDW DUH ODUJHU WKDQ ) DQG FORVHG XQGHU S 6n LV QRQHPSW\ EHFDXVH LW FRQWDLQV 7 %\ &RQVWUXFWLRQ ) 3_7 H[LVWV 6LQFH 6 S$f6f LV DQ RUGHU SUHVHUYLQJ PDS 6XE7Yf fÂ§r 6XE/f VR IRU HDFK $ S$f)$f & )$ Â’ 7KH SUHFHGLQJ JLYHV DQ HDV\ FDQGLGDWH IRU WKH IXQFWRU SDUW RI WKH PRQDG JHQHUDWHG E\ ) 2QH QHHGV D PRUH FRPSOH[ DUJXPHQW LI RQH ZDQWV GHWDLOHG LQIRUPDWLRQ DERXW WKH QDWXUDO WUDQVIRUPDWLRQV UHODWHG WR Uc DQG L ZKLFK PDNH ) LQWR D PRQDG 7KH SURRIV RI WKH IROORZLQJ OHPPDV UHTXLUH GHWDLOHG FRPSXWDWLRQ )RU FODULW\ WKH JRDO RI HDFK SDUDJUDSK LQ WKH SURRI LV ZULWWHQ LQ EROGIDFH /HPPD $VVXPH QRWDWLRQ DQG K\SRWKHVHV IURP /HPPD ,If 7KHUH DUH QDWXUDO WUDQVIRUPDWLRQV LGA fÂ§} ) DQG P ) fÂ§! ) ZKLFK PDNH ) Q Pf D PRQDG 3URRI 2QH GHILQHV IRXU VHTXHQFHV RI QDWXUDO WUDQVIRUPDWLRQV I? )? fÂ§! 7 WKH VXEIXQFWRU JHQHUDWHG DW VWDJH $ Q? LGA fÂ§} )? D QDWXUDO WUDQVIRUPDWLRQ REWDLQHG IURP UM E\ VXLWDEO\ PRGLI\LQJ WKH GRPDLQ DQG FRGRPDLQ P? )I fÂ§! )? D QDWXUDO WUDQVIRUPDWLRQ REWDLQHG IURP S E\ VXLWDEO\ PRGLI\LQJ GRPDLQ DQG FRGRPDLQ DQG F? )? fÂ§! )?L WKH LQFOXVLRQ 'HILQH ) ) XVH WKH QRWDWLRQ I ) fÂ§r 7 P DQG &R DUH GHILQHG DFFRUGLQJ WR WKH VDPH SDWWHUQ WKDW GHILQHV ODWHU fPfV DQG fFfV ZKLFK LV GHVFULEHG EHORZ
PAGE 58
'HILQLQJ I;n)[fÂ§!7 $ f P$ )[r r )[L DQG F[ )[ fÂ§! )[L $VVXPH WKDW I[ KDV EHHQ SUHYLRXVO\ GHILQHG &RQVLGHU WKH IROORZLQJ GLDJUDP 0fr"f 8 );$ *LYHQ I[ DQG cL$ WKH VTXDUH LV REWDLQHG E\ HSL H[WUHPDO PRQRfIDFWRUL]DWLRQ RI WKH FRPn SRVLWH >L$fI[$f $OO PDSV LQ WKH WULDQJOH DUH H[WUHPDO PRQRPRUSKLVPV REWDLQHG E\ FRPSDULQJ VXEREMHFWV RI 7$ 'HILQH )[$ I[$f?)c$f8)[$ DQG P[ );$ fÂ§! )[L$ WR EH WKH PDS VKRZQ RQ WKH OHIW VLGH RI WKH GLDJUDP ,W IROORZV IURP &RQVWUXFWLRQV DQG WKDW $L )$ fÂ§ 7 LV D VXEIXQFWRU 1RWH WKDW ^I[LfP[$f Q$fIO$f DQG P[ LV QDWXUDO )RU ERRNNHHSLQJ SXUSRVHV OHW XV FDOO F? )[ fÂ§! )[ WKH VXEIXQFWRU ZKLFK H[KLELWV )[ & )[L ,I $W LV D OLPLW RUGLQDO GHILQH ). ?^)[ $ W`
PAGE 59
I.n).fÂ§!7 LV D VXEIXQFWRU E\ 'HILQLQJ WKH VHTXHQFH RI fQrf V 7KH IROORZLQJ GHILQLWLRQ RI Q[ LV QRW UHFXUVLYH Q[ FDQ EH GHILQHG RQFH ZH NQRZ I[ EXW WKH GHILQLWLRQ RI I? GRHV QRW LQYROYH Q[ DW DOO &RQVLGHU WKH IROORZLQJ FRPPXWDWLYH GLDJUDP RI IXQFWRUV DQG QDWXUDO WUDQVIRUPDWLRQV 7KH QDWXUDO WUDQVIRUPDWLRQV ] DQG H DUH GHILQHG E\ FRQGLWLRQ WKDW UM H] LV WKH HSL H[WUHPDO PRQRfIDFWRUL]DWLRQ RI DV GHVFULEHG LQ %\ DVVXPSWLRQ DQG WKH FRQVWUXFWLRQ RI WKH VHTXHQFH )[f UfLG$f &)& )[ /HW L? GHQRWH WKH QDWXUDO WUDQVIRUPDWLRQ VXFK WKDW H I[L;f L[ H[LVWV EHFDXVH H & L 'HILQH Q[ L?] (YLGHQWO\ 9 I;Q[ DQG ADL F[Q[ :KDW KDSSHQV ZKHQ WKH VHTXHQFH WHUPLQDWHV 6LQFH $ LV H[WUHPDOO\ ZHOOSRZ HUHG IRU HDFK $ WKH VHTXHQFH )?$f? RI VXEREMHFWV RI 7$ HYHQWXDOO\ WHUPLQDWHV VD\ ZKHQ $ m $JDLQ XVLQJ H[WUHPDO ZHOOSRZHUHGQHVV WKHUH LV DQ RUGLQDO VD\ N VXFK WKDW WKH VHTXHQFH )?).$ff[ RI VXEREMHFWV RI 7).O$ WHUPLQDWHV DW N 'HILQH Q. I fÂ§ I. DQG P P. 7KHVH DVVLJQPHQWV JLYH QDWXUDO WUDQVIRUPDWLRQV WR FKHFN WKLV RQH FRQVLGHUV D PDS $ fÂ§r % DQG FKRRVHV N ODUJH HQRXJK WKDW WKH VXEREMHFW VHTXHQFHV GHVFULEHG DERYHf WHUPLQDWH IRU ERWK $ DQG % ,W VKRXOG EH FOHDU WKDW HDFK VXEREMHFW VHTXHQFH ^);$f[ WHUPLQDWHV DW )$
PAGE 60
9HULI\LQJ WKDW ) Q Pf LV D PRQDG 7R SURYH WKH XQLW ODZV LW VXIILFHV WKDW IRU DOO $ nP[$f^)[Q[$f P[$fQ[)[$f F[$ EHFDXVH RQFH WKH VHTXHQFH WHUPLQDWHV WKH fFD$fV EHFRPH LGHQWLW\ PDSV RQ )?$f )L[ DQ\ RUGLQDO $ 1DWXUDOLW\ RI I[ DQG Q[ LPSOLHV 7Q[$fI[$f I[)[$f)[Q[$f DQG &Q[7$f^I[$f )[I[$fQ[)[$f 1RWH WKDW I[Q[ DQG E\ GHILQLWLRQ RI KRUL]RQWDO FRPSRVLWLRQ IO$ 7I[$fI[)[$f I[7$f)[I[$f &DOFXODWLQJ RQH ILQGV IO$f^)[Q[$f I[7$f)[I[$f)[Q[$f GHI QRWHG DERYH I[7$f)[U@$f QRWHG DERYH 7UM$fI[$f D LV QDWXUDO 7KXV 8O$f^)[Q[$f 79$fI[$f DQG E\ D VLPLODU FRPSXWDWLRQf I[$fQ[)[$f U7$fI[$f 7R VKRZ P[$f)[Q[$f F[$ FRQVLGHU WKH IROORZLQJ GLDJUDP )$,Â‹A8 )O$
PAGE 61
WKH XSSHU VTXDUH FRPPXWHV E\ DQ LGHQWLW\ SURYHG DERYH WKH ORZHU VTXDUH LV WKH GHILQLWLRQ RI P\ WKH XSSHU WULDQJOH LV WULYLDO WKH ORZHU WULDQJOH LV WKH XQLW ODZ IRU 7 Uc Lf 5HDGLQJ WKH SHULPHWHU RI WKH GLDJUDP RQH ILQGV I[ $ I[$fP[$f)[Q[$f %\ XQLTXHQHVV RI PDSV PDQLIHVWLQJ LQHTXDOLWLHV EHWZHHQ VXEREMHFWV F[$ P[$f^)[Q[$f 7KH GLDJUDP QHHGHG WR SURYH P[$fQ[)[$f F[$ LV VLPLODU DQG RPLWWHG 7KLV HVWDEOLVKHV WKH XQLW ODZV 7R SURYH WKH DVVRFLDWLYH ODZ IRU )Pf RQH PD\ FKRRVH VXIILFHQWO\ ODUJH $ WKHQ GUDZ D GLDJUDP FRPSDULQJ WKH DVVRFLDWLYH VTXDUHV IRU 7 DQG ) XVLQJ DSSURSULDWH SRZHUV RI I[ WR FRPSDUH WKH FRUQHUV 7KH FRPSDULVRQ VTXDUHV FRPPXWH EHFDXVH RI WKH GHILQLWLRQV RI P[ DQG KRUL]RQWDO FRPSRVLWLRQ Â’ 'HILQLWLRQ $VVXPH WKDW $ LV FRFRPSOHWH ,Q $ XQLRQV RI FKDLQV DUH FROLPLWV LI ZKHQHYHU LV D GLDJUDP ZKHUH 2EM'f FRQVLVWV RI D FKDLQ RI H[WUHPDO VXEREMHFWV RI $ DQG 0DS'f FRQVLVWV RI DOO LQFOXVLRQV LQ 6XE$ff WKDW H[LVW DPRQJ HOHPHQWV RI 2EM'f WKH PDS F FROLP' fÂ§! $ LQGXFHG E\ FROLPLW SURSHUWLHV LV DQ H[WUHPDO PRQRPRUSKLVP ,Q WKLV FDVH F FROLP' fÂ§! $ LV DQ H[WUHPDO VXEREMHFW ZKLFK FRQWDLQV HDFK HOHPHQW RI 2EM'f WKXV WKHUH LV DQ H[WUHPDO PRQRPRUSKLVP FROLP' fÂ§! 2EM'f 6LQFH 2EM'f LV WKH VXSUHPXP RI 2EM'f LQ WKH 6XE$f WKLV PDS PXVW EH DQ LVRPRUSKLVP 5HPDUN 7KH FRQGLWLRQ WKDW XQLRQV RI FKDLQV DUH FROLPLWV LV VDWLVILHG LQ PDQ\ QDWn XUDOO\ RFFXULQJ FDWHJRULHV )RU H[DPSOH LQ fUHODWLRQDOf FDWHJRULHV OLNH 7RS 3 DQG 6HW H[WUHPDO VXEREMHFWV DUH MXVW VXEVHWV ZLWK WKH LQGXFHG VWUXFWXUH VR DOO XQLRQV DUH FROLPLWV $VVXPH $ LV D FDWHJRU\ RI ILQLWDU\ DOJHEUDV 7KH VHW WKHRUHWLFfXQLRQ RI D FKDLQ $Lf RI VXEREMHFWV LV D VXEREMHFW 7KH VHW WKHRUHWLFfXQLRQ $n RI $f OLHV LQVLGH D VPDOOHVW
PAGE 62
H[WUHPDO VXEREMHFW ZKLFK E\ RXU QRWDWLRQ LV $Â (YLGHQWO\ WKH FRSURGXFW RI WKH FKDLQ $Âf RI VXEREMHFWV LV $n 6R WKH LVVXH DERXW ZKHWKHU fXQLRQV RI FKDLQV DUH FRSURGXFWVf DPRXQWV WR FKHFNLQJ ZKHWKHU $ $ 7KLV GHSHQGV RQ WKH GHOLFDWH LVVXH RI ZKHWKHU HSLPRUSKLVPV DUH VXUMHFWLYH ,I HSLV DUH QRW QHFHVVDULO\ VXUMHFWLYH WKHQ $ ZLOO EH WKH ODUJHVW VXEREMHFW LQ ZKLFK $n LV HSL ZKLFK ZLOO JHQHUDOO\ EH ODUJHU WKDQ $n 'HILQLWLRQ $V DERYH DVVXPH $ LV DQ 6)FDWHJRU\ LV D PRQDG DQG ) LV D VXEIXQFWRU RI 7 ZKLFK FRQWDLQV U" LGA $ SDUWLDO DOJHEUD $ Df LV DQ REMHFW $ HTXLSSHG ZLWK D PDS D )$ fÂ§! $ VXFK WKDW DQR$f LGA /HW S$OJ)RQRf GHQRWH WKH FDWHJRU\ RI DOO SDUWLDO DOJHEUDV ZLWK PDSV $ Df fÂ§! % Ef VXFK WKDW $ fÂ§! % LV DQ $PDS DQG ID fÂ§ E)RIf /HPPD &RQWLQXH ZLWK VLWXDWLRQ DQG K\SRWKHVHV IURP I 6XSSRVH $ LV FRFRPSOHWH DQG XQLRQV RI FKDLQV DUH FROLPLWV 6XSSRVH $ Df LV D SDUWLDO DOJHEUD 7KH SDUWLDO DOJHEUD VWUXFWXUH D H[WHQGV WR DQ )DOJHEUD VWUXFWXUH PDS D )$ fÂ§! $ LI DQG RQO\ LI IRU HDFK RUGLQDO $ WKHUH LV D XQLTXH D[? PDNLQJ WKH GLDJUDP EHORZ FRPPXWH )O$f fÂ§fÂ§ P? r); $f );$ffÂ§A $ 3URRI 7R SURYH D PDS D )$ fÂ§! $ H[LVWV XVH WUDQVILQLWH LQGXFWLRQ 7KH K\SRWKHVLV JLYHV D[ DW VXFHVVRU RUGLQDOV 7R FRQVWUXFW D[ ZKHQ $ LV D OLPLW RUGLQDO XVH WKH FRSURGXFW SURSHUW\ RI );$ fÂ§ 8mD )T$f ,QGXFWLYHO\ RQH VKRZV WKDW H[Wf IRU HDFK $ D[ D$LF$$f HDFK D[L H[WHQGV WKH SUHFHGLQJ D[f XQLWf IRU HDFK $ D[Q[$f N8
PAGE 63
,I XQLWf KROGV IRU VXFHVVRU RUGLQDOV WKHQ LW KROGV IRU DOO RUGLQDOV 6XSSRVH N LV D OLPLW RUGLQDO 7KH PDS D. LV HSL EHFDXVH LW LV D OLPLW RI HSLPRUSKLVPV 1RWH WKDW D. DQG D.Q.D. DUH ERWK FRPSDWLEOH PDSV ).$ fÂ§! $ WKH GHILQLWLRQ RI FROLPLWV LPSOLHV WKDW EHFDXVH D. LV HSL LGA D.Q. 7KH DOJHEUD DVVRFLDWLYH ODZ KROGV IRU A Df ,I D fÂ§ D$ WKHQ D$ D$ VR E\ WKH K\SRWKHVLV UHJDUGLQJ WKH H[LVWHQFH RI D[ VXFK WKDW D?LUUL?$f D?)?D?f SURYHV WKH DVVRFLDWLYH ODZ ,I XQLWf KROGV IRU $ WKHQ H[Wf KROGV IRU $ 1RWH WKDW WKH HTXDWLRQ D??UQ?$f fÂ§ D?^)?D?f GHILQHV D? %\ LQGXFWLRQ K\SRWKHVLV XQLWf ^)?D?f)?Q?$f fÂ§ LGUAA WKHUHIRUH SUHFRPSRVLQJ ERWK VLGHV RI WKH GHILQLQJ HTXDWLRQ IRU D$L ZLWK )?UL?$f VKRZV D$ D[LP?$f)[Q[$f D[L^F[$f 7KH VHFRQG HTXDOLW\ IROORZV IURP P[$f^)[Q[$f F[$f ZKLFK ZDV SURYHG LQ YHULI\LQJ WKH XQLW ODZV IRU ) ,I XQLWf KROGV DW $ WKHQ LW KROGV IRU $ &RQVLGHU WKH IROORZLQJ GLDJUDP 7KH OHIW WULDQJOH FRPPXWHV EHFDXVH XQLWf KROGV IRU $ WKH ERWWRP WUDSH]RLG FRPPXWHV E\ GHILQLWLRQ RI D[ WKH WRS WUDSH]RLG FRPPXWHV EHFDXVH Q[$f)[Q[$fP[$f Q[$fF[$f Q[L$
PAGE 64
7KH ILUVW HTXDOLW\ KROGV E\ SURRI RI WKH XQLW ODZ IRU ) WKH VHFRQG HTXDOLW\ KROGV EHFDXVH RI WKH FRPSDWLELOLW\ EHWZHHQ fQ$fV 7KXV D$LQ$Lf D[LP[$f)[Q[$fQ[$f D[)[D[f)[Q[$fQ[$f D[Q[$f LG D WKLV HVWDEOLVKHV WKH XQLW ODZ IRU $ Df Â’ 8VH WKH QRWDWLRQ $OJ) QRf WR GHQRWH WKH IXOO VXEFDWHJRU\ RI S$OJ)R QRf FRQWDLQLQJ REMHFWV VDWLVI\LQJ WKH K\SRWKHVHV RI /HPPD /HPPD 6XSSRVH WKH SUHFHGLQJ OHPPDV DSSO\ $OVR VXSSRVH HDFK P[ LV HSL 7KH PDS $ Df LfÂ§! $ Df LQGXFHV DQ HTXLYDOHQFH RI FDWHJRULHV WKH IXQFWRUV LQYROYHG DUH ( $OJ) Qf rf $7 GHILQHG E\ H[WHQVLRQ RI VWUXFWXUH DQG )$)fÂ§! $OJ)Qf JLYHQ E\ UHVWULFWLRQ 3URRI ( LV OHJLWPDWHO\ GHILQHG /HPPD GHILQHV ( RQ REMHFWV 0DSV LQ $) DUH $PDSV FRPSDWLEOH ZLWK WKH VWUXFWXUH 6R LW VXIILFHV WR VKRZ WKDW IRU DQ\ $OJ)RQf PDS If $ % );$ );% E[ %
PAGE 65
FRPPXWHV IRU HDFK RUGLQDO $ %\ GHILQLWLRQ RI $OJ) Qf WKH VTXDUH FRPPXWHV IRU $ 6XSSRVH LW FRPPXWHV IRU $ QRWH WKDW MfD[LP[$f !D?)[D?f GHI DAL E[)[W!f)[D[f LQG K\S E[)[E[f)tf LQG K\S E[^P[%f^)O!f GHI E[L E[L^)[L!f^P[$f QDWXUDOLW\ P[ %HFDXVH P$ LV HSL RQH FRQFOXGHV SD?L fÂ§ E?L)?LM!f $W OLPLW RUGLQDOV SURSHUWLHV RI FROLPLWV LQVXUH WKDW WKH GLDJUDP FRPPXWHV %\ WUDQVILQLWH LQGXFWLRQ DQG WKH GHILQWLRQ RI WKH H[WHQGHG VWUXFWXUH DQ\ PDS ZKLFK SUHVHUYHV )R QfVWUXFWXUH SUHVHUYHV )VWUXFWXUH 2QH DOVR QRWHV WKDW UHVWULFWLRQ UHVSHFWV PDSV EHFDXVH )R LV D VXEIXQFWRU RI ) ( DQG 5 IRUP DQ HTXLYDOHQFH 2EYLRXVO\ 9$ Df 2EM$OJ)R Qff 5($DRf fÂ§ $Df 7KH FRQVWUXFWLRQ RI WKH H[WHQGHG VWUXFWXUH VKRZV WKH UHVWULFWLRQ RI DQ\ )VWUXFWXUH WR )R XQLTXHO\ GHWHUPLQHV WKH )VWUXFWXUH KHQFH (5$ Df $ Df Â’ $ 3DUWLDO $OJHEUD :KLFK 'RHV 1RW ([WHQG 7KH FDWHJRU\ 6HW LV DQ 6)FDWHJRU\ DQG XQLRQV DUH FROLPLWV VR 0HVHJXHUfV /HPPD DSSOLHV WR 6HW :H VKRZ WKH QHFHVVLW\ RI /HPPD fV K\SRWKHVLV WKDW D PDS D?M VXFK WKDW D[?P? D?)?D?f FDQ EH GHILQHG 7KH VHFWLRQ GLVFXVVHV D VXEIXQFWRU RI WKH IUHHPDJPD PRQDG ZKLFK KDV DQ DOJHEUD WKDW FDQQRW H[WHQG WR D PRQDG DOJHEUD $ PDJPD LV D VHW ZLWK D ELQDU\ RSHUDWLRQ VXEMHFW WR QR HTXDWLRQV 7KH IUHH PDJPD PRQDG 7 UM cLf KDV WKH IROORZLQJ SDUWV *LYHQ D VHW ; 7; FRQVLVWV RI DOO ZRUGV ZLWK YDULDEOHV LQ $ $ PDJPD ZRUG LV DQ\ H[SUHVVLRQ IRUPHG E\ ILQLWHO\ PDQ\ DSSOLFDWLRQV RI WKH UXOHV
PAGE 66
Lf ,I [ ; WKHQ ALV D PDJPD ZRUG LLf ,I V W DUH PDJPD ZRUGV WKHQ LV D PDJPD ZRUG WKH SURGXFW RI V DQG W )RU HDVH LQ UHDGLQJ ZH XVH ELQDU\ WUHH QRWDWLRQ IRU SURGXFWV 7 GHILQHV D IXQFWRU JLYHQ ; fÂ§! < WR FRPSXWH 7I ZH DSSO\ WR DOO PHPEHUV RI ; LQ D JLYHQ ZRUG OHDYLQJ WKH WUHH DQG FLUFOH VWUXFWXUH XQFKDQJHG 7KH fLQVHUWLRQ RI YDULDEOHVf PDS U@; ; fÂ§! 7; UM;f[f A 7KH fVHPDQWLF FRPSRVLWLRQf PDS >L; 7; fÂ§! 7; VHQGV D ZRUG V 7; RI ZRUGV WR D ZRUG LQ 7; E\ UHPRYLQJ WKH FLUFOHV DURXQG HDFK HOHPHQW RI 7; XVHG LQ PDNLQJ V Q LV DOVR D QDWXUDO WUDQVIRUPDWLRQ 7KH QRWDWLRQ WDNHV D OLWWOH ZKLOH WR VRDN LQ WR H[SHGLWH WKH SURFHVV ZH FRQVLGHU D FDOFXODWLRQ ZLWK 71 6XSSRVH V 71 LV WKH ZRUG DQG WL VV 71f LV WKH ZRUG
PAGE 67
WKHQ Â1fW[f LV 1RZ ZH GHILQH D VXEIXQFWRU ) RI 7 'HILQH WKH GHSWK GHSQf RI D QRGH Q LQ D ELQDU\ WUHH LQGXFWLYHO\ E\ WKH GHSWK RI WKH GHSURRWf LI D LV LPPHGLDWHO\ EHORZ E WKHQ WKH GHSDf GHSEf 'HILQH D OHDI WR EH D QRGH WKDW KDV QRWKLQJ EHORZ LW /HW ) FRQVLVW RI DOO URRWHG ODEHOHG ELQDU\ WUHHV LH PDJPD ZRUGVf VXFK WKDW WKH GHSWK RI HDFK OHDI LV WKH VDPH )RU H[DPSOH V DQG U1fWLf DUH LQ )1f EXW LV QRW EHFDXVH GHS^f DQG GHSOf GHSf (YLGHQWO\ ) LV D VXEIXQFWRU DQG IRU DQ\ ; Uc;f;f & );f %XW IRU DQ\ QRQHPSW\ ; S;f);ff LV QRW FRQWDLQHG LQ );f 2QH UHDGLO\ YHULILHV WKDW ) 7 1RZ ZH GHILQH DQ SDUWLDO DOJHEUD VWUXFWXUH RQ 1 WKDW GRHV QRW H[WHQG WR D 7DOJHEUD VWUXFWXUH $V LQ /HPPD D VWUXFWXUH PDS IRU D SDLU )UM LGA fÂ§! )f LV D PDS D )$ fÂ§! $ VDWLVI\LQJ LGÂ DUM$f 'HILQH Q ) 1f W 1RQ D WUHH U DV IROORZV LI WKH GHSWK RI HDFK OHDI RI U LV RGG WKHQ QUf LV WKH OHIWPRVW ODEHO LI WKH GHSWK RI HDFK OHDI RI U LV HYHQ RU ]HURf WKHQ QUf LV WKH ULJKWPRVW ODEHO )RU H[DPSOH QVf fÂ§ V GHILQHG DERYHf DQG QL1fWLf
PAGE 68
,Q RUGHU IRU Q )1f fÂ§! 1 WR H[WHQG WR Q )L1f PDNLQJ ) 1f )Q Â‘)^ 1f [1 ) 1f Q L Â‘r1 FRPPXWH 1R VXFK D FDQ H[LVW IRU ]1fWLf LV 1 WKHUH PXVW EH D IXQFWLRQ ULL ZKLFK HTXDOV L1fWf ZKHUH W LV )LQDOO\ QRWH WKDW )QfWLf LV VR Q)QfWf 1RWH WKDW )QfWf U1fWf VR Q)QfWf 7KXV ZH KDYH HOHPHQWV )1 WKDW [1 LGHQWLILHV DQG Q)Qf GRHV QRW LGHQWLI\ VR WKHUH LV QR IXQFWLRQ Q? VXFK WKDW Q?Q1f Q)Qf
PAGE 69
&+$37(5 )5(( $/*(%5$6 7KLV FKDSWHU H[SORUHV D JHQHUDOL]DWLRQ RI FRPSOHWH GLVWULEXWLYLW\ IRU 3AREMHFWV ,Q VHYHUDO VWHSV WKH IUHH FRPSOHWH GLVWULEXWLYH ODWWLFH PRQDG LV FRQVWUXFWHG WKHQ 0HVHJXHUfV /HPPDV DUH XVHG WR FRQVWUXFW DQ DSSURSULDWH VXEPRQDG ZKRVH DOJHEUDV JHQHUDOL]H FRPn SOHWHO\ GLVWULEXWLYH ODWWLFHV 6HFWLRQ GHVFULEHV PRQDGV IRU FRPSOHWH VHPLODWWLFHV LV WKH PRQDG IRU FRPSOHWH MRLQ VHPLODWWLFHV LV WKH PRQDG IRU FRPSOHWH PHHW VHPLODWWLFHV 7KH JLVW RI VHFWLRQ LV WKDW WKHUH LV D OLIWLQJ RI nOO RYHU 9 DQG D OLIWLQJ RI RYHU f 7KHUHIRUH 89 DQG 98 DUH PRQDGV RYHU 3 6HH UHVXOWV RI %HFN >@ VXPPDUL]HG LQ 6HFWLRQ f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f RI VXEVHW V\VWHPV JLYHV ULVH WR D VXEIXQFWRU RI ,Wn' 0HVHJXHUfV /HPPDV DUH DSSOLHG WR WKLV VXEIXQFWRU WR SURGXFH D PRQDG ) $Q\ 3AREMHFWV ZKLFK
PAGE 70
LV 3AHPEHGGDEOH LQ D FRPSOHWHO\ GLVWULEXWLYH FRPSOHWH ODWWLFH LV DQ )DOJHEUD $Q\ ) DOJHEUD KDV D QDWXUDO VWUXFWXUH %HFDXVH RI FRPSXWDWLRQDO GLIILFXOWLHV QR H[DFW DOJHEUDLF FKDUDFWHUL]DWLRQ RI )DOJHEUDV LV JLYHQ KHUH $ ZRUG DERXW QRWDWLRQ WKH IXQFWRU SDUWVf RI WKH PRQDGV GHVFULEHG EHORZ DUH JLYHQ E\ IDPLOLHV RI VHWV 7KXV FKHFNLQJ WKH XQLW DQG DVVRFLDWLYH ODZV UHTXLUHV ZRUNLQJ ZLWK PDQ\ OHYHOV RI WKH SRZHU VHW WRZHU 5RPDQ OHWWHUV 67 Â‘ Â‘ GHQRWH VHWV $ VXEVFLSW GHVLJQDWHV WKH fSRZHU VHW FRPSOH[LW\f 6L &3$f 6 6f$ 6 LV D IDPLO\ RI VHWV 6 $ 6 LV D IDPLO\ RI IDPLOLHV RI VHWV HWF &RPSOHWH VHPLODWWLFHV ,Q WKLV VHFWLRQ ZH GHVFULEH WKH IUHH FRPSOHWH MRLQ UHVS PHHWf VHPLODWWLFH RQ D SRVHW $ XVLQJ WKH PRQDG 9 G ILf UHVS n8L[ff 7KH GHVFULSWLRQ RI LV ZHOO NQRZQ IRU H[DPSOH 0HVHJXHU >@ XVHV LW 7KH UHDGHU ZLOO KDYH QRWLFHG WKDW T LV XVHG DV D QDPH IRU WZR GLIIHUHQW QDWXUDO WUDQVIRUPDWLRQV WKLV ZRXOG QRUPDOO\ EH KRUULEOH QRWDWLRQ EXW LQ WKLV FDVH WKH IRUPXOD IRU \ LV WKH VDPH 7KXV RXU QRWDWLRQDO HFRQRQ\ VKRXOG FDXVH QR FRQIXVLRQ 7KH IXQFWRUV nOO DQG 7! DFW RQ D SRVHW $ E\ $f ^6&$[!\6 e! [ ` WKH VHW RI LQFUHDVLQJ VXEVHWV RI $ RUGHUHG E\ UHYHUVH LQFOXVLRQ DQG 7!$f ^6&$[\H6 [ ` WKH VHW RI GHFUHDVLQJ VXEVHWV RI $ RUGHUHG E\ LQFOXVLRQ *LYHQ PRQRWRQH I $ fÂ§! % ZH GHILQH 9I n'$f r 9^%f 6 ^E e % V f 6 E IVf` DQG 8I 8$f } 8%f 6 A ^EH % V H 6E! I^Vf` 7KH IROORZLQJ IDFWV ZLOO EH RI ODWHU XVH
PAGE 71
W DQG DUH IXQFWRUV WULYLDOO\ WKH\ UHVSHFW LGHQWLW\ DUURZV 2EVHUYH WKDW IRU DQ\ 6 $ DQG PRQWRQH IXQFWLRQV $ fÂ§! % J % fÂ§! & 7! JIf6f ^Ff&VH6FJIVf` ^FH&IVfHn'I6fFJIVf` e"fef DQG VLPLODUO\ 8JIf6f 8Jf8If6f )RU HDFK SRVHW $ '$f LV D FRPSOHWH ODWWLFH ZLWK VXSUHPXP RSHUDWLRQ JLYHQ E\ VHW WKHRUHWLF XQLRQ DQG LQILPXP JLYHQ E\ LQWHUVHFWLRQ WW$f LV D FRPSOHWH ODWWLFH ZLWK LQILPXP JLYHQ E\ XQLRQ DQG VXSUHPXP JLYHQ E\ LQWHUVHFWLRQ )RU HDFK 6 & '$f ,Q SDUWLFXODU WKLV KROGV LI LV HPSW\ 6R 9I SUHVHUYHV DOO VXSUHPD 6LPLODUO\ LI 6 & n8$f WKHQ 8I?Mf ?M^8If6f6HH` VR WKDW 8I SUHVHUYHV DOO ÂQILPD 'HILQH G$ $ fÂ§! 7!$f D fÂ§! ^[ $ [ D` L$ $ fÂ§} n8$f D+^LeLL!D` DQG IL$ 7!^$f } 9$f 6 f .$f A 6 ,W PD\ EH SX]]OLQJ WKDW 7!$f LV RUGHUHG E\ VXEVHW LQFOXVLRQ DQG OLf LV RUGHUHG E\ UHYHUVH VXEVHW LQFOXVLRQ 7KH IDFW WKDW [ \ G$[f & G$\f L$\f & L$[f
PAGE 72
PRWLYDWHV WKH FKRLFH EHFDXVH RQH ZDQWV L$ WR EH RUGHU SUHVHUYLQJ 0RUHRYHU RQH ZDQWV LI WR SUHVHUYH DOO LQWLPD ZKLFK RQO\ KDSSHQV LI WKH LQILPXP RSHUDWLRQ LQ 8$ LV XQLRQ 7KH UHDGHU PD\ FKHFN WKDW L G DQG S DUH QDWXUDO WUDQVIRUPDWLRQV 7KH PRQDG DVVRFDWLYH ODZ 9\ Â‘' \9 Â‘9 KROGV IRU ERWK DQG nOO EHFDXVH LI 6 H 7Af WKHQ 88V 8^O!6AVV` 7KH PRQDG XQLW ODZV fÂ§fÂ§r fÂ§fÂ§ KROG IRU 9 EHFDXVH LI 6 7f$f WKHQ ^G$f[f [ f @ MU f 7!$f 7 & ` 6LPLODU FRPSXWDWLRQ VKRZV WKH XQLW ODZV KROG IRU 8 1RZ WR GHVFULEH DOJHEUDV RYHU WKHVH PRQDGV /HPPD /HW 7 & !$f VXFK WKDW IRU DOO [ ( $ G$f[f f 7 7KHQ WKH IROORZLQJ DUH HTXLYDOHQW WKHUH LV DQ RUGHU SUHVHUYLQJ PDS D 7 fÂ§! $ VXFK WKDW LGA DG$f HDFK 6 7 KDV D VXSUHPXP G$ $ fÂ§r 7!$ KDV D ULJKW DGMRLQW
PAGE 73
7KH DQDORJRXV FRQGLWLRQV LQYROYLQJ 7 & /$ DUH DOVR HTXLYDOHQW 3URRI :H JLYH WKH SURRI IRU 7 & !$ OHDYLQJ WKH XSVLGHGRZQ DUJXPHQW IRU 7 & 8$ WR WKH UHDGHU f 'HILQH D6f ?6 DQG FRPSXWH f D6f LV DQ XSSHU ERXQG IRU 6 EHFDXVH LI [ 6 WKHQ G$f[f & ZKHQFH [ DG$f[f D6f 6XSSRVH IRU DOO [ 6 [ X 7KHQ G$f^[f & G$fXf VR 6 8^J/f[f [ ` & G$fXf WKHUHIRUH Tf X f %\ GHILQLWLRQ RI 9 6 M [ & 6 r !f [ ?6 Â’ 3URSRVLWLRQ 3rr LV HTXLYDOHQW WR WKH FDWHJRU\ RI 7f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rr LW KDV VWUXFWXUH PDS D 7!$f $ 6 !r?6 ,Q /HPPD LW ZDV QRWHG WKDW D VDWLVILHV WKH XQLW ODZ IRU VWUXFWXUH PDSV D VDWLVILHV WKH DVVRFLDWLYH ODZ E\ WKH RUGHU WKHRUHWLF IDFW SURYHG LQ /HPPD EHORZ Â’ /HPPD )RU DQ\ 6 '$f 9 ^[ f $ 6 f 6 ; 96L` 98An 3URRI 7R YHULI\ WKLV HTXDOLW\ RQH QRWHV WKDW IRU DOO [ 6 VXFK WKDW [ f VR [ 9 6? 7KXV WKH OHIW KDQG VLGH GRPLQDWHV WKH ULJKW ,I X LV DQ XSSHU ERXQG IRU 6 DQG L 6 [ 9 6L WKHQ [ X 6R WKH OHIW KDQG VLGH LV WKH OHDVW XSSHU ERXQG RI >6 WKLV HVWDEOLVKHV WKH OHPPD Â’
PAGE 74
&RPSOHWHO\ 'LVWULEXWLYH &RPSOHWH /DWWLFHV ,Q WKH IROORZLQJ DUJXPHQWV WKH IRUJHWIXO IXQFWRUV DUH VXSUHVVHG IURP WKH QRWDWLRQ 7KH OHPPDV WKDW IROORZ HVWDEOLVK WKDW WKHUH LV D OLIWLQJ RI nOO RYHU 9 DQG VLPLODU DUJXPHQWV VKRZ WKDW WKHUH LV D OLIWLQJ RI 9 RYHU ,Wf ([SOLFLWO\ WKLV PHDQV )RU DQ\ 8DOJHEUD $ WKHUH LV D OWVWUXFWXUH RQ 9$f ,W PDSV 'DOJHEUD PDSV WR 'DOJHEUD PDSV VR ZH PD\ YLHZ nOO DV D IXQFWRU 3p fÂ§! 3p %RWK QDWXUDO WUDQVIRUPDWLRQV L DQG S 8 fÂ§! 8 SUHVHUYH DOO MRLQV VLPLODUO\ G DQG [ fÂ§! SUHVHUYH DOO PHHWV /HPPD )RU DQ\ SRVHW $ ERWK $f DQG 9^$f DUH FRPSOHWH ODWWLFHV 7KXV $f DQG '$f DUH ERWK ,WfÂ§ DQG 7!fÂ§ DOJHEUDV /HPPD )RU DQ\ SRVHW $ G$ $ fÂ§! '$f SUVHUYHV DOO H[LVWLQJ ÂQILPD L$ $ fÂ§} $f SUVHUYHV DOO H[LVWLQJ VXSUHPD 3URRI 6XSSRVH 6 & $ DQG X I?6 H[LVWV 6LQFH G$ LV RUGHU SUHVHUYLQJ G$fXf & 3_ MF/Uf [ )RU WKH UHYHUVH LQHTXDOLW\ VXSSRVH L I@^G$[f [ ( m6` LH L LV D ORZHU ERXQG RI 6 7KHQ O X VR e ( G$Xf 7KH SURRI IRU X$ LV XSVLGHGRZQ EXW RWKHUZLVH LGHQWLFDO Â’ /HPPD ,I I $ fÂ§r % SUHVHUYHV DOO VXSUHPD WKHQ ,W Af fÂ§! 8%f DOVR SUHn VHUYHV DOO VXSUHPD 6LPLODUO\ LI I SUHVHUYHV ÂQILPD VR GRHV 7!f 7KHUHIRUH IRU DQ\ RUGHU SUHVHUYLQJ PDS I ERWK 87!I DQG 9,,I SUHVHUYH DOO QILPD DQG VXSUHPD 3URRI $ SURRI RI WKH ILUVW IDFW LV JLYHQ WKH VHFRQG LV YHU\ VLPLODU EXW QRWDWLRQDOO\ HDVLHU 8VH WKH IDFW WKDW OW$f LV D FRPSOHWH VXEODWWLFH RI 7$f DQG WKHUHIRUH FRPSOHWHO\ GLVWULEXWLYH
PAGE 75
/HW 6?f[H/ EH DQ LQGH[HG VXEVHW RI OL$f IRU HDFK 6? FKRRVH D IDPLO\ [[. WKDW 6? 7 [?. Â‘ Nf .f 5HFDOO WKDW f MRLQV LQ ,/ DUH LQWHUVHFWLRQV f L$ SUHVHUYHV VXSUHPD ZKLFK UHDGV 7 ?\[ \[f 3_ 7 \? $ IRU DQ\ LQGH[HG IDPLO\ \[f & $ 2QH FRPSXWHV DV IROORZV 0QLEAf $ A/.Q F/N`nf M^X?[[F:f &/.@ nD 8 ^ W 9 rrrff 8 ^ 3L 7 ADADf F/AN` $ Q XIA &RUROODU\ )RU DQ\ SRVHW $ S 7!$f fÂ§! 7f$f SUVHUYHV DOO QILPD S OL$f fÂ§! 8$f SUVHUYHV DOO VXSUHPD N f .f VXFK Â’
PAGE 76
3URRI 6XSSRVH 6D e ;!Af IRU $ f / FODLP DPRXQWV WR XQVZ QXVZ $ $ &OHDUO\ WKH OHIW KDQG VLGH LV FRQWDLQHG LQ WKH ULJKW 6XSSRVH WKDW [ LV D PHPEHU RI WKH ULJKW KDQG VLGH WKDW LV IRU HDFK $ WKHUH H[LVWV 6LD ÂD ZLWK [ f 6LD 7KHQ 6L 3L 6LD $ LV LQ HDFK 6D EHFDXVH HDFK 6D LV GRZQZDUG FORVHG UHODWLYH WR WKH LQFOXVLRQ RUGHU LQ k$f 6LQFH D f 6L WKLV SURYHV WKDW [ LV LQ WKH ULJKW KDQG TXDQWLW\ DQG WKH GHVLUHG HTXDOLW\ RI VHWV KROGV 7KH SURRI IRU WKH ,W ZRUNV VLPLODUO\ Â’ 7KH SUHFHGLQJ OHPPDV VKRZ WKDW WKHUH DUH OLIWLQJV RI nOO RYHU k DQG YLFH YHUVDf VR ZH KDYH &RUROODU\ ,Wk ,W&' ZKHUH @ LGS fÂ§ ,Wk D L! ^6 f 9$ _D&6` DQG Y 8'OWk ,Wk 6 A XQr LV D PRQDG RYHU ,3 ZKRVH DOJHEUDV DUH FRPSOHWHO\ GLVWULEXWLYH FRPSOHWH ODWWLFHV 3URRI 7KH RQO\ WKLQJ OHIW WR EH SURYHG LV WKH IRUPXODV IRU WKH QDWXUDO WUDQVIRUPDWLRQV %\ %HFNfV UHDVRQLQJ DV VXPPDUL]HG LQ 6HFWLRQ J LV WKH KRUL]RQWDO FRPSRVLWLRQ L Â‘ G 7R FRPSXWH Y OWkOWk ,Wk RQH XVHV WKH IDFW WKDW Y$ LV WKH VWUXFWXUH PDS IRU WKH IUHH DOJHEUD 89$ &RQVLGHU WKH VLWXDWLRQ LQ OLJKW RI WKH GLVFXVVLRQ IROORZLQJ 7KHRUHP +HUH ,W SOD\V WKH UROH RI ZKLOH k SOD\V WKH UROH RI 7f 7KH UHODWLRQ D DU^7DVf DSSOLHG WR WKH REMHFW 89$ LPSOLHV Y$fVf XpQfrf 8^UH%$6HI_6UFI_6` 8L!
PAGE 77
7KH DVVHUWLRQ DERXW DOJHEUDV IROORZV IURP 7KHRUHP Â’ 5HPDUN 89 LV D PRQDG EHFDXVH WKHUH LV D OLIWLQJ RI 8 RYHU 3p 2QH FRXOG XVH WKH FRUUHVSRQGHQFH EHWZHHQ OLIWLQJV DQG GLVWULEXWLYH ODZV RXWOLQHG LQ WR ILQG D QDWXUDO WUDQVIRUPDWLRQ 98 fÂ§! 87! 7KLV GLVWULEXWLYH ODZ LV QRW QHHGHG IRU WKH FDOFXODWLRQV ZKLFK IROORZ DQG LV VRPHZKDW FXPEHUVRPH VR LWV H[SOLFLW GHVFULSWLRQ LV RPLWWHG 6RPH FDWHJRULHV RI DOJHEUDV ,Q WKLV VHFWLRQ 0HVHJXHUfV /HPPDV DUH DSSOLHG WR WKH PRQDGV GHVFULEHG DERYH 1RWH WKDW 3 WKH FDWHJRU\ RI SRVHWV DQG RUGHU SUHVHUYLQJ PDSV LV D FRFRPSOHWH 6)FDWHJRU\ LQ ZKLFK XQLRQV RI FKDLQV DUH FROLPLWV 1RWH WKDW H[WUHPDO PRQRPRUSKLVPV LQ 3 DUH LQFOXVLRQV RI VXEVHWV ZLWK WKH LQGXFHG RUGHU HDFK RUGHU SUHVHUYLQJ PDS $ fÂ§! % IDFWRUV DV $ fÂ§! I$f fÂ§\ % ZKHUH I$f LV WKH VHWWKHRUHWLF LPDJH RI $ ZLWK WKH RUGHU LQGXFHG IURP % 7KXV WKH UHVXOWV RI 6HFWLRQ GR DSSO\ WR 3 7KH UHDGHU LV DGYLVHG WR UHYLHZ WKH GHILQLWLRQ RI VXEVHW V\VWHP JLYHQ LQ LI QHFHVVDU\ $VVXPH DOO VXEVHW V\VWHPV = DUH QRQWULYLDO LQ WKH VHQVH WKDW IRU HDFK $ DQG D f $ ^D` f =$f WKLV GRHV QRW UHGXFH WKH JHQHUDOLW\ RI WKH DUJXPHQW EHFDXVH JLYHQ DQ\ VXEVHW V\VWHP RQH PD\ DGMRLQ DOO VLQJOHWRQV WR LW ZLWKRXW FKDQJLQJ ZKLFK RSWLPXP ERXQGV DUH SUHVHUYHG 7KHRUHP >@ 3 DQG 3P DUH PRQDGLF 3URRI *LYHQ D VXEVHW V\VWHP M $f ^L66HM$f` GHILQHV D VXEIXQFWRU RI 9 7KLV H[WHQGV WR D VXEIXQFWRU M RI 7! VXFK WKDW M G Of LV D PRQDG 7KH QDWXUDO WUDQVIRUPDWLRQV G DQG / DUH REWDLQHG IURP G DQG L E\ VXLWDEO\ PRGLI\LQJ WKH GRPDLQV DQG FRGRPDLQV 7KH SUHFLVH GHILQLWLRQ RI WKHVH PDSV LV FRQWDLQHG LQ WKH SURRI RI /HPPD
PAGE 78
,I $ Df KDV D VWUXFWXUH /HPPD VKRZV WKDW DR6f ? IRU DOO H R$f 7KXV WR GHILQH D[ RQH PXVW VKRZ VRPH PDS ;$ $ PDNHV WKH IROORZLQJ GLDJUDP FRPPXWH 6$ r9f r$ 8 ;$ /HPPD VKRZV WKDW ?6 LV GHILQHG IRU DOO 6 f ?$ DQG PRUHRYHU WKDW IRU 6 f T$ 98 6 9m9 fm!ff ,GHQWLFDO DUJXPHQWV VKRZ VKRZ WKDW IRU HDFK $ LI D?6f ?6 WKDW D$6f 9 ,W IROORZV WKDW DQ\ SRVHW $ LQ ZKLFK HDFK PHPEHU RI T$ KDV D VXSUHPXP HDFK PHPEHU RI $f KDV D VXSUHPXP 0RUHRYHU HDFK PDS SUHVHUYLQJ VXSUHPD DOVR SUHVHUYHV VXSUHPD 7KXV RQH REWDLQV DQ HTXLYDOHQFH RI FDWHJRULHV EHWZHHQ 3 DQG 3$Q XSVLGHGRZQ YHUVLRQ RI WKLV DUJXPHQW VKRZV WKDW 3P LV D FDWHJRU\ RI DOJHEUDV Â’ *LYHQ VXEVHW V\VWHPV P DQG M ) 0IRUPV D VXEIXQFWRU RI 7 ,/' %\ &RUROODU\ 7XUMf LV D PRQDG RQ 3 %\ WKHUH H[LVWV D VPDOOHVW ) DERYH ) WKDW LV FORVHG XQGHU Y 7KLV PRQDG LV XVHG WR GLVFXVV WKH FDWHJRULHV GHILQHG EHORZ
PAGE 79
'HILQLWLRQ DQG 5HPDUNV /HW '3A GHQRWH WKH FDWHJRU\ RI FRPSOHWHO\ GLVWULEXWLYH FRPSOHWH ODWWLFHV ZLWK PDSV ZKLFK SUHVHUYH RUGHU DQG DOO RSWLPXP ERXQGV '3A GHQRWHV WKH IXOO VXEFDWHJRU\ RI FRQWDLQLQJ REMHFWV VXFK WKDW WKHUH LV D 3P PDS c! $ fÂ§! % ZLWK 3H[WUHPDO PRQR DQG % f 2EM'3Af GHQRWHV WKH FDWHJRU\ RI )DOJHEUDV ZKHUH ) LV WKH PRQDG GHVFULEHG DERYH ,W ZLOO EH VKRZQ WKDW 6S)P & PL F 0O 7R VHH WKH ILUVW LQFOXVLRQ RQH QRWHV WKDW DQ\ IDPLO\ 7 RI VXEVHWV RI D VHW ; ZKLFK LV DOVR D 3A REMHFW LV LQ '3A WKH LQFOXVLRQ 7 & ;f LV D 3AHPEHGGLQJ DQG 7;f LV D FRPSOHWHO\ GLVWULEXWLYH FRPSOHWH ODWWLFH 3URSRVLWLRQ VKRZV WKH VHFRQG LQFOXVLRQ /HW $ %3A DQG c! $ fÂ§ % EH JLYHQ DV LQ WKH GHILQLWLRQ RI 23A OHW E 7% fÂ§! % EH WKH 7VWUXFWXUH IRU % 7KH YHULILFDWLRQ RI WKH K\SRWKHVLV RI SURFHHGV E\ FRPSDULQJ WKH GLDJUDP WR EH FRPSOHWHG ZLWK WKH OW'DOJHEUD DVVRFLDWLYLW\ GLDJUDP IRU % 3URP WKLV SRLQW RQ WKH QRWDWLRQ RI /HPPD LV DGRSWHG ZLWK VOLJKW PRGLILFDWLRQ Y 7 fÂ§ 7 LV WKH PRQDG PXOWLSOLFDWLRQ Y? )b fÂ§ )?L LV GHILQHG LQGXFWLYHO\ DQG SOD\V WKH UROH WKDW ZDV SOD\HG E\ P? LQ WKH SURRI RI f Uc LGS fÂ§ 7 LV WKH XQLW Uc? LGS fÂ§! )? DUH GHILQHG LQGXFWLYHO\ DQG SOD\ WKH UROH WKDW ZDV SOD\HG E\ Q? LQ 7KH QDWXUDO WUDQVIRUPDWLRQV I[f)?fÂ§r7 DQG F? )? fÂ§! )? SOD\ WKH VDPH UROHV DV LQ /HPPD 5HPDUN %\ /HPPD WKH VWUXFWXUH PDS E 7% fÂ§! % LV JLYHQ E\ $ ^9L Lf 0 6LQFH I! LV D 3APDS WKH IROORZLQJ GLDJUDP FRPPXWHV )G! I?% )$ fÂ§r)% !7% DR % %
PAGE 80
7KH OHIW VTXDUH H[SUHVVHV WKDW c! FRPPXWHV ZLWK WKH )VWUXFWXUH WKH ULJKW VTXDUH H[SUHVVHV WKDW WKH )VWUXFWXUH RQ % LV WKH UHVWULFWLRQ RI WKH 7VWUXFWXUH /HPPD 'HILQH DR )$ fÂ§! $ E\ D6f $&A 9;A2} WKLV PDS VDWLVILHV D7@R LGA 3URRI /HW XV EHJLQ E\ PRUH H[SOLFLWO\ FDOFXODWLQJ DR 8VLQJ WKH GHILQLWLRQ RI OL RQ PDSV ZH ILQG WKDW D6f $89frf I?^[H$6H6 ?6L[` ?^?66H6` 7KLV GHILQLWLRQ PDNHV VHQVH HDFK 6L T$f VR HDFK 9 6L H[LVWV EHFDXVH $ 3A 8 9f LV DQ RUGHU SUHVHUYLQJ PDS DQG 6 PRMR$f ZKLFK LPSOLHV WKDW 8?f6f H 0$f 7KXV WKH PHHW GHILQLQJ DR6f H[LVW 5HFDOO UM[f ^6 '$ [ & ` 1RZ RQH FDOFXODWHV DR}"rff ?^?66H'$O[46` [ Â’ )RU ODWHU FDOFXODWLRQV LW LV FUXFLDO WR NQRZ WKDW VRPH RI WKH PDSV DUH HSL RU PRQR /HPPD )RU HDFK RUGLQDO ; X?$ LV VXUMHFWLYH 7KHUHIRUH Y?$ LV 3HSL 3URRI %HFDXVH RI WKH VWUXFWXUH RI 3 XQLRQV RI VXESRVHWV DFWXDOO\ DUH VHW WKHRUHWLF XQLRQV /HW )? )? 8 L?$f)b$f ,I 6 Y?$f)I$f WKHQ WKHUH LV DQ 6 )b$ VXFK WKDW Y?$f6f 6 6LQFH F?$ Ln[$fU@[)[$f VR IRU DQ\ 6L )[$ 7 U@[)[$f6f LV D PHPEHU RI )I$ VXFK WKDW X[$f^7f fÂ§ 6L 7KLV OHPPD KROGV LQ DQ\ FDWHJRU\ ZKHUH XQLRQV RI VXEREMHFWV DFWXDOO\ DUH VHWWKHRUHWLF XQLRQV Â’
PAGE 81
$IWHU QRWLQJ WKDW H[WUHPDO PRQRPRUSKLVPV LQ 3 DUH LQMHFWLRQV $ fÂ§! % ZKHUH I$f KDV WKH RUGHU LQGXFHG DV D VXEVHW RI % RQH HDVLO\ YHULILHV /HPPD 7KH IXQFWRUV 8 7! 7 DQG )? DOO SUHVHUYH H[WUHPDO PRQRPRUSKLVPV 1RZ RQH FDQ YHULI\ WKH K\SRWKHVLV RI LV VDWLVILHG 3URSRVLWLRQ /HW $ f 2EM'3Af % H 2EM%3Af I! $ fÂ§! % DV LQ WKH GHILQLWLRQ RI %3Â 7KHQ H[f IRU HDFK $ WKHUH LV D XQLTXH PDS f$L f )?L$ $ VXFK WKDW D?)?D?f fÂ§ DL,nD FRPf IRU HDFK $ ILD[ EI?%f)?!f 7KXV 23A & 3URRI 5HPDUN HVWDEOLVKHG FRPf IRU $ ,I FRPf KROGV IRU VXFHVVRU RUGLQDOV SURSHUWLHV RI XQLRQV LQVXUH WKDW FRPf KROGV DW DOO RUGLQDOV
PAGE 82
6XSSRVH FRPf KROGV DW VWDJH $ ZH ZLOO HVWDEOLVK FRPf DQG H[f KROG IRU VWDJH $ &RQVLGHU WKH IROORZLQJ GLDJUDP 7% !7% 7KH GLDJUDP FRPPXWHV WKH RXWHU VTXDUH FRPPXWHV EHFDXVH RI WKH 7DOJHEUD DVVRFLDWLYH ODZ IRU %? WKH OHIW RXWHU WUDSH]RLG FRPPXWHV EHFDXVH RI WKH GHILQLWLRQ RI Y?%? FRPPXWDWLYLW\ RI WKH ULJKW RXWHU WUDSH]RLG LV REYLRXV WKH OHIW LQQHU WUDSH]RLG H[SUHVVHV QDWXUDOLW\ RI Y?? WKH WRS DQG ULJKW LQQHU WUDSH]RLGV FRPPXWH E\ FRPf (VWDEOLVKLQJ FRPPXWDWLYLW\ RI WKH WRS RXWHU WUDSH]RLG UHTXLUHV D ELW PRUH 6LQFH I[ )? fÂ§} 7 LV QDWXUDO 7EfI[7%f fÂ§ I[%f)[Ef %\ GHILQLWLRQ RI KRUL]RQWDO FRPSRVLWLRQ IO% I[7%f)[I[%f 7KXV &7Ef^IO%f 7EfI[7%f)[I[%f I[%f)[Ef)[I[%f 7KLV SURYHV WKDW WKH WRS RXWHU VTXDUH FRPPXWHV &RQVLGHU WKH HSLH[WUHPDO PRQRfIDFWRUL]DWLRQ RI EI[%f)[LW!\
PAGE 83
VD\ LW IDFWRUV DV )[O$ $Ln&% 6LQFH LV HSL WKH HSL H[WUHPDO PRQRfIDFWRUL]DWLRQ RI T EI[%f)[I!fX[$f LV );$ $ & % 7KH SURRI RI /HPPD VKRZV WKDW HDFK H[LVWLQJ D[ LV VSOLW HSL ZKLFK LPSOLHV ERWK D[ DQG )D[f DUH HSL 6LQFH T !D[)[D[f WKH HSL H[WUHPDO PRQRf IDFWRUL]DWLRQ RI T LV DOVR );$ fDfDf $&% 8QLTXHQHVV RI IDFWRUL]DWLRQ LPSOLHV WKHUH LV DQ LVRPRUSKLVP L $ fÂ§! $ FRPSDWLEOH ZLWK WKH IDFWRUL]DWLRQV 6R DAL LD LV D PDS PDNLQJ FRPf KROG IRU $ 5HDGLQJ RII WKH GLDJUDP RQH ILQGV WKDW W!D[)D[f E7Ef8[%f)[W!f E^Y%f^I[%f)[M!f EI[%f)[I!fY[$f SD[LY[$f 7KH ODVW VWHS XVHG WKDW FRPf KROGV IRU $ 6LQFH LV D PRQRPRUSKLVP RQH FRQFOXGHV WKDW H[f LV VDWLVILHG IRU $ D[ LV WKH RQO\ PDS ZKLFK PDNHV WKH LQQHU VTXDUH FRPPXWH EHFDXVH Y[$ LV HSL E\ /HPPD Â’
PAGE 84
&RUROODU\ '3A LV D IXOO VXEFDWHJRU\ RI03P 3URRI %\ /HPPD DQ\ 3APDS SUHVHUYHV WKH )VWUXFWXUH RI D 23A REMHFW /HPPD LPSOLHV WKDW DQ\ 3APDS H[WHQGV WR D PDS Â’ &RUROODU\ 7KH IRUJHWIXO IXQFWRU '3A fÂ§! 3 KDV D OHIW DGMRLQW 3URRI )LUVW QRWH WKDW )$ 2EM'3Af IRU DQ\ $ EHFDXVH I$ )$ fÂ§! 7$ LV DQ HPEHGGLQJ RI )$ LQ D FRPSOHWHO\ GLVWULEXWLYH ODWWLFH %HFDXVH )$ LV WKH IUHH )DOJHEUD RQ $ LI $ fÂ§ % LV DQ RUGHU SUHVHUYLQJ PDS IURP $ 2EM3f WR %H 2EM'3Af & 2EM3)f WKHQ WKHUH LV D '3AXQLTXH PDS Ir )$ fÂ§! % 7KLV SURYHV WKDW UM KDV WKH XQLYHUVDO SURSHUW\ GHVFULEHG LQ WKXV ) LV WKH OHIW DGMRLQW WR WKH IRUJHWIXO IXQFWRU '3A fÂ§}3 Â’ &RUROODU\ 7KH IRUJHWIXO IXQFWRUV '3A fÂ§! 23A DQG '3A fÂ§! KDYH OHIW DGMRLQWV 3URRI $SSO\ 3URSRVLWLRQ DQG &RUROODU\ 1RWH WKH SURRI RI GRHV QRW GHSHQG RQ WKH DUJXPHQWV LQ WKLV VHFWLRQ VR WKH UHVXOW LV OLVWHG KHUH 7KHUH LV QR FLUFXODU DUJXPHQWf Â’ 7KHUH LV VRPH UDWKHU OLPLWHG LQIRUPDWLRQ DERXW PHPEHUV RI 0A 3URSRVLWLRQ ,I $ 2EM0Af WKHQ $ 2EM3Af 3URRI 'HILQH $ 0$ D DQG E\ WKH UHTXLUHPHQW WKDW WKH VTXDUHV EHORZ DUH SXOOEDFNV $ A )$ D 7!$ fÂ§fÂ§f8+$ 0$ )$ I$ 8$ A8'$
PAGE 85
%\ DVVXPSWLRQ M$f DQG UQ$f FRQWDLQ VLQJOHWRQV WKHUHIRUH Â‘L7!$f^$f & 0T R $ & )$ DQG n8$f0$f & 0T R $ & )$ ,W IROORZV WKDW $ & $ L'$fB)$ DQG 0R $ & 0$ OWG$fB)$ $JDLQ XVLQJ WKH IDFW WKDW M DQG P FRQWDLQ VLQJOHWRQV $f$f & $ DQG Â$f$f & 0$ VR FRUHVWULFWLRQV G$? $ fÂ§! $ DQG L$_ $ fÂ§! 0$ VDWLVI\LQJ DG$f? G$ DQG ML$f_ I$ H[LVW ,I D )$ fÂ§! $ LV DQ )DOJHEUD WKHQ DG$?f fÂ§ LGA DQG DL$?f LGA /HPPD VKRZV WKDW IRU DQ\ VXEVHW & '$ FRQWDLQLQJ HDFK SULQFLSDO GRZQVHJPHQW > [ DQ RUGHU SUHVHUYLQJ PDS \ fÂ§rÂ‘ $ VDWLVILHV \G$?f fÂ§ LG LI DQG RQO\ LI \^6f 96 IRU HDFK f DQG DQDORJRXVO\ IRU VXEVHWV RI $f 7KXV ZH FRQFOXGH D ?DQG D ? Â’ 5HPDUN ,I M P X! WKH VXEVHW V\VWHP ZKLFK VHOHFWV DOO ILQLWH VXEVHWV WKHQ 6S)A '3A )RU JHQHUDO M DQG P WKH DXWKRU GRHV QRW NQRZ LI HTXDOLW\ KROGV EXW KH VXVSHFWV WKDW WKH HTXDOLW\ GRHV QRW DOZD\V KROG 5HPDUN $ IXQGDPHQWDO GLIILFXOW\ ZRUNLQJ ZLWK 0AQ LV WKDW RQH GRHV QRW LPPHn GLDWHO\ NQRZ DQ\ RUGHU WKHRUHWLF IRUPXOD IRU WKH )VWUXFWXUH PDSV 2QH PLJKW KRSH WKDW HDFK )VWUXFWXUH PDS LV 6 ?8 ?If6 EXW WKH DXWKRU FDQQRW SUHVHQWO\ VXEVWDQWLDWH VXFK KRSHV
PAGE 86
5HPDUN 7KH DXWKRU NQRZV QR H[DPSOHV RI REMHFWV LQ 2EM0Af ? 2EM'3Af ,Q OLJKW RI &RUROODU\ DQ\ %3A REMHFW LV IUHHO\ HPEHGGHG LQ D FRPSOHWHO\ GLVWULEXWLYH ODWWLFH ,Q IDFW $ f 2EM0Af LV LQ 2EM%3Pf LI DQG RQO\ LI WKH XQLW RI WKH DGMXQFWLRQ PHQWLRQHG LQ LV 3H[WUHPDO PRQR /L>@ H[SOLFLWO\ FRQVWUXFWV D PDS X 3 fÂ§! ,6)3f QRW QHFHVVDULO\ PRQRf ZLWK WKH IROORZLQJ SURSHUWLHV f ,6)3f LV D FRPSOHWH FRPSOHWHO\ GLVWULEXWLYH ODWWLFH f X SUHVHUYHV GHVLJQDWHG PHHWV DQG MRLQV f ,I 3 fÂ§! $ LV D PDS SUHVHUYLQJ GHVLJQDWHG PHHWV DQG MRLQV DQG $ LV FRPSOHWH FRPSOHWHO\ GLVWULEXWLYH WKHQ WKHUH LV D XQLTXH PDS r ,6)3f fÂ§ $ VXFK WKDW IrX 7R H[SODLQ WKH UHODWLRQVKLS EHWZHHQ /LfV UHVXOWV DQG WKH UHVXOWV KHUH QRWH f /L XVHV IDPLOLHV 63 & 63 DQG ,3 & 63 ZKLFK DUH RQO\ UHTXLUHG WR FRQWDLQ VLQJOHWRQV RLQV RI 63VHWV DQG PHHWV R ,3 VHWV DUH UHTXLUHG WR EH SUHVHUYHG ,Q WKLV GRFXPHQW WKH FKRLFH RI GLVWLQJXLVKHG RSWLPXP ERXQGV LV PDGH IRU DOO SRVHWV DW RQFH YLD D VXEVHW V\VWHP f &RPSDULQJ XQLYHUVDO SURSHUWLHV RQH VHHV WKDW /LfV FRQVWUXFWLRQ DSSOLHG WR 3 f 2EM0Af ZLWK 63 M3f DQG ,3 fÂ§ P3f \LHOGV WKH OHIW DGMRLQW PHQWLRQHG LQ 8QIRUWXQDWHO\ LW LV YHU\ GLIILFXOW WR VHH ZKHQ X LV DQ HPEHGGLQJ /HW >@ GHQRWH WKH VPDOOHVW GRZQZDUG FORVHG IDPLO\ RI $ FRQWDLQLQJ 6 ZKLFK LV FORVHG XQGHU 63MRLQV /HW 63 GHQRWH WKH FODVV RI LQFUHDVLQJ VHWV ZKLFK DUH FORVHG XQGHU ,3PHHWV /L GHILQHV ,63f WR EH WKH IDPLOO\ RI DOO GHFUHDVLQJ VHWV LQ 63 RUGHUHG E\ 6 Â‘ 7 LI DQG RQO\ LI WKHUH LV DQ LQGH[HG FKDLQ nÂfÂ ZLWK 4 IO >@ VXFK WKDW 6 6 6c 7 DQG ZKHQHYHU L M 6L IO >3 ? 6M@ 7KH PDS X 3 fÂ§! ,63 LV GHILQHG E\ X[f ^8 8$ 8 [`
PAGE 87
1RQHWKHOHVV %3A LW LV D IDLUO\ QLFH FDWHJRU\ ,Q IDFW LW LV FRPSOHWH 3URSRVLWLRQ %3A LV FRPSOHWH 3URRI 1RWH WKDW D SRVHW $ LV D 3A REMHFW LI DQG RQO\ LI WKHUH LV DQ RUGHU SUHVHUYLQJ PDS D $ [ 0$ fÂ§! $ VXFK WKDW DG$ [ L$f fÂ§ N8 6XSSRVH $LfLfL LV D IDPLO\ RI 3A REMHFWV DQG IRU HDFK L f MfL f A 3L LV DQ HPEHGGLQJ RI $L LQWR D FRPSOHWHO\ GLVWULEXWLYH ODWWLFH 7KH SURRI RI /HPPD LPSOLHV WKDW WKH SRVHW SURGXFW $W LV D 3AREMHFW 0RUHRYHU WKH SURGXFWLQGXFHG PDS QÂH $r r QÂf %L LV DQ HPEHGGLQJ ZKLFK SUHVHUYHV M PfRSWLLQXP ERXQGV (VVHQWLDOO\ ZKDW LV JRLQJ RQ LV WKDW PHHWV DQG MRLQV DUH FRPSXWHG FRRUGLQDWHE\FRRUGLQDWHf ,I J $ fÂ§! % D[H 3Â PDSV WKHQ WKH SURRI RI /HPPD LPSOLHV HTIJf LV D M PfFRPSOHWH VXEVHW RI $ ,I $ f 2EM%3Af WKHQ $ FDQ EH MPfHPEHGGHG LQWR D FRPSOHWHO\ GLVWULEXWLYH ODWWLFH VR HT Jf PD\ DOVR EH VR HPEHGGHG 7KXV DQ\ VHW RI REMHFWV LQ %3A KDV D SURGXFW DQG DQ\ SDLU RI %3A PDSV KDV DQ HTXDOL]HU %\ %RUFHX[ > 9ROXPH @ WKLV LPSOLHV %3ID LV FRPSOHWH Â’
PAGE 88
&+$37(5 &2(48$/,=(56 (SLV DQG (TXDOL]HUV LQ 3 7KH UHVXOWV RI WKLV VHFWLRQ FKDUDFWHUL]H HSLV DQG UHJXODU HSLV LQ WKH FDWHJRU\ 3 RI SRVHWV 7KH SUHVHQWDWLRQ DQG SURRIV H[FHSW &RQVWUXFWLRQ ZKLFK LV GLVFXVVHG LQ 0HVHJXHU >@f DUH WKH ZRUN RI WKH DXWKRU EXW WKH DXWKRU EHOLHYHV LW OLNHO\ WKDW WKH\ DUH QRW QHZ ,Q DOO VWDWHPHQWV $ DQG % DUH DUELWUDU\ SRVHWV /HPPD /HW I $ fÂ§! % EH D PRQRWRQH PDS I LV HSL LI DQG RQO\ LI I LV RQWR 3URRI 6LQFH WKH IRUJHWIXO IXQFWRU 3 fÂ§! 6HW LV IDLWKIXO LI LV RQWR WKHQ LV HSL )RU WKH FRQYHUVH VXSSRVH E % ? I$f 'HILQH 6L L I$fQ Ef DQG 6 E (YLGHQWO\ ERWK 6? DQG 6 KDYH WKH VDPH LQWHUVHFWLRQ ZLWK I$f EXW E 6 ? 6L 7KHUHIRUH FKDUDFWHULVWLF IXQFWLRQV RI % ? 6L DQG %? 6 DUH GLVWLQFW PDSV VD\ FLF % fÂ§r DUH UHVSHFWLYHO\ WKH FKDUDFWHULVWLF IXQFWLRQV RI % ? 6L DQG % ? 6f VXFK WKDW &LI FI %RWK 4 DUH RUGHU SUHVHUYLQJ EHFDXVH 6L DQG 6 DUH GHFUHDVLQJ VHWV Â’ 'HILQLWLRQ DQG 5HPDUNV 7KH IROORZLQJ FRQVWUXFWLRQ LV SDUDSKUDVHG IROORZLQJ 0HVHn JXHU >@ SS ,I $ fÂ§! % LV DQ\ RUGHU SUHVHUYLQJ PDS ZH PD\ GHILQH D SUHRUGHU ; RQ E\ DL D D[f Df
PAGE 89
7KLV UHODWLRQ LV REYLRXVO\ UHIOH[LYH DQG WUDQVLWLYH EXW EHFDXVH PD\ QRW EH LQMHFWLYH PD\ QRW EH DQWLV\PPHWULF 6XSSRVH $ f LV D SRVHW DQG A LV D SUHRUGHU VWUHQJWKHQLQJ LH L D DL Â‘ 7KHQ ZH KDYH DQ RUGHU SUHVHUYLQJ PDS D $ fÂ§! $ Â‘ 5HFDOO WKDW $ Â‘ LV WKH VHW RI HTXLYDOHQFH FODVVHV ^[ f $ [ D DQG D Â‘ [` SDUWLDOO\ RUGHUHG E\ f 'HILQH PDSV K $ %W WR EH HTXLYDOHQW LI WKHUH LV DQ LVRPRUSKLVP L %[ fÂ§! % VXFK WKDW % FRPPXWHV 2QH YHULILHV WKDW 7KH PDSV I $ r %f DQG f! D ? $ $ DUH PXWXDOO\ LQYHUVH FRUUHVSRQGHQFHV EHWZHHQ f WKH VHW PRGXOR HTXLYDOHQFHf RI VXUMHFWLYH PDSV ZLWK GRPDLQ $ f WKH VHW RI SUHRUGHUV RQ $ ZKLFK VWUHQJWKHQ )RU DQ\ $fÂ§r % I FD ZKHUH D $ fÂ§! $ A DQG FDDff Df *LYHQ VXUMHFWLRQV ILKn$fÂ§r %Â WKHUH LV D F %? fÂ§! % LI DQG RQO\ LI L K DO A D DO A D
PAGE 90
5HPDUN 1RWH WKDW DQ\ LQWHUVHFWLRQ RI SUHRUGHUV LV D SUHRUGHU 7KXV WKH FODVV RI SUHRUGHUV VWUHQJWKHQLQJ LV D FRPSOHWH ODWWLFH :H VD\ D VHW 6 RI RUGHUHG SDLUV JHQHUDWHV Â‘ LI Â‘ LV WKH VPDOOHVW SUHRUGHUHG FRQWDLQLQJ 6 1RWH WKDW LI JHQHUDWHV Â‘ ZH FDQ GHVFULEH H[SOLFLWO\ D Â‘ DQ LI f FOT GP RU f WKHUH LV D ILQLWH VHTXHQFH DLD DfBM VXFK WKDW IRU DOO L ZLWK L Q fÂ§ pÂf pWLf e 6 7KH ILUVW EXOOHWHG FRQGLWLRQ LQVXUHV WKDW LV UHIOH[LYH 7KH VHFRQG UHTXLUHPHQW LQVXUHV WKDW Â‘ LV WUDQVLWLYH JHQHUDOO\ LI 6 LV D UHODWLRQ WKH UHODWLRQ REWDLQHG E\ DSSO\LQJ WKH VHFRQG EXOOHWHG LWHP RQO\ LV FDOOHG WKH WUDQVLWLYH FORVXUH RI 6f 2QH PD\ YHULI\ WKH IROORZLQJ FRQVWUXFWLRQ XVLQJ &RQVWUXFWLRQ /HW IJ $ fÂ§! % EH RUGHU SUHVHUYLQJ PDSV 7KH FRHTXDOL]HU RI I DQG J LV WKH TXRWLHQW RI % f E\ WKH VPDOOHVW SUHRUGHU FRQWDLQLQJ DQG &J Kf fÂ§ ^J[K[fK[J[f [ $f /HPPD /HW I $ fÂ§! % EH D VXUMHFWLYH RUGHU SUHVHUYLQJ PDS 7KH IROORZLQJ DUH HTXLYDOHQW I LV WKH FRHTXDOL]HU RI VRPH SDLU J K $R K $ 7KHUH H[LVW D SRVHW $R DQG PDSV J K $R fÂ§! $ VXFK WKDW WKH SUHRUGHU Â‘ RQ $ f LV JHQHUDWHG E\ DQG &JKf 7KH SUHRUGHU A RQ $ f LV JHQHUDWHG E\ DQG .If fÂ§ ^[\f ( %[% I[f I\f` I 7KH SUHRUGHU Â‘ RQ $ f LV JHQHUDWHG E\ DQG VRPH HTXLYDOHQFH UHODWLRQ 3URRI f LV HYLGHQW IURP f LV WULYLDO EHFDXVH .^I f LV DQ HTXLYDOHQFH UHODWLRQ :H KDYH &JKf&.If F
PAGE 91
EHFDXVH IJ IK DQG [\f H.If A I[f I^\f DQG I^\f I[f I I 6LQFH &JKf DQG D JHQHUDWH ; .If DQG JHQHUDWH ; )RU f OHW ; EH JHQHUDWHG E\ DQG WKH HTXLYDOHQFH UHODWLRQ ( & $ [ $ ZLWK WKH WULYLDO RUGHU 7KHQ WKH SURMHFWLRQ PDSV WW? 7 ( fÂ§! $ JLYHQ E\ 79L[L;f ;L DUH RUGHU SUHVHUYLQJ 7KH FRHTXDOL]HU RI 7L DQG U LV Â’ ([DPSOH 6RPH UHDGHUV PD\ EH VXUSULVHG E\ WKH IDFW WKDW VXUMHFWLRQV DUH QRW DOZD\V 3UHJXODU HSL $Q H[DPSOH RI WKLV VLWXDWLRQ LV DQ\ PDS ÂD fÂ§} IURP WKH WULYLDOO\ RUGHUHGf WZR SRLQW VHW RQWR D WZRSRLQW FKDLQ $Q\ VXFK PDS IDLOV FRQGLWLRQ IURP /HPPD )DFWRUL]DWLRQ RI 0DSV 8VLQJ 3UHRUGHUV 7KLV VHFWLRQ PRGLILHV WKH SUHRUGHU IDFWRUL]DWLRQ WR JLYH D ILUVW DSSUR[LPDWLRQ WR FRHTXDOL]HUV LQ )P 'HILQLWLRQ /HW = EH D VXEVHW V\VWHP = LV VDLG WR DGPLW FRQJUXHQFHV LI IRU DOO SRVHWV $ =$ [ $f ^6 f $ [ $ f =$fL ` /HPPD 6XSSRVH ERWK M DQG P DUH VXEVHW V\VWHPV ZKLFK DGPLW FRQJUXHQFHV /HW $ f EH D SRVHW DQG ; EH D SUHRUGHU VWUHQJWKHQLQJ 7KHQ D $ fÂ§r $ ; SUHVHUYHV M PfRSWLPXP ERXQGV LI DQG RQO\ LI ; LV D )PVXEREMHFW RI $[ $ 3URRI :H VKRZ WKDW T9[Âf 9DUÂf IRU DQ\ [Â L ` f M^$f &OHDUO\ D9[f LV DQ XSSHU ERXQG RI D[Âf L f ,` 6LQFH ; LV D 3AVXEREMHFW RI $ [ $ ^[c L f ` M$f DQG IRU DOO L ;L D 9^[M Df L ( ` 9[LDf
PAGE 92
LV D PHPEHU RI Gn 9[Â D 7KXV DQ\ XSSHU ERXQG RI ^D[Âf L e ` GRPLQDWHV D9[Âf 7KH SURRI IRU PHHWV LV LGHQWLFDO VR RPLWWHGf &RQYHUVHO\ LI D SUHVHUYHV MMRLQV DQG PPHHWV DQG IRU H[DPSOHf ^[LnL f ` ^\W L e ` e M^$f DQG IRU DOO L e [Â G WKHQ 9Â[ G 6R A LV D 3AVXEREMHFW RI [ $ Â’ 'HILQLWLRQ DQG 5HPDUNV 6XSSRVH M DQG P DGPLW FRQJUXHQFHV /HW $ f e 2EM3Af ,I A LV D SUHRUGHU VWUHQJWKHQLQJ DQG VLPXOWDQHRXVO\ D 3AVXEREMHFW RI $ [ $ WKHQ ZH VD\ Â‘ LV D M PfSRFRQJUXHQFH 3RFRQJUXHQFHV RI 3AREMHFWV EHKDYH YHU\ PXFK OLNH SUHRUGHUV RI SRVHWV 7KH PDSV I $ fÂ§r %f !! DQG Af D $ fÂ§! $ Â‘ DUH PXWXDOO\ LQYHUVH FRUUHVSRQGHQFHV EHWZHHQ f WKH VHW PRGXOR HTXLYDOHQFHf RI VXUMHFWLYH PDSV ZKLFK SUHVHUYH M PfRSWLPXP ERXQGV DQG KDYH GRPDLQ $ f WKH VHW RI M PfSRFRQJUXHQFHV RQ $ )RU DQ\ $ fÂ§! % I FD ZKHUH D $ fÂ§! $ Â‘ DQG FDDff Df ,I SUHVHUYHV M PfRSWLPXP ERXQGV VR GR F DQG D *LYHQ MPfRSWLPXP ERXQG SUHVHUYLQJ VXUMHFWLRQV ILIn$fÂ§!%O WKHUH LV F %? fÂ§ % LI DQG RQO\ LI $ 6L DL GL D mL GL D &RQVLGHUDWLRQ RI /HPPD VKRZV WKDW LI $ fÂ§! % LV VXUMHFWLYH DQG 6 e M$f UHVS 6 e UQ$ff WKHQ I6f KDV D VXSUHPXP UHVS LQILPXPf LQ % 0RUHRYHU VHWV RI WKH IRUP I6f IRU 6 e M$f UHVS 6 e P$ff DUH WKH RQO\ VHWV ZKRVH RSWLPXP ERXQGV DUH JXDUn DQWHHG WR H[LVW 7KXV ZH DUH PRWLYDWHG WR RIIHU WKH IROORZLQJ FRQGLWLRQV RQ VXEVHW V\VWHPV WR JXDUDQWHH VXUMHFWLYH UHVS 3UHJXODU HSLf LPDJHV RI 3Â REMHFWV DUH M PfFRPSOHWH
PAGE 93
'HILQLWLRQ DQG 5HPDUNV /HW = EH D VXEVHW V\VWHP :H VD\ = SUHVHUYHV VXUMHFWLRQV LI ZKHQHYHU $ fÂ§r % LV D SRVHW VXUMHFWLRQ =%f ^I6f6H=$f` 6LPLODUO\ ZH VD\ = SUHVHUYHV UHJXODU HSLV LI ZKHQHYHU $ fÂ§r % LV D SRVHW UHJXODU HSLPRUSKLVP =%f ^I6f f =$f` 6LQFH SRVHW UHJXODU HSLPRUSKLPV DUH VXUMHFWLYH DQ\ = WKDW SUHVHUYHV VXUMHFWLRQV DOVR SUHVHUYHV UHJXODU HSLPRUSKLVPV $Q\ FDUGLQDO N WKH DVVRFLDWHG VXEVHW V\VWHP N SUHVHUYHV VXUMHFWLRQV )RU LI I $ fÂ§! % LV VXUMHFWLYH WKHQ LV VSOLW HSL LQ 6HWf +HQFH WKHUH LV V % fÂ§! $ QRW QHFHVVDULO\ RUGHU SUHVHUYLQJf VXFK WKDW LGJ IV 7KHUHIRUH DQ\ VHW 6 & % ZLWK FDUGLQDOLW\ OHVV WKDQ Q LV WKH LPDJH RI VRPH VHW V6f & $ 7KH IROORZLQJ UHPDUN SURYLGHV DQ H[DPSOH RI D VXEVHW V\VWHP WKDW GRHV QRW SUHVHUYH VXUMHFWLRQV RU UHJXODU HSLV 7KH DXWKRU FXUUHQWO\ GRHV QRW NQRZ RI DQ\ VXEVHW V\VWHP WKDW SUHVHUYHV UHJXODU HSLV ZLWKRXW SUHVHUYLQJ VXUMHFWLRQV EXW LW VHHPV OLNHO\ VXFK D VXEVHW V\VWHP H[LVWV $ PRUH LQWHUHVWLQJ TXHVWLRQ ZKLFK WKH DXWKRU DOVR FDQQRW FXUUHQWO\ DQVZHU LV fGR WKHUH H[LVW VXEVHW V\VWHPV RWKHU WKDQ FDUGLQDOV ZKLFK SUHVHUYH UHJXODU HSLPRUSKLVPV"f 5HPDUN /HW M GLU EH WKH VXEVHW V\VWHP ZKLFK VHOHFWV DOO XSZDUG GLUHFWHG VXEVHWV RI D SRVHW /HW 1 GHQRWH D FRXQWDEOH GLVMRLQW XQLRQ RI WZRSRLQW FKDLQV 6HH GUDZLQJ EHORZ IRU KHOS YLVXDOL]LQJ DQG WR IL[ QRWDWLRQf WR W? K ER (YLGHQWO\ 1 LV MMRLQ FRPSOHWH
PAGE 94
&RQVLGHU WKH IXQFWLRQ 1 fÂ§} 1 ZKHUH 1 ^ fff` ZLWK WKH XVXDO RUGHU GHILQHG E\ IFÂf L IRU L f DQG ÂBLf L IRU L f 6SHDNLQJ URXJKO\ fI VWDFNV WKH WZR SRLQW FKDLQVf LV UHJXODU HSL DV D PDS RI SRVHWV 0RUHRYHU A LV D MVXEREMHFW RI 1 [ 1 EHFDXVH SUHVHUYHV DOO H[LVWLQJ MRLQV %XW 1 LV QRW MMRLQ FRPSOHWH EHFDXVH 1 LV XSZDUG GLUHFWHG EXW KDV QR MRLQ %HFDXVH RI WKH SRFRQJUXHQFH IDFWRUL]DWLRQ IRU M PfERXQG SUHVHUYLQJ PDSV RXWOLQHG LQ f ZH KDYH WKH IROORZLQJ ILUVW DSSUR[LPDWLRQ WR WKH FRHTXDOL]HU LQ 3A 7KH RQO\ WKLQJ VWRSSLQJ PDS D % fÂ§! % Â‘ GHVFULEHG EHORZf IURP DFWXDOO\ EHLQJ D FRHTXDOL]HU LQ 3A LV WKDW % Â‘ LV JHQHUDOO\ QRW ^M PfFRPSOHWH &RQVWUXFWLRQ 6XSSRVH IJ$fÂ§! % f DUH gP PDSV DQG &IJf LV GHILQHG DV LQ ,Â‘ /HW ; EH WKH VPDOOHVW SRFRQJUXHQFH FRQWDLQLQJ DQG & Jf 7KHQ D % fÂ§! % ; KDV WKH IROORZLQJ XQLYHUVDO SURSHUW\ LI K % fÂ§! & SUHVHUYHV M PfRSWLPXP ERXQGV DQG KI KJ WKHQ WKHUH LV D XQLTXH PDS L % fÂ§! & VXFK WKDW K fÂ§ LD 6LQFH WKH TXRWLHQW PDS D % fÂ§! % A LV D VXUMHFWLRQ ZH KDYH &RUROODU\ &RQWLQXH ZLWK WKH QRWDWLRQ IURP &RQVWUXFWLRQ ,I M DQG P SUHVHUYH VXUMHFWLRQV WKHQ % LV MPfFRPSOHWH 7KXV D % fÂ§ % LV WKH FRHTXDOL]HU RI I DQG J LQ )P ,Q SDUWLFXODU LI M N DQG P $ DUH FDUGLQDOLW\ VXEVHW V\VWHPV WKHQ D LV WKH FRHTXDOL]HU &RUROODU\ '3f KDV FRHTXDOL]HUV 3URRI 7KH SUHFHGLQJ VKRZV WKDW DQ\ SDLU IJ$fÂ§!%H 3A KDV D FRHTXDOL]HU 6XSSRVH % LV FRPSOHWHO\ GLVWULEXWLYH 6LQFH D % fÂ§! FRHT3RR Jf LV D VXUMHFWLRQ ZKLFK SUHVHUYHV DOO PHHWV DQG MRLQV FRHTSr Jf LV FRPSOHWHO\ GLVWULEXWLYH ,Q SDUWLFXODU LI ERWK $ DQG % DUH FRPSOHWHO\ GLVWULEXWLYH WKHQ FRHTSr1IJf LV DOVR WKH FRHTXDOL]HU LQ '3f Â’ )DFWRUL]DWLRQ RI 0HHWVHPLODWWLFH PDSV 6HYHUDO VLPSOLILFDWLRQV RFFXU GHVFULELQJ FRHTXDOL]HUV LQ 3Â LI WKH REMHFWV KDYH D PHHWn VHPLODWWLFH VWUXFWXUH LH P! XM
PAGE 95
/HPPD /HW I $ fÂ§! % EH D PHHWVHPLODWWLFH PDS I LV )UHJXODU HSL LI DQG RQO\ LI I LV VXUMHFWLYH 3URRI 6LQFH 3UHJXODU HSLPRUSKLVPV DUH DOZD\V VXUMHFWLYH RQH LPSOLFDWLRQ LV WULYLDO )RU WKH RWKHU VXSSRVH $ fÂ§r % LV D VXUMHFWLRQ SUHVHUYLQJ ELQDU\ PHHWV 7KHQ [ N 9 I[$\f I[f 7KXV ZH KDYH WKH VHTXHQFH [[ $ \ \ ZKHUH I[f I[ $ \f DQG [ $\ \ %\ /HPPD WKLV SURYHV LV UHJXODU HSL Â’ $V QRWHG LQ WKH SURRI WKH UHODWLRQ Â‘ LV FRPSOHWHO\ GHVFULEHG E\ .If ,I REMHFWV KDYH D PHHWVHPLODWWLFH VWUXFWXUH RQH PD\ XVH FRQJUXHQFHV HTXLYDOHQFH UHODWLRQV WKDW DUH VLPXOWDQHRXVO\ M PfVXEDOJHEUDVf UDWKHU WKDQ WKH PRUH FRPSOH[ SRFRQJUXHQFHV )LQDOO\ ZH KDYH 7KHRUHP 6XSSRVH P WR DQG M DQG P SUHVHUYH UHJXODU HSLV /HW I $ fÂ§! % EH D 3@P PDS 7KH IROORZLQJ DUH HTXLYDOHQW I LV VXUMHFWLYH I LV D TXRWLHQW E\ VRPH FRQJUXHQFH I LV 3 UHJXODU HSL I 3AUHJXODU HSL I LV )IH[WUHPDO HSL 3URRI 7KH EDVLF IDFWV DERXW SRFRQJUXHQFHV VKRZ f 7KH SUHFHGLQJ OHPPD VKRZV f f KROGV LQ DQ\ FDWHJRU\ 7R VKRZ f DVVXPH $ fÂ§! $. 6LQFH & $ [ $ ZH KDYH SURMHFWLRQ PDSV LU L W fÂ§! $ I LV SODLQO\ WKH FRHTXDOL]HU RI ULU
PAGE 96
! f KROGV EHFDXVH ZH PD\ IDFWRU DQ\ PDS WKURXJK LWV DVVRFLDWHG SRFRJUXHQFH ,I LV H[WUHPDO HSL WKHQ PXVW FRLQFLGH ZLWK WKH TXRWLHQW E\ LWV DVVRFLDWHG SRFRQJUXHQFH Â’ &RUROODU\ /HW M DQG P EH VXEVHW V\VWHPV ZKLFK SUHVHUYH UHJXODU HSLV /HW P Z 7KH IRUJHWIXO IXQFWRUV 8S )P fÂ§! 3 DQG 86HW 3Â an 6HW SUHVHUYH DQG UHIOHFW UHJXODU HSLV /HPPD 7KH IRUJHWIXO IXQFWRU 8PA '3A fÂ§! 6HW UHIOHFWV NHUQHO SDLUV 3URRI 2QH YHULILHV WKDW WKH NHUQHO SDLU RI D PDS $ fÂ§r % FDOFXODWHG LQ '3ÂFRQVLVWV RI WKH SURMHFWLRQV IURP $[ $f UHVWULFWHG WR WKH VHW ^[\f f $ [ $ I[f I\f` Â’ $SSO\LQJ 7KHRUHP RQH REWDLQV WKH IROORZLQJ &RUROODU\ /HW M DQG P EH VXEVHW V\VWHPV ZKLFK SUHVHUYH UHJXODU HSLV /HW P XM 7KH IRUJHWIXO IXQFWRU 8VHW nÂ‘ 03Â a 6HW LV PRQDGLF &RHTXDOL]HUV LQ '3Â 7KHRUHP )RU DQ\ SDLU JK $ fÂ§! % f ,03A WKH FRHTXDOL]HU FRHT" Kf H[LVWV 3URRI 7KH FRQVWUXFWLRQ RI WKH FRHTXDOL]HU RI PDSV J DQG K LV LOOXVWUDWHG E\ WKH IROORZLQJ GLDJUDP 7KH PDSV RQ WKH GLDJUDP ZLOO EH GHILQHG EHORZ a'^&f
PAGE 97
/HW D % fÂ§! % EH WKH PDS JLYHQ E\ &RQVWUXFWLRQ /HW % fÂ§! & EH DQ\ %3A PDS VXFK WKDW IJ IK 7KH H[LVWHQFH DQG XQLTXHQHVV RI LV JLYHQ E\ 7R FRQVWUXFW WKH OHIWPRVW VTXDUH ZH DSSO\ WKH IUHH FRPSOHWHO\ GLVWULEXWLYH FRPSOHWH ODWWLFH IXQFWRU O' Uc LV WKH QDWXUDO WUDQVIRUPDWLRQ ZKLFK LQMHFWV D SRVHW + LQWR '+ 7R FRQVWUXFW WKH ULJKWRXWVLGH VTXDUH ZH QRWH WKDW RR RRfSRFRQJUXHQFHV DUH FORVHG XQGHU LQWHUVHFWLRQ 6R WKHUH LV D VPDOOHVW LQ WKH VHQVH WKDW LW PDNHV WKH IHZHVW SRVVLEOH LGHQWLILFDWLRQVf RR RRfTXRWLHQW N '% Â‘f fÂ§! (% ;f VXFK WKDW NU@% ;ffD SUHVHUYHV DOO M PfRSWLPXP ERXQGV 6LPLODUO\ GHILQH e '&f (&f WR EH WKH VPDOOHVW TXRWLHQW VXFK WKDW eUc&fI SUHVHUYHV DOO M PfERXQGV 6LQFH e'f9% OffD e9&fID eUM&fI SUHVHUYHV DOO M PfERXQGV ZH KDYH WKH LQGXFHG PDS (If 'HILQH (R WR EH WKH VPDOOHVW %3AVXEREMHFW RI (% Â‘f WKURXJK ZKLFK NUM% Aff IDFWRUV %\ 3URSRVLWLRQ WKH LQWHUVHFWLRQ GHILQLQJ (R H[LVWVf 7KH PDSV M % fÂ§! (T DQG ( fÂ§! (% ;f DUH REWDLQHG E\ IDFWRULQJ Uc% ffDf WKURXJK ( %\ FRQVWUXFWLRQ WKH PDS eUc&f & fÂ§! (&f KDV WKH XQLYHUVDO SURSHUW\ WKDW DQ\ PDS M PfRSWLPXP ERXQG SUHVHUYLQJ I! & fÂ§r % ZLWK % f 2EM%3Af IDFWRUV DV If WeUc&f IRU D XQLTXHO\ GHWHUPLQHG PDS W 6LQFH & f 2EM%3AU>f WKHUH LV VXFK D Mf ZKLFK LV 3H[WUHPDO PRQR 7KXV eU@&f LV 3H[WUHPDO PRQR 7KHUHIRUH WKH 3HSL H[WUHPDO PRQRf IDFWRUL]DWLRQ RI eUf&f SURGXFHV WKH IDFWRUL]DWLRQ eMJ&f TP %\ FRQVWUXFWLRQ P LV ERWK 3HSL DQG 3H[WUHPDO PRQR VR P LV D 3LVRPRUSKLVP 0RUHRYHU T & fÂ§! (&f LV QHFHVVDULO\ WKH VPDOOHVW %3AVXEREMHFW RI (&nf WKURXJK ZKLFK e^J&f IDFWRUV :H FODLP WKDW FRHTJKf (R YLD M % fÂ§! (R ,W VXIILFHV WR VKRZ WKH H[LVWHQFH RI D XQLTXH FRPSDWLEOH PDS r %R fÂ§} & 7KH FRPPXWDWLYLW\ RI WKH ODUJHVW UHFWDQJOH
PAGE 98
LQ WKH GLDJUDP DERYH LPSOLHV WKDW NUM% ff IDFWRUV XQLTXHO\ WKURXJK WKH .03AREMHFW (Ifa&f & (%f %HFDXVH ( LV WKH VPDOOHVW %3AREMHFW WKURXJK ZKLFK NU@% ff IDFWRUV (T & (IfaO&f 7KLV LQVXUHV WKH H[LVWHQFH DQG XQLTXHQHVV RI r Â’
PAGE 99
&+$37(5 MPf63$&(6 7KLV FKDSWHU VWXGLHV VSDFHV REH\LQJ D 7RVW\OH VHSDUDWLRQ D[LRP 6HFWLRQ GHILQHV VSDFHV DQG GHVFULEHV D IXQFWRULDO *DORLV FRQQHFWLRQ ZKLFK VSHFLDOL]HV WR *DORLV FRQn QHFWLRQV EHWZHHQ '3A DQG 6P 6HFWLRQ GHYHORSV D FRQYHQLHQW GHVFULSWLRQ RI HSLPRUSKLVPV LQ 6P ZKLFK JHQHUDOL]HV D NQRZQ FKDUDFWHUL]DWLRQ RI HSLPRUSKLVPV RI 7RVSDFHV 6HFWLRQ JLYHV FRQVWUXFWLRQV RI OLPLWV VLPLODU WR WKRVH IRU WRSRORJLFDO VSDFHV 6HFWLRQ GHVFULEHV TXRWLHQW PDSV DQG FKDUDFWHUL]HV H[WUHPDO DQG UHJXODU HSLV DV TXRWLHQW PDSV 7KH ODVW WZR VHFWLRQV DUH UHODWHG WR WKH SUREOHP RI ILQGLQJ UHIOHFWLRQV LQ 6P@ 6HFWLRQ JLYHV WKH IODW VSHFWUXP FRfUHIOHFWLRQ RQ VSDWLDO REMHFWV ZKLFK LV HTXLYDOHQW WR D UHIOHFWLRQ RQ 6ÂM 7KLV UHIOHFWLRQ RQ VSDFHV LV D JHQHUDOL]DWLRQ RI WKH 7LUHIOHFWLRQ RI WRSRORJLFDO VSDFHV /DVW EXW QRW OHDVW 6HFWLRQ SDUWLDOO\ GHVFULEHV WKH HSLFRPSOHWH 6UQ REMHFWV WKH GHVFULSWLRQ LV FRPSOHWH IRU 7VSDFHV (SLFRPSOHWH 7RVSDFHV DUH FKDLQV ZLWK WKH VSHFLDOL]Dn WLRQ RUGHU 3URGXFWV RI HSLFRPSOHWH 7VSDFHV DUH QRW HSLFRPSOHWH VR WKHUH LV QR IXQFWRULDO HSLFRPSOHWLRQ LQ WKH FDWHJRU\ RI 7RVSDFHV $V PHQWLRQHG LQ WKH LQWURGXFWRU\ FKDSWHU WKH UHVHDUFK OHDGLQJ WR WKLV GLVVHUWDWLRQ EHJDQ LQ DQ DWWHPSW WR ILQG UHIOHFWLRQV DQG FRUHIOHFWLRQV LQ FDWHJRULHV RI JHQHUDOL]HG WRSRn ORJLFDO VSDFHV $IWHU SURYLQJ WKH UHVXOWV RI WKLV FKDSWHU DQG UHDGLQJ 0HVHJXHU >@ WKH DXWKRU UHDOL]HG WKDW DGGLWLRQDO DVVXPSWLRQV ZKHUH UHTXLUHG RQ VXEVHW V\VWHPV WR LQVXUH WKDW VXEVSDFHV FRXOG EH UHDVRQDEO\ GHILQHG 7KLV UHDOL]DWLRQ SURPSWHG PXFK RI WKH WKRXJKW VXPn PDUL]HG LQ &KDSWHU LQ SDUWLFXODU 6HFWLRQ WKH DXWKRU ZDQWHG WR ILQG ZKHQ WKH WKHRU\ RI WKLV FKDSWHU ZDV YDOLG 7KHVH FRQVLGHUDWLRQV DQG FRQVWUXFWLRQ RI IUHH '3AREMHFWV EHn FDPH WKH PDLQ IRFXV RI WKH GLVVHUWDWLRQ +RZHYHU WKLV VWDWH RI DIIDLUV OHDYHV PDQ\ TXHVWLRQV FRQFHUQLQJ UHIOHFWLRQV DQG FRUHIOHFWLRQV LQ 6P XQWRXFKHG
PAGE 100
7KLV FKDSWHU DVVXPHV VOLJKWO\ PRUH EDFNJURXQG WKHQ WKH UHVW RI WKH WH[W &ORVXUH RSHUDWRUV DUH XVHG ZLWKRXW FRPPHQW ,I F LV D FORVXUH RSHUDWRU IL[Ff ^[ F[f [` 6SDWLDO6REHU )XQFWRULDO *DORLV &RQQHFWLRQ 'HILQLWLRQ DQG 5HPDUNV /HW $3RV GHQRWH WKH FDWHJRU\ RI ZKRVH REMHFWV $ FRQVLVW RI RELf DQ XQGHUO\LQJ SDUWLDOO\ RUGHUHG VHW GHQRWHG $ REf D IDPLO\ RI GHVLJQDWHG VXEVHWV $ VXFK WKDW ?H[LVWV IRU DOO 6H$ REf D IDPLO\ RI GHVLJQDWHG VXEVHWV 0$ VXFK WKDW I? 6 H[LVWV IRU DOO 6f0$ ,WHPV REf DQG REf RI WKH GDWD GHILQLQJ D PHPEHU RI 2EM $3RVf ZLOO EH UHIHUUHG WR DV WKH VLJQDWXUH RI $ 1RWH WKDW D JLYHQ SRVHW $ PD\ KDYH VHYHUDO SRVVLEOH VLJQDWXUHV $ $3RVPDS nÂ‘ $ fÂ§! % LV D IXQFWLRQ $ fÂ§r % VXFK WKDW PDS2f FMf LV PRQRWRQH LH D E H $ !^Df I!Ef PDSOf !>$@ ^!f $` & 3 DQG 9 f $ FI!^?6f ?I!^6f PDSf f>0$? & 0% DQG 9 0$ $ 6f $ 7KH FDWHJRU\ $6 KDV REMHFWV $f';f 6;f $;ff ZKHUH ; LV D VHW e!;f LV D IDPLO\ RI VXEVHWV RI ; DQG 6$f $;f DUH IDPLOLHV RI VXEVHWV RI 2;f VXFK WKDW '$f ($7f $;ff LV D $3RVREMHFW LQ ZKLFK WKH RSWLPXP ERXQG RSHUDWLRQV PHHW DQG MRLQ DUH WKH VHW WKHn RUHWLF RSHUDWLRQV RI LQWHUVHFWLRQ DQG XQLRQ $6 PDSV DUH IXQFWLRQV I ; fÂ§r < VXFK WKDW
PAGE 101
' $6 fÂ§A $3RV LV WKH FRQWUDYDULDQW IXQFWRU ZKLFK VHQGV D VSDFH ; WR WKH $3RVREMHFW 2;f ($f $$ff DQG D PDS WR WKH DVVRFLDWHG LQYHUVH LPDJH PDS ZKLFK VHQGV 8 f '\f WR Ia?8f e ';f $ SULPH ILOWHU RQ $ f 2EM $3RVf LV D VHW 3 & $ VXFK WKDW S)LOOf 3 LV LQFUHDVLQJ WKDW LV LI D e 3 DQG D E WKHQ E e 3 S)LOf 3 LV LQDFFHVVLEOH WKDW LV LI 6 e $ DQG ?6 e 3 WKHQ 6I@3 LV QRQHPSW\ S)LOf 3 LV 0FORVHG WKDW LV LI 0$ DQG 6 & 3 WKHQ I?6 e 3 )RU $ e 2EM $3RVf WKHUH LV D QDWXUDO ELMHFWLRQ EHWZHHQ SULPH ILOWHUV SULPH LGHDOV GHFUHDVn LQJ FORVHG DQG 0LQDFFHVVLEOH VXEVHWV RI $f DQG FKDUDFWHUV $ fÂ§r e $6 LI [ $ fÂ§! LV D PDS WKHQ [Bf LV D SULPH LGHDO DQG [BOf LV D SULPH ILOWHU [ FDQ EH UHFRYHUHG HLWKHU IURP [f RU [Of )RU D e $ OHW FR]Df GHQRWH WKH FROOHFWLRQ RI DOO SULPH ILOWHUV FRQWDLQLQJ D /HW n,n$ GHQRWH WKH FROOHFWLRQ RI DOO SULPH ILOWHUV RI $ n/$ LV D $6REMHFW LI ZH GHILQH 'A$ff FR]>$@ ^FR]Df D e $` ($f FR]>>$@@ ^^FR]Df D e 6` 6 e $` DQG $$f FR]>S9W$@@ %\ GHILQLWLRQ RI PDSV LQ $6 n/ H[WHQGV WR D FRQWUDYDULDQW IXQFWRU ,Q %DQDVFKHZVNL DQG %UXQV >@ LW LV YHULILHG WKDW n/ DQG DUH DGMRLQW RQ WKH ULJKW ZKHUH WKH XQLW QDWXUDO WUDQVIRUPDWLRQV KDYH FRPSRQHQWV W@D^Rf FR] Df H[[f ^8 e 2;f [ e 8`
PAGE 102
,Q YHULI\LQJ WKH DGMXQFWLRQ RQH VKRZV WKDW FR] Yr+^ FR]Df D VM IRU DQ\ 6 $f DQG FR] $}fQ^ FR]Df D IRU DQ\ 6 0$f :H DOVR IROORZ %DQDVFKHZVNL DQG %UXQV >@ LQ GHILQLQJ 6S$3RV WR EH WKH IXOO VXEFDWHn JRU\ RI $3RV FRQWDLQLQJ DOO $ VXFK WKDW Uc? LV DQ LVRPRUSKLVP DQG 6R $6 WR EH WKH IXOO VXEFDWHJRU\ RI $6 FRQWDLQLQJ ; DOO ; VXFK WKDW H[ LV DQ LVRPRUSKLVP 7KHVH FDWHJRULHV DUH VDLG WR FRQVLVW RI VSDWLDO SRVHWV DQG VREHU VSDFHV UHVSHFWLYHO\f %DQDVFKHZVNL DQG %UXQV >@ VKRZ WKDW 6S$3RV LV RQWRUHIOHFWLYH LQ $3RV DQG WKDW 6R $6 LV UHIOHFWLYH LQ $6 ,Q RUGHU WR DSSO\ WKH UHVXOWV RI WKLV VHFWLRQ ZH LGHQWLI\ 3P ZLWK WKH IXOO VXEFDWHJRU\ RI $3RV FRQWDLQLQJ DOO M PfFRPSOHWH REMHFWV ZLWK DSSURSULDWH VLJQDWXUH LH 2EM)Pf ^$ $3RV $ M$f 0$ P$f` 5DWKHU WKDQ $6 ZH GLVFXVV VXEFDWHJRULHV WKDW URXJKO\ FRUUHVSRQG WR WKH 3MQ FDWHJRULHV :H XVH WKH IROORZLQJ QRWDWLRQV JLYHQ ; 2EM$6f DQG [ ; H[[f GHQRWHV WKH FODVV RI DOO RSHQ VHWV PHPEHUV RI e!f WKDW FRQWDLQ [ :H UHIHU WR PHPEHUV RI 2EM$6f DV VSDFHV :H DVVXPH DOO VSDFHV VDWLVI\ D VHSDUDWLRQ D[LRP ZKLFK JHQHUDOL]HV WKH XVXDO 7 D[LRP 1DPHO\ ZH UHTXLUH VHSf 9[ \ ; H[f H\f [ \ $V LQ JHQHUDO WRSRORJ\ ZKHQHYHU WKH D[LRP VHSf KROGV IRU ; WKH VREHULILFDWLRQ PDS H[ LV LQMHFWLYH 'HILQH 6P E\ 6L ^; $6 ';f VHSf KROGV IRU ;`
PAGE 103
5HPDUN 7R VWUHQJWKHQ WKH DQDORJ\ EHWZHHQ WRSRORJLFDO VSDFHV DQG 3PREMHFWV ZH DVVXPH WKDW WKH VXEVHW V\VWHPV M DQG P SUHVHUYH UHJXODU HSLPRUSKLVPV 7KH IXQFWRULDO *DORLV FRQQHFWLRQ RI %DQDVFKHZVNL DQG %UXQV >@ UHVWULFWV WR D GXDOLW\ EHWZHHQ DQG 6S3P %HFDXVH RI VHSf H LGcM fÂ§! 'n e!P LV D PRQRUHIOHFWLRQ RQ 6A 7KH 6NXOD 7RSRORJ\ DQG ([WUHPDO 0RQRV $VVXPH M DQG P SUHVHUYH UHJXODU HSLPRUSKLVPV ,I 6 & ; f 2EM6Af ZH PD\ SXW D 6AVWUXFWXUH RQ 6 E\ GHILQLQJ '^6f ^8 Q 6 8 f ';f` %\ DQDORJ\ ZLWK JHQHUDO WRSRORJ\ ZH FDOO 6 ZLWK WKLV 6P D M PfVXEVSDFH %HFDXVH WKH SRVHW PDS GXDO WR WKH LQFOXVLRQ 6 & ; LV RQWR DQG M DQG P DUH VXEVHW V\VWHPV WKDW SUHVHUYH UHJXODU HSLV '6f LV M PfFRPSOHWH 3URPSWHG E\ WKH WUHDWPHQW RI VLPLODU SUREOHPV LQ 7R DQG VREHU WRSRORJLFDO VSDFHV ZH GHILQH D WRSRORJ\ RQ XQGHUO\LQJ VHWVf RI REMHFWV IURP 6A VHH 6NXOD >@ :H GHILQH WKH 6NXOD WRSRORJ\ RQ $ f 2EM6Af WR EH WKH WRSRORJ\ ZLWK EDVH '^;f 8 ^; ? 8 8 e ';f` ,W PD\ VHHP VRPHZKDW FRQIXVLQJ WR SXW DQ DFWXDO WRSRORJ\ RQ D JHQHUDOL]HG WRSRORJLFDO VSDFH EXW LW LV DOVR XVHIXO &RUROODULHV DQG LGHQWLI\ HSLPRUSKLVPV DQG H[WUHPDO PRQRPRUSKLVPV XVLQJ WKH 6NXOD WRSRORJ\ :LWKRXW LW WKH UHVXOWV EHFRPH PRUH FXPEHUVRPH WR VWDWH DQG OHVV LQWXLWLYH /HPPD 7KH 6NXODFORVXUH RI 6 & ; GHQRWHG E6f LV JLYHQ E\ WKH IRUPXOD E6f ^[ f ; 9*; f ^;f *AQ 6 *Q 6 [ *[ $*`
PAGE 104
3URRI /HW En6f GHQRWH WKH ULJKW KDQG VLGH RI WKH JLYHQ IRUPXOD 7KH RSHUDWRU En LV RUGHU SUHVHUYLQJ DQG LQFUHDVLQJ ,I 6 & 8 ZH KDYH D WUDFH PDS IURP 8f fÂ§! f JLYHQ E\ WU$f $ IO 1RWH WKDW En6f LV WKH ODUJHVW VXEVSDFH RI ; VR WKDW WKH WUDFH PDS LV LQMHFWLYH 7KXV En LV DOVR LGHPSRWHQW DQG LQ IDFW D FORVXUH RSHUDWRU ,W QRZ VXIILFHV WR VKRZ IL[nf IL[f :H ILUVW VKRZ IL[Enf & IL[f 6XSSRVH E?$f $ DQG OHW 8 EH WKH FRPSOHPHQW RI $ DQG [ ( 8 :H ZLOO JLYH D 6NXODRSHQ QHLJKERRUKRRG RI [ FRQWDLQHG LQ 8 ,I [ e En$f WKHQ ZH KDYH *L f 2;f ZLWK [ ( *L?* DQG *[ IO $ IO $ (YLGHQWO\ *L IO ; ? *f LV D 6NXOD QHLJKERRUKRRG RI [ FRQWDLQHG LQ 8 )RU IL[f & IL[nf LW VXIILFHV WR VKRZ 6';?7f LV En FORVHG ZKHQHYHU 7 ( ';f /HW [ J E?6 Q ^; ? 7ff ZLWK 7 H 2;f 7KHQ ZH KDYH *8* ( ';f ZLWK *QVQ;?7f *QVQ;?7f DQG [ ( *L? ,W IROORZV WKDW [ 6 IO ; ? 7f Â’ &RUROODU\ /HW M DQG P EH VXEVHW V\VWHPV ZKLFK SUHVHUYH UHJXODU HSLV $ PDS I ; fÂ§! < ( LV HSL LI DQG RQO\ LIEI;ff < :H VD\ I ; fÂ§r < LV D VXE VSDFH HPEHGGLQJ LI FRUHVWULFWHG WR I;f FRQVLGHUHG DV D M PfVXEVSDFH RI < LV DQ iP LVRPRUSKLVP &RUROODU\ /HW M DQG P EH VXEVHW V\VWHPV ZKLFK SUHVHUYH UHJXODU HSLV $ PDS I ; fÂ§} < LV H[WUHPDO PRQR LI DQG RQO\ LI I LV D MPfVXEVSDFH HPEHGGLQJ DQG EI;ff fÂ§ rff 3URRI 6XSSRVH LV H[WUHPDO PRQR :H PD\ IDFWRU ; fÂ§r EI;ff fÂ§ < WKH ILUVW IDFWRU LV HSL HUJR DQ LVRPRUSKLVP &RQYHUVHO\ FRQVLGHU WKH PDS EI;ff & < ,I WKLV IDFWRUV DV JV ZKHUH V LV HSL ZLWK FRGRPDLQ = ZH FDQ IDFWRU J DV = fÂ§r EJ=ff & < 7KH FRPSRVLWLRQ EI;ff fÂ§! = fÂ§! EI;ff LV DQ HSL FRQWDLQPHQW RI VXEVSDFHV RI < %\ EI;ff fÂ§ EJ=ff VR V LV DQ LVRPRUSKLVP Â’
PAGE 105
&RUROODU\ /HW M DQG P EH VXEVHW V\VWHPV ZKLFK SUHVHUYH UHJXODU HSLV 6XSSRVH ; 2EM6R6ÂLV VREHU 7KHQ 6 & ; LV VREHU LI DQG RQO\ LIE6f 6 3URRI 6LQFH 6R6P LV PRQRUHIOHFWLYH LQ iP DQG 6NXODFORVHG VXEREMHFWV DUH H[WUHPDO LI E$f $ WKHQ $ LV VREHU )RU WKH FRQYHUVH VXSSRVH $ LV VREHU :H KDYH WKH HSL FRQn WDLQPHQW $ & E$f RI VREHU VSDFHV 6LQFH WKH FRQWDLQPHQW LV HSL LWV GXDO LV LQMHFWLYH 6LQFH LW LV D FRQWDLQPHQW WKH GXDO LV VXUMHFWLYH :H QRZ DSSO\ WKH IDFW WKDW ELMHFWLYH PDSV LQ 3A DUH LVRPRUSKLVPV Â’ /HPPD /HW IJ$fÂ§!%f DQG F FRHTPLQIJf % & 7KH 6S3PFRHTXDOL]HU RI IJf LV REWDLQHG E\ IROORZLQJ F ZLWK WKH VSDWLDO UHIOHFWLRQ 7KHRUHP /HW M DQG P EH VXEVHW V\VWHPV ZKLFK SUHVHUYH UHJXODU HSLV 6XSSRVH I ; fÂ§} < 6R6P LV D PDS RI VREHU REMHFWV ZLWK GXDO I! 7I 7KH IROORZLQJ DUH HTXLYDOHQW I LV H[WUHPDO PRQR Mf LV H[WUHPDO HSL I LV D VXEVSDFH LQFOXVLRQ M I LV D TXRWLHQW PDS LQ 3Pf VXFK WKDW WKH FRGRPDLQ LV VSDWLDO I! LV RQWR Mf LV UHJXODU HSL LV UHJXODU PRQR 3URRI R f DQG m f KROG EHFDXVH 6R6P DQG 6S)P DUH GXDO R f IROORZV IURP DQG !f f KROGV EHFDXVH WKH 6E6AVWUXFWXUH RQ D VXEVSDFH LV GHILQHG E\ D 3AFRQJUXHQFH A m f KROGV E\ DQG f KROGV LQ DQ\ FDWHJRU\ VHH /HPPD Â’
PAGE 106
&RPSXWLQJ /LPLWV :H GHVFULEH KRZ WR FDOFXODWH OLPLWV LQ 6A 7R WKLV HQG ZH SURYH D UHVXOW DQDORJRXV WR WKH IDFW LQ JHQHUDO WRSRORJ\ WKDW DQ\ IDPLO\ RI VXEVHWV JHQHUDWHV D VPDOOHVW WRSRORJ\ /HPPD /HW ; EH D VHW ,I LV D FROOHFWLRQ RI IDPLOLHV RI VXEVHWV RI ; VXFK WKDW HDFK W e LV FORVHG XQGHU MXQLRQV DQG PLQWHUVHFWLRQV WKHQ D LV DOVR FORVHG XQGHU MXQLRQV DQG PLQWHUVHFWLRQV 3URRI )RU DQ\ UH6 WKH LQFOXVLRQ PDS IURP D fÂ§! U SUHVHUYHV VXEVHW LQFOXVLRQ 7KXV IRU DQ\ W e M^Df & MMf DQG PDf & PUf VR LI 6 f MDf UHVS 6 UQDff WKHQ _f W UHVSHFWLYHO\ S_ 6 Uf +HQFH D If 6n LV FORVHG XQGHU XQLRQV RI MIDPLOLHV DQG LQWHUVHFWLRQV RI PIDPLOLHV Â’ ,I T LV D IDPLO\ RI VXEVHWV RI ; ZH FRQVLGHU WKH FROOHFWLRQ RI DOO M PfFRPSOHWH IDPLOLHV RI VHWV RI ; WKDW FRQWDLQ D IM LV FDOOHG WKH M PfFRPSOHWH IDPLO\ JHQHUDWHG E\ D 'HILQLWLRQ DQG 5HPDUNV 'HILQLQJ 6ASURGXFWV &RQVLGHU D VHW ^;` & 2EMif 7KH XQGHUO\LQJ VHW RI ; LV WKH FDUWHVLDQ SURGXFW 7,;b ZLWK FDQRQLFDO SURMHFWLRQV UÂ ; fÂ§ ;O ef;f LV WKH M PfFRPSOHWH IDPLO\ JHQHUDWHG E\ VHWV RI WKH IRUP LUaO8f ZKHUH 8 & ';Âf 7KH REMHFW ; WKXV GHILQHG VDWLVILHV VHSf LI [ \ f ; DQG [ A \ WKHUH H[LVWV L VXFK WKDW 7L[f 7L\f 6LQFH ;W VDWLVILHV VHSf WKHUH LV 8 f ';Âf ZLWK ?8 Q ^UÂ[UÂM`_ 7KXV WKHUH LV 9 2;f QDPHO\ 9 fÂ§ 7Weff ZLWK ?9 Q ^[\`? fÂ§ ; LV WKH FDWHJRULFDO SURGXFW RI ^;` VLQFH WKH XQGHUO\LQJ VHW RI ; LV WKH FDUWHVLDQ SURGXFW RI ^;Â` ZKHQHYHU ZH KDYH DQ REMHFW < f 2EM6Af DQG IXQFWLRQV IL < fÂ§ ;Wf WKHUH LV D XQLTXH IXQFWLRQ < fÂ§! ; VR WKDW IRU DOO L UÂ Â ,I HDFK Â LV DQ 6APDS WKHQ EHFDXVH RI RXU GHILQLWLRQ RI ';f LV DOVR DQ 6APDS 'HILQLWLRQ DQG 5HPDUNV 6AHTXDOL]HUV /HW M DQG P EH VXEVHW V\VWHPV ZKLFK SUHVHUYH UHJXODU HSLV /HW IJ ; fÂ§! < EH D SDLU RI PDSV LQ 6A 7KH VHW ( ^[ f ;
PAGE 107
I[f J[f` KDV D QDWXUDO 6AVWUXFWXUH (f ^8 IO ( 8 f ';f` 7KH UHODWLRQ 5 RQ ';f JLYHQ E\ XUY X Q ( Y Q ( LV D 3AFRQJUXHQFH EHFDXVH IfXLQ( I_LfQe L L DQG AQe WÂfQe L L IRU DQ\ VHW ( DQG IDPLO\ RI VXEVHWV ^f 6LQFH 2(f LV WKH TXRWLHQW RI 2;f E\ D 3A FRQJUXHQFH DQG M DQG P SUHVHUYH UHJXODU HSLV 2(f LV MPfFRPSOHWH ( VDWLVILHV VHSf EHFDXVH ; GRHV ,I K = fÂ§! ; HTXDOL]HV J WKHQ WKH LPDJH RI K PXVW EH FRQWDLQHG LQ ( 6LQFH K LV DQ 6APDS WKH LQFOXVLRQ RI WKH LPDJH RI K LQ ( LV DOVR DQ 6APDS 4XRWLHQWV ([WUHPDO DQG 5HJXODU (SLV 'HILQLWLRQ DQG 5HPDUNV /HW ; 2EM6Af (TY;f GHQRWH WKH VHW RI HTXLYDOHQFH UHODWLRQV RQ ; DQG GHQRWH WKH VHW RI VXEIDPLOLHV RI 2;f :H FRQVLGHU HTXLYDOHQFH UHODWLRQV WR EH VHWV RI RUGHUHG SDLUV ,I 5 f (TY;f DQG [ f ; ZH XVH WKH QRWDWLRQ 5[f >\ e ; [ \f 5` IRU WKH HTXLYDOHQFH FODVV RI [ ,I 5 H (TY;f ZH VD\ 8 f 2;f LV 5VDWXUDWHG LI [ \f f 5 DQG [ 8 LPSO\ WKDW \ f 8? OHW 5r GHQRWH WKH IDPLO\ RI DOO 5 VDWXUDWHG PHPEHUV RI 2;f ,I 5? & 5 WKHQ Ur F U? ,I D f "3 ZH GHILQH DQ HTXLYDOHQFH UHODWLRQ Dr RQ ; E\ [ \f Dr LI DQG RQO\ LI IRU HDFK 8 ZH KDYH [ 8 \ 8 1RWH WKDW FT & D FU & MM ,I D r` WKHQ HDFK PHPEHU RI D LV FUrVDWXUDWHG ,Q V\PEROV D & Drr 6XSSRVH 5 (TY;f 8 ';f LV 5VDWXUDWHG LI DQG RQO\ LI IRU DOO [\f 5 [ 8 Â‘ \ 7KXV 5 & 5rr
PAGE 108
7R VXPPDUL]H r JLYHV D *DORLV FRQQHFWLRQ EHWZHHQ (TY;f DQG r` *LYHQ 5 f (TY;f ZH GHILQH DQ 6AVWUXFWXUH RQ WKH VHW ;5 RI HTXLYDOHQFH FODVVHV E\ GHILQLQJ ;5f WR EH WKH IDPLO\ ^^5[f [ 8` 8 5r` 6LQFH DUELWUDU\ XQLRQV DQG LQWHUVHFWLRQV RI VDWXUDWHG VHWV DUH VDWXUDWHG DQG VLQFH WKH LQFOXVLRQ RI 5r LQ 2;f LV RUGHU SUHVHUYLQJ ';5f LV M PfFRPSOHWH :H RPLW WKH SURRIV RI WKH IROORZLQJ OHPPDV ZKLFK DUH VWUDLJKWIRUZDUG YHULILFDWLRQV /HPPD ,I I ; fÂ§r < 0DS6Af 5 (TY;f DQG I LV FRQVWDQW RQ HDFK HTXLYDn OHQFH FODVV 5[f WKHQ I IDFWRUV DV ; fÂ§! ;5 fÂ§r < ZKHUH ; fÂ§} ;5 LV WKH TXRWLHQW PDS DQG ;5 fÂ§! < LV GHILQHG E\ 5[f I^[f )RU DOO 8 '7f IaO8f LV 5VDWXUDWHG DQG LQ 2;f /HPPD /HW 5?5 e (TY;f ZLWK TXRWLHQW PDSV ; fÂ§! ;5L 7KHUH LV D TXRWLHQW PDS T ;5? fÂ§r ;5 VXFK WKDW T TT? LI DQG RQO\ LI 5? & 5 /HPPD )RU 5 (TY;f WKH IROORZLQJ DUH HTXLYDOHQW 5rr 5 WKHUH LV FU ( S VXFK WKDW 5 Dr ;5 VDWLVILHV VHSf 3URRI 7KH HTXLYDOHQFH RI WKH ILUVW WZR FRQGLWLRQV LV WULYLDO ,I 5 Dr DQG 5[f A 5\f WKHQ WKHUH H[LVWV 8 D VXFK WKDW _8 IO ^D\`_ 6R f } f 6LQFH 5 & 5rr ZH KDYH D TXRWLHQW PDS T ;5 fÂ§ ;5rr DV GHVFULEHG LQ WKH SUHFHGLQJ OHPPD ,W VXIILFHV WR VKRZ WKDW T LV LQMHFWLYH ,I 5[f 5\f DQG ;5 VDWLVILHV VHSf WKHQ WKHUH LV 8 5r ZLWK ?8 ^[\`? fÂ§ 7KXV 5rr[f s 5rr^\f Â’ &RUROODU\ ,I ; LV D 7R WRSRORJLFDO VSDFH DQG 5 (TY;f WKHQ ;5 LV 7R LI DQG RQO\ LI 5 fÂ§ 5rr
PAGE 109
/HPPD ,I I ; fÂ§! < f 6P DQG 5 (TY;f LV GHILQHG E\ 5 ^]\f IL[f I^\f` WKHQ 5rr 5 3URRI :H XVH /HPPD 6LQFH < VDWLVILHV VHSf DQG IRU DOO 8
PAGE 110
M^& [ & & ;5M D GLVMRLQW XQLRQ DQG GHILQH 2=f =f 2=f LV FRPSOHWH KHQFH M PfFRPSOHWHf VR = 2EM6Af PRUHRYHU DQ\ VHW PDS IURP = fÂ§! ; LV D 6APDS :H QRZ GHILQH WZR PDSV J K = fÂ§! ; E\ J[ \f [ DQG K[ \f \ %\ FRQVWUXFWLRQ LV WKH FRHTXDOL]HU RI JK DV FRPSXWHG LQ 6HW 6LQFH LV DOVR D 6ATXRWLHQW PDS LV DOVR WKH FRHTXDOL]HU RI J K LQ iP f KROGV LQ DQ\ FDWHJRU\ > @ 7R VKRZ Â‘ f VXSSRVH LV H[WUHPDO HSL :H PD\ IDFWRU LQWR D TXRWLHQW PDS T IROORZHG E\ DQ LQMHFWLRQ L XVLQJ /HPPD 6LQFH LV HSL WKHQ L LV DOVR HSL LW QRZ IROORZV E\ WKH H[WUHPDO SURSHUW\ RI WKDW L LV DQ LVRPRUSKLVP 6R LV D TXRWLHQW PDS Â’ )ODW 6SHFWUD /HW M DQG P EH VXEVHW V\VWHPV ZKLFK SUHVHUYH UHJXODU HSLV :H QRZ H[DPLQH D PRQRFRUHIOHFWLYH FDWHJRU\ RI 6SgP 7R GHILQH WKH FDWHJRU\ ZH QHHG WR SXW D SDUWLDO RUGHU RQ $ IRU $ f gP 5HFDOO WKDW n/$ LV WKH VHW RI IXQFWLRQV $ fÂ§! ZKLFK SUHVHUYH RUGHU DQG M PfRSWLPXP ERXQGVf :H WKLQN RI [ \ ?/$ DV FKDUDFWHUV [ \ $ fÂ§r :H GHILQH [ \ WR PHDQ 9D $[Df \Df &OHDUO\ [ \ LI DQG RQO\ LI H[f ^D $ [Df ` & H\f 7KH UHODWLRQ LV DQWLV\PPHWULF E\ YLUWXH RI WKH IDFW WKDW RXU VSDFHV VDWLVI\ VHSf 1RZ ZH GHILQH WR EH WKH IXOO VXEFDWHJRU\ RI FRQWDLQLQJ DOO $ ZLWK WKH SURSHUW\ WKDW IRU DOO [ \ WI\$f [ \ LPSOLHV [ fÂ§ \ 7R VKRZ )6@P PRQRFRUHIOHFWLYH ZH ZLOO XVH WKH FDWHJRU\ WKHRUHWLF GXDO RI WKH IROORZLQJ WKHRUHP 7KHRUHP > @ 6XSSRVH $ LV D FRZHOOSRZHUHG FRPSOHWH FDWHJRU\ DQG 6LVD IXOO LVRPRUSKLVP FORVHG VXEFDWHJRU\ % LV HSLUHIOHFWLYH LI DQG RQO\ LI LW LV FORVHG XQGHU SURGXFWV DQG H[WUHPDO VXEREMHFWV 3URSRVLWLRQ LV PRQRFRUHIOHFWLYH LQ 6S3A 3URRI 6S)P LV FRZHOOSRZHUHG DQG FRFRPSOHWH EHFDXVH LW LV ZHOOSRZHUHG FRPSOHWH DQG KDV D FRVHSDUDWRU > @ :H VKRZ WKDW )A LV FORVHG XQGHU FRSURGXFWV 6XSSRVH WKH
PAGE 111
FRSURGXFW 8 $L KDV FKDUDFWHUV [ \ 6LQFH [ A \ WKHUH LV DQ LQGH[ L VXFK WKDW [SL s \SW %XW [UÂ\LÂ DUH FKDUDFWHUV RI $L 7KXV VRPH $W KDV FKDUDFWHUV [ \ :H VKRZ WKDW LV FORVHG XQGHU H[WUHPDO HSLV 6XSSRVH $ KDV IODW VSHFWUXP DQG $ fÂ§! $. LV D TXRWLHQW ,I $ LV QRQWULYDO WKHQ LW FDQQRW KDYH FRQVWDQW FKDUDFWHUV 7KXV LI $ KDV IODW VSHFWUXP DOO FKDUDFWHUV DUH RQWR 1RZ ZH XVH WR FRQFOXGH WKDW FKDUDFWHUV RI $. DUH SUHFLVHO\ FKDUDFWHUV RI $ ZKRVH NHUQHO FRQWDLQV ,W IROORZV WKDW $. KDV IODW VSHFWUXP Â’ 5HPDUN 7KH WKHRUHP ZRXOG UHPDLQ WUXH LI ZH UHGHILQHG WR EH D IXOO VXEFDWHJRU\ RI FRQWDLQLQJ IODW VSHFWUXP REMHFWV 7KH UDWLRQDOH IRU QRW XVLQJ WKLV GHILQLWLRQ RI )Â LV WKDW QRQVSDWLDO IODW VSHFWUXP REMHFWV EHKDYH UDWKHU GLIIHUHQWO\ WKDQ VSDWLDO REMHFWV :LWKRXW DVVXPLQJ $ VSDWLDO WKH VSHFWUXP JLYHV LQFRPSOHWH LQIRUPDWLRQ )RU LQVWDQFH WKHUH DUH $ ZLWK QR FKDUDFWHUV ZKDWVRHYHU 7KHVH REMHFWV WULYLDOO\ KDYH IODW VSHFWUXP :H ZLVK WR DYRLG WKLV SUREOHP (SLFRPSOHWH REMHFWV LQ 6P $VVXPH WKURXJKW WKLV VHFWLRQ WKDW M DQG P SUHVHUYH UHJXODU HSLPRUSKLVPV 'HILQLWLRQ $Q REMHFW $ LQ DQ\ FDWHJRU\ LV HSLFRPSOHWH LI DQ\ PDS I $ fÂ§! % WKDW LV ERWK HSL DQG PRQR LV DQ LVRPRUSKLVP 1RWH DQ\ PRQRUHIOHFWLRQ PDS LV HSL VHH +HUUOLFK DQG 6WUHFNHU > 6HFWLRQ @ 6REHULILFDWLRQ LV D PRQRUHIOHFWLRQ RQ iP 7KHUHIRUH DQ\ HSLFRPSOHWH REMHFW LQ 6P PXVW EH VREHU :H QRZ H[DPLQH ZKLFK VREHU VSDFHV DUH HSLFRPSOHWH :H UHFDOO WKH SDUWLDO RUGHU RQ FKDUDFWHUV $ fÂ§r WKDW ZDV XVHG LQ WKH SUHFHGLQJ VXEVHFWLRQ )RU D VREHU VSDFH WKH SRLQWV [ f ; DUH LQ ELMHFWLYH FRUUHVSRQGHQFH WR 6APDSV ';f fÂ§! ,I $ LV D VSDFH [ \ PHDQV H[f & H\f 1RWH WKDW LI 8 f ^;f [ 8 DQG [ \ WKHQ \ f 8@ LQ RWKHU ZRUGV HDFK 8 2;f LV D LQFUHDVLQJ VXEVHW RI ; $OVR QRWH WKDW DQ\ FRQWLQXRXV PDS SUHVHUYHV WKLV RUGHU /HPPD 6XSSRVH ; f 2EM6Af ZLWK X & M IO P
PAGE 112
; LV D FKDLQ LI DQG RQO\ LI ';f LV D FKDLQ ,I ; LV QRW D FKDLQ WKHQ WKHUH LV D MPfVWUXFWXUH U RQ ; VXFK WKDW U & ';f DQG ;Wf VDWLVILHV VHSf 3URRI ,I ; LV D FKDLQ WKHQ WKH IDPLO\ RI DOO LQFUHDVLQJ VXEVHWV RI ; LV D FKDLQ RUGHUHG E\ LQFOXVLRQf 6LQFH e!;f & e!;f LV D FKDLQ &RQYHUVHO\ LI ef;f LV D FKDLQ WKHQ WKH VHW RI SULPH LGHDOV RI ';f LV D FKDLQ VR ; LV D FKDLQ 6XSSRVH ^;f LV QRW D FKDLQ 7KHQ ZH KDYH LQFRPSDUDEOH 89 f ';f /HW U ^: ( ';f : & 8 RU 8 & :` 2QH PD\ HDVLO\ YHULI\ WKDW U LV FORVHG XQGHU DOO XQLRQV DQG LQWHUVHFWLRQV ZKLFK H[LVW LQ 2;f VLQFH MUf & M;ff DQG PUf & P';ff WKLV VKRZV WKDW U LV D M PfFRPSOHWH :H QRZ YHULI\ WKDW ; Uf VDWLILHV VHSf 6XSSRVH [ \ f ; DUH GLVWLQFW SRLQWV ZLWKRXW ORVV RI JHQHUDOLW\ WKHUH LV ( 2;f VXFK WKDW [ f DQG \ ,I [ ( 8 WKHQ IO 8 U VHSDUDWHV [ DQG \ ,I [ e 8 DQG \ t 8 WKHQ 8 8 VHSDUDWHV [ DQG \ /DVWO\ LI [ A 8 DQG \ ( 8 WKHQ 8 VHSDUDWHV [ DQG \ Â’ 7KHRUHP /HW ; 2EMR6Af OR & M Q P :H KDYH WKH LPSOLFDWLRQV Lf LLf +Lf } LYf ,I M fÂ§ RR WKHQ LYf Lf Lf ; LV HSLFRPSOHWH LLf ,I I ; fÂ§r< LV 6R3PPRQLF WKHQ I;f E I;f DQG ; DUH DOO KRPHRPRUSKLF +Lf 7KHUH LV QR MPfVWUXFWXUH U RQ ; VXFK WKDW U & 2;f DQG ;Wf VDWLVILHV VHSf LYf ; LV D FKDLQ VDWLVI\LQJ +Lf 3URRI L LLf 6XSSRVH ; fÂ§! < LV PRQLF %\ FRUHVWULFWLRQ RI ZH REWDLQ D 6APDS ; fÂ§! I;f &RPSRVLQJ ZLWK WKH LQFOXVLRQ PDS RI I;f LQ KI;f ZH REWDLQ D PRQLF HSLF 6R3P PDS ZLWK GRPDLQ ;? Lf IRUFHV WKLV PDS WR EH DQ LVRPRUSKLVP LL LLLf 6XSSRVH U LV D MPfVWUXFWXUH RQ ; DQG U & ';f 7KH LGHQWLW\ PDS RQ ; LV D FRQWLQXRXV ELMHFWLRQ ; ef;ff fÂ§! ; Uf )ROORZLQJ WKLV PDS E\ er} ; Uf fÂ§!
PAGE 113
Ân7 ZH REWDLQ D PRQR HSL PDS ;';ff fÂ§} LnU LLf LPSOLHV WKDW DOO PDSV LQYROYHG DUH LVRPRUSKLVPV LLL LYf 7KLV IROORZV IURP /HPPD LY Lf :KHQ M RR ZH KDYH PRUH LQIRUPDWLRQ DERXW ZKLFK M UHVWUXFWXUHV FRUUHVSRQG WR D JLYHQ SDUWLDOO\ RUGHUHG VHW 6XSSRVH ; f LV D SRVHW DQG U LV D WRSRORJ\ RQ ; VXFK WKDW [ \ LI DQG RQO\ LI HDFK RSHQ VHW FRQWDLQLQJ [ DOVR FRQWDLQV \ 7KH XQLRQ RI WKH IDPLO\ ^8 e W [ 8` LV RSHQ DQG HTXDOV ^\ ; \ [` 7KXV U PXVW FRQWDLQ WKH ZHDN WRSRORJ\ JHQHUDWHG E\ WKH VHWV ^\ f ; \ A [` IRU [ f ; 6XSSRVH ; LV D FKDLQ DQG ; fÂ§! < LV DQ HSLF PRQRPRUSKLVP 6LQFH LV LQMHFWLYH DQG SUHVHUYHV RUGHU I;f LV D FKDLQ RUGHU LVRPRUSKLF WR ; (UJR WKH I;f KDV D WRSRORJ\ ILQHU WKDQ WKH ZHDN WRSRORJ\ LQGXFHG E\ WKH SDUWLDO RUGHULQJ 6LQFH WKH FRUHVWULFWLRQ ; fÂ§ I;f LV FRQWLQXRXV WKH WRSRORJ\ RQ ; LV ILQHU WKDQ WKH WRSRORJ\ RQ I;f 7KXV ; DQG I;f DUH KRPHRPRUSKLF 6LQFH ; LV VREHU I;f LV VREHU ZKLFK LPSOLHV EI;ff fÂ§ I;f 6LQFH LV HSL EI;ff < VR ; DQG < DUH KRPHRPRUSKLF Â’
PAGE 114
5()(5(1&(6 >@ 6HPLQDU RQ 7ULSOHV DQG &DWHJRULFDO +RPRORJ\ 7KHRU\ (GLWHG E\ % (FNPDQQ /HFWXUH 1RWHV LQ 0DWKHPDWLFV 6SULQJHU9HU ODJ %HUOLQ1HZ @ 5 %DOEHV DQG 3 'ZLQJHU 'LVWULEXWLYH /DWWLFHV 8QLYHUVLW\ RI 0LVVRXUL 3UHVV &ROXPELD 0R f >@ % %DQDVFKHZVNL DQG %UXQV 7KH IXQGDPHQWDO GXDOLW\ RI SDUWLDOO\ RUGHUHG VHWV 2UGHU f QR >@ 0 %DUU DQG & :HOOV 7RSRVHV 7ULSOHV DQG 7KHRULHV 6SULQJHU9HUODJ %HUOLQ1HZ @ %HFN 'LVWULEXWLYH ODZV LQ 6HPLQDU RQ 7ULSOHV DQG &DWHJRULFDO +RPRORJ\ 7KHRU\ /HFWXUH 1RWHV LQ 0DWKHPDWLFV 6SULQJHU9HUODJ %HUOLQ1HZ @ ) %RUFHX[ +DQGERRN RI &DWHJRULFDO $OJHEUD YROXPHVf &DPEULGJH 8QLYHUVLW\ 3UHVV &DPEULGJH f >@ 0 (UQH=FRQWLQXRXV SRVHWV DQG WKHLU WRSRORJLFDO PDQLIHVWDWLRQ $SSOLHG &DWHJRULFDO 6WUXFWXUHV f >@ + *DLIPDQ ,QILQLWH %RROHDQ SRO\QRPLDOV )XQG 0DWK f >@ $ : +DOHV 2Q WKH QRQH[LVWHQFH RI IUHH FRPSOHWH %RROHDQ DOJHEUDV )XQG 0DWK f >@ 3 5 +DOPRV 1DLYH 6HW 7KHRU\ 8QGHUJUDGXDWH 7H[WV LQ 0DWKHPDWLFV 6SULQJHU9HUODJ %HUOLQ1HZ @ + +HUUOLFK 5HJXODU FDWHJRULHV DQG UHJXODU IXQFWRUV &DDG 0DWK f >@ + +HUUOLFK 7RSRORJLVFKH 5HIOH[LRQHQ XQG &RUHIOH[LRQHQ /HFWXUH 1RWHV LQ 0DWKHPDWLFV 6SULQJHU9HUODJ %HUOLQ1HZ @ + +HUUOLFK t 6WUHFNHU &DWHJRU\ 7KHRU\ +HOGHUPDQQ9HUODJ %HUOLQ f >@ + +RIIPDQ 0 0LVORYH )UHH 2EMHFWV LQ WKH FDWHJRU\ RI FRPSOHWHO\ GLVWULEXWLYH ODWWLFHV LQ &RQWLQXRXV /DWWLFHV DQG WKHLU $SSOLFDWLRQV /HFWXUH 1RWHV LQ 3XUH DQG $SSOLHG 0DWKHPDWLFV 0DUFHO 'HNNHU 1HZ @ 5 ,VEHOO $WRPOHVV SDUWV RI VSDFHV 0DWK 6FDQG f >@ 3 RKQVWRQH 6WRQH 6SDFHV &DPEULGJH 6WXGLHV LQ $GYDQFHG 0DWKHPDWLFV &DPEULGJH 8QLYHUVLW\ 3UHVV &DPEULGJH f
PAGE 115
>@ $ R\DO 0 7LHUQH\ $Q ([WHQVLRQ RI WKH *DORLV 7KHRU\ RI *URWKHQGLHFN 0HP $PHU 0DWK 6RF f QR >@ < /L *HQHUDOL]HG 6,QFRPSOHWH IUHH FRPSOHWH FRPSOHWHO\ GLVWULEXWLYH ODWWLFHV JHQHUn DWHG E\ SRVHWV 6HPLJURXS )RUXP f >@ / ( /LQWRQ $SSOLHG IXQFWRULDO VHPDQWLFV ,, LQ 6HPLQDU RQ 7ULSOHV DQG &DWHJRULFDO +RPRORJ\ 7KHRU\ /HFWXUH 1RWHV LQ 0DWKHPDWLFV 6SULQJHU9HU ODJ %HUOLQ1HZ @ / ( /LQWRQ &RHTXDOL]HUV LQ FDWHJRULHV RI DOJHEUDV LQ 6HPLQDU RQ 7ULSOHV DQG &DWHJRULFDO +RPRORJ\ 7KHRU\ /HFWXUH 1RWHV LQ 0DWKHPDWLFV 6SULQJHU9HUODJ %HUOLQ 1HZ @ 6 0DF/DQH &DWHJRULHV IRU WKH :RUNLQJ 0DWKHPDWLFLDQ *UDGXDWH 7H[WV LQ 0DWKHPDWLFV 6SULQJHU9HUODJ %HUOLQ1HZ @ 0DGGHQ .IUDPHV RXUQDO RI 3XUH DQG $SSOLHG $OJHEUD f >@ ( 0DQHV $OJHEUDLF 7KHRULHV *UDGXDWH 7H[WV LQ 0DWKHPDWLFV 6SULQJHU9HUODJ %HUOLQ1HZ @ 0DUNRZVN\ )UHH FRPSOHWHO\ GLVWULEXWLYH ODWWLFHV 3URF $PHU 0DWK 6RF f >@ 0HVHJXHU 2UGHU FRPSOHWLRQ PRQDGV $OJHEUD 8QLYHUVDOLV f >@ 3DUH $EVROXWH FRHTXDOL]HUV LQ &DWHJRU\ 7KHRU\ +RPRORJ\ 7KHRU\ DQG WKHLU $Sn SOLFDWLRQV /HFWXUH 1RWHV LQ 0DWKHPDWLFV 6SULQJHU9HUODJ %HUOLQ1HZ @ 1 5DQH\ &RPSOHWHO\ GLVWULEXWLYH ODWWLFHV 3URF $PHU 0DWK 6RF f >@ 1 5DQH\ 7LJKW *DORLV FRQQHFWLRQV DQG FRPSOHWH GLVWULEXWLYLW\ 7UDQV $PHU 0DWK 6RF ,f >@ 1 5DQH\ $ VXE GLUHFWXQLRQ UHSUHVHQWDWLRQ IRU FRPSOHWHO\ GLVWULEXWLYH FRPSOHWH ODWWLFHV 3URF $PHU 0DWK 6RF YRO f >@ / 6NXOD 2Q D UHIOHFWLYH VXEFDWHJRU\ RI WKH FDWHJRU\ RI DOO WRSRORJLFDO VSDFHV 7UDQV $PHU 0DWK 6RF f >@ : 7KDWFKHU % :ULJKW t ( :DJQHU $ XQLIRUP DSSURDFK WR LQGXFWLYH SRVHWV DQG LQGXFWLYH FORVXUH 7KHRUHWLFDO &RPSXWHU 6FLHQFH f >@ : 5 7XQQLFOLII 7KH IUHH FRPSOHWHO\ GLVWULEXWLYH ODWWLFH RYHU D SRVHW $OJHEUD 8QLYHUn VDOLV f
PAGE 116
%,2*5$3+,&$/ 6.(7&+ (ULF 5LFKDUG =HQN ZDV ERUQ RQ )HEUXDU\ LQ )ULGOH\ 01 +H OLYHG LQ $Qn GRYHU 01 IRU WKH IROORZLQJ HLJKWHHQ \HDUV 7KHQ (ULF DWWHQGHG 6W RKQV 8QLYHUVLW\ LQ &ROOHJHYLOOH 01 ZKHUH KH PDMRUHG LQ PDWKHPDWLFV DQG SK\VLFV 8SRQ JUDGXDWLQJ LQ (ULF PRYHG WR )ORULGD ZKHUH KH VWXGLHG PDWKHPDWLFV DW WKH JUDGXDWH OHYHO DW WKH 8QLYHUVLW\ RI )ORULGD
PAGE 117
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n$n 6FRWW 0F&XOORXJK 3URIHVVRU RI 0DWKHPDWLFV WKDW KDYH UHDG WKLV VWXG\ DQG WKDW LQ P\ RSLQLRQ LW FRQIRUPV WR DFFHSW RI VFKRODUO\ SUHVHQWDWLRQ DQG LV IXOO\ DGHTXDWH LQ VFRSH DQG TXDOLW\ DV D WKH GHJUHH RI 'RFWRU RI 3KLORVRSK\ 5H\LDOG LPHQH] $VVRFLDWH 3URIHVVRU RI 5RPDQFH /DQJXDJHV DQG /LWHUDWXUHV 7KLV GLVVHUWDWLRQ ZDV VXEPLWWHG WR WKH *UDGXDWH )DFXOW\ RI WKH 'HSDUWPHQW RI 0DWKn HPDWLFV LQ WKH &ROOHJH RI /LEHUDO $UWV DQG 6FLHQFHV DQG WR WKH *UDGXDWH 6FKRRO DQG ZDV DFFHSWHG DV SDUWLDO IXOILOOPHQW RI WKH UHTXLUHPHQWV IRU WKH GHJUHH RI 'RFWRU RI 3KLORVRSK\ $XJXVW 'HDQ *UDGXDWH 6FKRRO
PAGE 118
/' 2$ A 81,9(56,7< 2) )/25,'$
LD
1780
20 OA
.75^
UNIVERSITY OF FLORIDA
3 1262 08554 2651
xml version 1.0 encoding UTF8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchemainstance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID ETRHCXIK6_XQBS3I INGEST_TIME 20110929T20:49:21Z PACKAGE AA00004701_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
SUBSET SYSTEMS AND GENERALIZED DISTRIBUTIVE LATTICES
By
ERIC R. ZENK
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
2004
ACKNOWLEDGMENTS
I would like to thank 
â€¢ Jorge Martinez. Sometimes words are inadaquate. He has been kind, generous, patient,
and interesting to work with. I have grown under his guidance.
â€¢ The people who inhabited the University of Florida math department from 1999 to
2004. They make the department the wonderful place it is.
â€¢ My friend Cielo: earth is wonderful, when one can see into the sky.
â€¢ My M. L.
â€¢ My mathematical siblings  other graduate students who worked with Jorge, especially
Ricardo Carrera.
â€¢ My family  Wayne, Phyllis, Margie, Jeff, and Rob.
â€¢ Those teachers who, by challenging me, caused me to improve.
â€¢ Participants in the â€œfRings and Ordered Algebraic Structuresâ€ conferences in GainesÂ¬
ville and Nashville during my time in graduate school. Because of these conferences, I
feel like I am joining a family of researchers, rather than just â€œgetting a degreeâ€. It is
an honor to know them.
â€¢ Those who helped proofread this document: constructive comments were made by
Jorge Martinez, Scott McCullough, and Pham Tiep.
â€¢ Anyone taking the time to read these words: a dissertation, like any book, is meant to
be read.
u
TABLE OF CONTENTS
ACKNOWLEDGEMENTS Ã¼
ABSTRACT v
CHAPTERS
1 INTRODUCTION 1
1.1 Distributive Lattices 1
1.2 Subset Systems
1.3 Methods and Results 5
2 PRIMER ON CATEGORIES AND POSETS 8
2.1 Distinguished Maps
2.2 Bounds 12
2.3 Natural Transformations 17
2.4 (Co)Limits 19
2.5 Adjoint Functors 23
3 ALGEBRAS OF A MONAD 28
3.1 Categories of Algebras 28
3.2 Adjoint Connections induce Monads 32
3.3 Detecting Categories of Algebras 34
3.4 Distributive Laws 37
4 GENERATING SUBMONADS 42
4.1 Subfunctors 42
4.2 Meseguerâ€™s Lemmas 50
4.3 A Partial Algebra Which Does Not Extend 59
5 FREE ALGEBRAS 63
5.1 Complete semilattices 64
5.2 Completely Distributive Complete Lattices 68
5.3 Some categories of algebras 71
6 COEQUALIZERS 82
6.1 Epis and Equalizers in P 82
iii
6.2 Factorization of Maps Using Preorders 85
6.3 Factorization of Meetsemilattice maps 88
6.4 Coequalizers in DUâ€, 90
7 (j,m)SPACES 93
7.1 Spatial/Sober Functorial Galois Connection 94
7.2 The Skula Topology and Extremal Monos 97
7.3 Computing Limits 100
7.4 Quotients, Extremal and Regular Epis 101
7.5 Flat Spectra 104
7.6 Epicomplete objects in 105
REFERENCES 108
BIOGRAPHICAL SKETCH HO
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
SUBSET SYSTEMS AND GENERALIZED DISTRIBUTIVE LATTICES
By
Eric R. Zenk
August 2004
Chairman: Jorge Martinez
Major Department: Mathematics
Distributive lattices  alone, or with enriched structure  are mathematical objects
of fundamental importance. This text studies generalized distributive lattices: the generalÂ¬
ization is that certain infinite meets and joins are required to exist. Subset systems (natural
rules which select a family of subsets of each poset) j and m label which sets have joins and
meets, respectively.
A calculus of subfunctors is developed: using this calculus, it is shown that any
subfunctor F of a monad (containing the image of the unit) generates a submonad F. Under
suitable conditions, any partial Falgebra extends to an Falgebra. The monad F for the free
distributive (j. m)complete lattice is the submonad of the completely distributive complete
lattice monad generated by a subfunctor obtained from j and m.
The category DP^ of (j. m)complete lattices which can be embedded in a completely
distributive complete lattice is a full subcategory of Falgebras. DP(n is complete and has
coequalizers.
(j, m)complete families of subsets of a set (generalized topological spaces) are invesÂ¬
tigated in analogy to classical pointset topology. Assuming suitable restrictions on j and
m, subspaces can be defined. Assuming these restrictions, there are wellbehaved categories
corresponding to To and sober spaces.
CHAPTER 1
INTRODUCTION
1.1 Distributive Lattices
A semilattice is a set with an operation A satisfying the following universally quantified
equations
â€¢ a A a = a,
â€¢ a A b â€” b A a, and
â€¢ (a A b) A c = a A (b A c).
A and V, which are related by the condition
a V b â€” b.
< b to mean a A b = a; with this order a A b
is the largest thing smaller than both a and b and a V b is the smallest thing larger than both
a and b.
A lattice is distributive if either of the following, equivalent, universally quantified
equations hold
a A (b V c) â€” (a A b) V (a A c),
a V (6 A c) = (a V b) A (a V c).
Let us consider the concept and its relevance.
A distributive lattice bears some resemblance to ordinary arithemetic where A and
V correspond to addition and multiplication; the principal difficulty with this view is that
a A a = a V a = a, which does not hold in arithmetic. There is more symmetry in the
A lattice is a set with two semilattice operations
that
a A b = a 4=>
A lattice may be partially ordered by defining a
1
2
equations defining distributive lattices than in ordinary arithmetic; formally â€œAâ€ and â€œVâ€ are
interchangable, and switching them reverses the order. Distributive lattices are interesting
algebraic structures in the same right as rings  structures with + and â– sensibly defined.
Another perspective is that distributive lattices are models of logic, with â€œAâ€ repreÂ¬
senting â€œandâ€ and â€œVâ€ representing â€œor.â€ Obviously, the connectives â€œandâ€ and â€œorâ€ both
satisfy the semilattice rules. In this interpretation, the distributive laws are tautologies and
a < b means the proposition a implies b. Conventional logics are often described by disÂ¬
tributive lattices obeying extra equations, which correspond to additional tautologies to be
modeled.
A perspective particularly revelant to the author is that distributive lattices (with
some additional structure) describe topological situations. Intuitively, a topological space
is an amorphous blob, from which certain pieces can be cleanly removed. The removable
pieces are called closed parts and the complements (i.e, things left over after a closed part
has been removed) are called open parts. The usual definition of a topological space is a set
X, together with a designated family of open subsets, such that
â€¢ X and the empty subset are open,
â€¢ if U and V are open, so is the intersection U ft V, and
â€¢ if (Ui) is a family of open subsets, then the union Jt is also open.
The lattice of open sets encodes how a space is woven together. Real analysis provides some
justification for the usual definition of a topological space. However, the author wondered
how the concept of a topology changes if one varies the definition by requiring either fewer
unions of open sets be open, or more intersections of open sets be open.
1.2 Subset Systems
A category is an abstract class of objects with structure, and maps (or, homomor
phisms) which preserve the relevant structure. The maps in a category allow comparisons
3
between objects. Category theory allows formal comparisons between various theories of
algebraic objects: e.g., one can compare the category of all rings with the category of all
distributive lattices.
Many categories of partially ordered sets (with additional structure) fit into a simple,
general pattern. The objects are partially ordered sets in which certain intentionally distinÂ¬
guished subsets have suprema or infima while the maps are order preserving functions which
preserve said infima and suprema. P denotes the category of all posets and order preserving
maps.
The challenge is how one selects subsets which have infima and suprema. We follow
Thatcher, Wright and Wagner [31], who introduced the useful (but blandly named) concept
of a subset system.
Definition and Remarks 1.2.1. A subset system Z, is a rule which assigns a family Z(A)
of subsets to any poset A, such that for any order preserving map / : A â€”> B,
{,f(S) : 5 G Z(A)} C Z(B).
Zcomplete posets are posets A in which each set S G Z(A) has a supremum. (See Erne [7]
for more extensive bibliography regarding subset systems.)
The following examples of subset systems may convey the generality and usefulness
of the concept.
1. If k is a cardinal, we use k to denote the subset system which selects all subsets
with cardinality less than k. We use oo for the functor which places no restriction on
cardinality. This is a subset system because for any function / : A â€”> B, S C A, with
A < k implies /(A) < k.
2. A subset S C A is (upward) directed if for each x,y G S, there exists u G S such that
x < u and y < it. The rule dir which selects all directed subsets of a poset is a subset
system: if f{x),f(y) G /(S') and S is directed, then there is u G S such that x < u
and y < u. Thus f(u) is a common upper bound for f{x),f(y).
4
3. A subset S C A is (upward) compatible if for any x,y â‚¬ S there is u â‚¬ A such that
x < u and y < u. â€œCompatibleâ€ differs from â€œdirectedâ€ because for the former u 6 S,
while for the later we only require u G A. Similar arguments show that compat, which
selects compatible subsets is a subset system.
4. A subset of S C A is a chain if x, y â‚¬ S implies x < y or y < x. The rule ch which
selects all chains in a poset is a subset system.
5.We say a subset 5 C A is (upward) selfbounded if there is s 6 S such that for all
x â‚¬ S, x < s; a selfbounded set contains a maximum element. The rule sb, which
selects all selfbounded sets, is a subset system. Note that any order preserving map
preserves joins of upward selfbounded sets.
6.A nonexample: An antichain is a set of pairwise incomparable elements. The rule
ac which selects antichains is not a subset system, because there is an order preserving
surjection / : D â€”> 2, where D is the twopoint antichain and 2 is the two point chain.
7.Generating examples: Let Q be any class of posets closed under order preserving
surjections. One may define a subset system Zq by
Za(A) = {S C A : S e Q}.
This construction shows there is a great multitude of subset systems. The subset sysÂ¬
tem, compat, described above, is not generated this way, because one cannot determine
if S C A is compatible merely by looking at the poset S with its induced order.
We use subset systems, which we generically call j and m, to select which subsets
have joins and meets, respectively. Now we enumerate some categories of interest in this
discussion.
 the category of all (j, m)complete posets: that is, posets in which jsuprema and
TOinfima exist and are preserved by all maps.
5
DP^  the category of all completely distributive complete lattices.
DP^  the full subcategory of P^ containing objects which can be P^embedded in a
completely distributive complete lattice. DP^ is discussed in Section 5.3.
MÂ¿,  the full subcategory of P^ containing posets with an F structure, where F is
the monad defined in Section 5.3.
SpFJm  the full subcategory of PÂ£, containing spatial posets. See Chapter 7.
SJm  the category of generalized spaces with (j, m)complete families of distinguished
subsets: see Chapter 7.
FÂ£,  the full subcategory of SpTPJm containing spatial posets with flat spectrum: A is
defined to have flat spectrum if the maps A â€”+ 2 are trivially ordered: see Section 7.5.
1.3 Methods and Results
The discussion of DP^ and Mjn, which offer generalizations of distributive lattices,
uses the language of category theory. A quick introduction occurs in Chapter 2. The crucial
notion of a free object is formalized by monads, which are introduced in Chapter 3.
Chapter 4 describes a theory of subfunctors. The class of subfunctors of a functor
F bears a strong similarity to the power set lattice of a set X. Given a monad, with
functor part T, Meseguerâ€™s Lemmas 4.2.1, 4.2.2, 4.2.6, and 4.2.7, show that any subfunctor
F of T containing all constants has a â€œmonadic closure,â€ i.e., a smallest submonad F of T
which exceeds F. (Meseguerâ€™s Lemmas were formulated and proved by the author, but the
technique is similar to one in Meseguer [25].) Intuitively, TX is the full set of polynomials (in
the sense of universal algebra) with variables in X, FX is a natural subset of polynomials,
and FX is the smallest natural subset of polynomials which is closed under composition
and contains FX. An algebra structure for T is a way of evaluating â€œall polynomialsâ€; a
partial algebra structure for F is a way of â€œevaluating polynomials in F." Under suitable
conditions, partial algebras extend to Falgebras.
6
In Chapter 5, Meseguerâ€™s Lemmas are brought to bear upon monads for free complete
semilattices and free completely distributive complete lattices. Given any subset systems j
and m, there is a submonad F of the free completely distributive complete lattice monad.
The category of Falgebras offers a (somewhat mysterious) generalization of the category
of distributive lattices. The subcategory BP^, containing all (j, m)complete posets which
may be embedded in a completely distributive lattice is somewhat easier to understand and
still well behaved.
The existence of free objects in OP^ contrasts with the nonexistence of free objects
in PÂ£Â£ [9] and the category of complete Boolean algebras [8, 9], A fundamental difference
between these categories and DPÂ¿, is the requirement that joins and meets obey a distributive
law.
The power of category theory comes as much from what it ignores as what it examines.
Significant conclusions are often obtained without examining the â€œgrubby detailsâ€ of what
is going on. But this innocence of â€œgrubby detailsâ€ limits the scope of investigation. In
the case of this document, several nicely behaved categories  M?m, BP^,, and SpPJrn  are
introduced. For general subset systems j and m, the author does not even know if these
categories differ! The end of Section 5.3  from Corollary 5.3.10 onwards  describes most
of the authorâ€™s knowledge on the relationship between these categories.
Chapter 6 explores congruences, quotients and coequalizers in P, P3m and OPÂ¿,. Much
classical algebra (ring theory, lattice theory, group theory, etc.) is simplified by the fact that
any surjection is a regular epimorphism. For the categories introduced here, the situation is
not so simple. Example 6.2.5 shows that a jjoin preserving surjective image of a jcomplete
poset need not be jcomplete.
The results of Chapter 7 predate the other results presented here. Herrlich [12]
contains a detailed examination of reflections (and coreflections) in categories of topological
spaces. This dissertation aimed to generalize results summarized in Herrlich [12], for (j,m)
spaces. A (j, m)space consists of an underlying set and a family of â€œopenâ€ subsets which are
7
closed under junions and mintersections. Continuous maps of (j, m)spaces are functions
such that preimages of open sets are open. The initial aim was to find reflections and
coreflections of the category of (j, m)spaces (obeying a Tostyle separation axiom), and
study how the existence and properties of reflections and coreflections varied depending
upon the subset systems j and m. An obvious prerequisite to such a project is knowledge
of factorizations of continuous maps.
The chapter contains a description of (j, m)subspaces and (j, m)quotients. In addiÂ¬
tion, Section 7.5 describes a reflection of (j, m)spaces that corresponds to the 7) reflection
of topological spaces. Lastly, Section 7.6 describes epicomplete (j, m)spaces.
CHAPTER 2
PRIMER ON CATEGORIES AND POSETS
The text assumes a familiarity with the theory of sets typically used in mathematical
arguments. So familiar constructions  unions, intersections, cartesian products, quotients
by equivalence classes, functions, Zornâ€™s Lemma, and transfinite induction  are used without
further comment. (See Halmos [10] if this background is needed.) A basic familarity with
general topology is helpful.
Also some comfort with category theory is assumed. Namely, the reader can fill in
the blanks in the following informal definitions.
â€¢ A category A consists of a class of objects Obj(A) and maps Map(A), such that each
object has an identity map, and there is an associative notion of composition of maps.
â€¢ The set of Amaps from A\ to A2 is denoted A(Ai, A2).
â€¢ A functor F : A â€”> 23 assigns each A G Obj(A) an object F(A) â‚¬ Obj(23) and
each Amap / : A\ â€”Â» A2 a 23map Ff : F(A1) â€”> F(A2). The assignment respects
composition and identity arrows.
â€¢ If A is a category, Aop is the category with the same objects as A, but all arrows
reversed. For a category theoretic concept C, the dual is obtained by applying C to
AÂ°p.
â€¢ A contravariant functor A â€”+ 23 is a functor A â€”â–º Bop.
â€¢ Diagrams are used to display the behavior of a collection of maps; a diagram commutes
if any composites with the same domain and codomain are equal. For example, the
8
9
diagram
W
f
h
9
Y
*Z
commutes if and only if gf â€” ih.
Recall the following properties of functors:
Definition 2.0.1. Let F : A â€”> B be a functor. For each Ai, A2 E A, F gives a function
from the homset A(Ai, A2) into B(FAi, FA2) by
(/ : Ax Â» A2) ~ (Ff : Aj.  A2).
If, for each Ai and A2 this map is onto, then F is said to be full. If, for each A\ and A2 this
map is onetoone, then F is said to be faithful.
A full subcategory of a category A is a category â€˜B such that Obj(S) C Obj(A) and
all / : A â€”> B with A, B G Obj(B) are Bmaps. B C A is full if, and only if, the inclusion
functor is full.
Good general references for category theory are MacLane [21], Borceux [6], and Her
rlich and Strecker [13]. MacLane [21] gives a concise, high level summary of most category
theory and includes a chapter on monads. Herrlich and Strecker [13] is quite user friendly and
concretely describes many examples of adjoint functors. Borceux [6] covers a large amount
of material; the exposition is clear and very detailed.
2.1 Distinguished Maps
Definition and Remarks 2.1.1. Begin by defining a dual pair of concepts which coincide
with the notions â€œinjectiveâ€ and â€œsurjectiveâ€ in the category Set.
1. A map / : A\ â€”> A2 is epi, a.k.a epic (in noun form, an epimorphism) if
whenever g and h are maps A2 â€”> A3 such that gf = hf, then g = h.
10
2. A map / : A2 â€”> A3 is mono, a.k.a. monic (in noun form, a monomorphism)
if whenever g and h are maps A\ â€”> A2 and such that fg = fh, then g â€” h.
One may verify that a composition of epimorphisms (resp. monomorphisms) is epi (resp.
mono). Moreover, if f = ab is epi (resp. mono), then a is also epi (resp. b is also mono).
Definition 2.1.2. A map / : A â€”> B is an isomorphism, if there is g \ B â€”* A such that
id.4 = gf and idB = fg.
In most categories of â€œsets with structureâ€: a map is mono if and only if it is injective,
surjective maps are epi, but epimorphisms may not be surjective.
Example 2.1.3. Consider tfAb  the category of torsionfree abelian groups, i.e., abelian
groups such that
na = 0 => a = 0
for any natural number n and group element a, together with group homomorphisms. The
inclusion i : Z â€”Â» Q of the integers in the rational numbers is epi, but not onto.
In categories of â€œsets with relational structure,â€ bijective maps are not necessarily
isomorphisms.
Example 2.1.4. Consider Top  the category of topological spaces and continuous maps.
The identity function i : â€”> R from the reals (with discrete topology) to the reals (with
the usual topology), is a continuous bijection. However, the inverse function i_1 is not
continuous.
Example 2.1.5. Consider P  the category of partially ordered sets and order preserving
maps. Either bijection
two elements is order preserving. But the inverse function 4>~l is not order preserving.
For further discussion and more examples of epimorphisms and monomorphisms see
Herrlich and Strecker [13, Section 6] and Borceux [6, Volume 1, Sections 1.7 and 1.8].
11
Definition and Remarks 2.1.6. Consider a pair of maps f,g:A\â€”* A2. A map i : Aq â€”â– *
A\ rightidentifies f and g if fi = gi. A map i : Aq â€”> A\ is called the equalizer of / and g
if:
(eql) i rightidentifies / and g, and
(eq2) i has the feature that whenever j : B0 â€”> A\ rightidentifies / and g, there
is a unique map e : Bo â€”> Aq such that j â€” ie.
The definite article is used for equalizers, because (eq2) implies that if i : Ao â€”* A and
i' : A'0 â€”> A are equalizers for / and g, then there is an isomorphism j \ Aqâ€”* Aq such that
i' â€” ij. Notation: i = eq(/,
If there are / and g such that i : A0 â€”* A\ is the equalizer of / and g, then i is regular
mono. Regular monomorphisms are monomorphisms. If / is epi and regular mono, then / is
an isomorphism. For proofs of the assertions in this paragraph and a discussion of examples,
see Borceux [6, Volume 1, Section 2] or Herrlich and Strecker [13, Section 16].
The definitions of â€œcoequalizerâ€ and â€œregular epiâ€ are dual to â€œequalizerâ€ and â€œregular
mono,â€ but are repeated for emphasis, e : A2 â€”* A3 leftidentifies f and g if ef = eg. A
map e : A2 â€”* A3 is the coequalizer of / and g if
(coeql) e leftidentifies / and g, and
(coeq2) e has the feature that whenever d : Ai â€”Â»â€¢ B3 leftidentifies / and g, there is
a unique map c : A3 â€”Â» B3 such that e = cd.
Notation: e = coeq(/, g). If e is the coequalizer of some pair of maps, then e is called
regular epi The duals of all basic properties of regular monomorphisms hold for regular
epimorphisms.
Definition 2.1.7. An epimorphism / is extremal if whenever f = gh and g is mono, then g
is an isomorphism. Dually, a monomorphism / is extremal if whenever / = gh and h is epi,
then h is an isomorphism. For more detailed discussions, see Borceux [6, Volume 1, Section
4.3] and Herrlich and Strecker [13, Section 17].
12
Definition 2.1.8. The map / : A â€”â–º B is split mono if there exists g : B â€”> A such that
idÂ¿ = gf. The map g : B â€”* A is split epi if there exists / : A â€”> B such that id^ = gf. For
more information, see Herrlich and Strecker [13, Section 5]. Note that our terminology differs
slightly from the reference; â€œsectionâ€ and â€œsplit monoâ€ are synonyms, as are â€œretractionâ€
and â€œsplit epi.â€
Lemma 2.1.9. For A either â€œmonoâ€ or â€œepi,â€ consider the following statements.
1. f is splitA.
2. f is regularA.
3. f is extremalA.
4. f is A.
The implications 1 => 2 =4* 3 => 4 always hold. None of the converses generally hold.
For proof see: (1 => 2) Herrlich and Strecker [13, 16.15], (2 => 3) Herrlich and Strecker
[13, 17.11] or Borceux [6, Volume 1, 4.3.3(1)], (3 =Â» 4) holds by definition.
2.2 Bounds
Recall that a partial order on a set A is a relation < satisfying:
(pol) For all a 6 A, a < a.
(po2) Whenever a < b and b < c, a < c.
(po3) Whenever a < b and b < a, a = b.
A preorder is a relation that satisfies (pol) and (po2). If ^ is a preorder on A, define an
equivalence relation ~ on A by
a ~ b <=> a ^ b and b â– < a.
The relation is a partial order on A/ ~. A set A with a partial order (resp. preorder) is
called a partially ordered set (resp. preordered set).
13
Definition and Remarks 2.2.1. Let A be a preordered set. Define the up and down
closures of x 6 A by
 x = {a â‚¬ A : a ^ x}
and
^ x = {a â‚¬. A : x â– < a}.
More generally, if 5 C A define
I 5 â€” U{ i:iGS} = {aâ‚¬yl:3sGS,aXs}
and
 5 = U{f i :i Â£ 5} = {a G j4 : 3s 6 S', s ^ a}.
If 5 C A an upper (resp. lower) bound for 5 is a Â£ A such that for all s G 5, s â– < a (resp.
a â– < s). Use the following notation for the set of upper bounds of 5,
1(5) = n{f x : x â‚¬ 5} = {a G A : Vs 6 5, s â– < a}
and similar notation for the set of lower bounds
1(5) = n{j x : x â‚¬ 5} = (a G A : Vs G 5, a â– < s}.
If a G A, S C A and f a = 1(5), then a is a least upper bound, a.k.a. join, a.k.a.
supremum of 5. An equivalent way to say this is
a < x <=> V s â‚¬ 5, s < x.
If both a and a' are suprema of 5, then a,a' â‚¬ 1(5). Thus a ~ a'. In a partially ordered
set, the (unique) supremum of 5 is written \/ 5. If a â‚¬ A, S C A and j a â€” 1(5), then a is
a greatest lower bound, a.k.a. meet, a.k.a. infimum of 5. An equivalent way to say a is a
least upper bound is
x < a V s â‚¬ 5, x < s.
14
In a partially ordered set, the (unique) infimum of 5 is denoted /\ 5. (In using this notion
with a preordered set (A, ^), one refers to the associated poset A/ For example, \J S
then denotes the equivalence class containing all suprema of 5.) If a = V 5 or a = f\ 5, a is
an optimum bound for 5.
The following are elementary properties of the boundoperators and optimum bounds.
Lemma 2.2.2. For any preordered set A (in particular any poset), the following properties
hold. Use IB to denote either operator B or B.
1. For x,y â‚¬ A, x ^ y [ x C j y <==> ] y C] x.
2. For S,T C A, S CT => B(T) C B(5).
3. For SC A, 1(5) = I( 5), 1(5)  1(T 5).
4 For (5j)j6/ a family of subsets of A, B(u5Â¿) = nB(5Â¿).
5 If f : A â€”y B is order preserving, then ffE^S)) C Bs(/(5)).
Definition and Remarks 2.2.3. Let A and B be preordered sets. A Galois connection
between A and B is a pair of functions / : A â€”> B and g : B â€”> A such that
(gel) / and g are order reversing.
(gc2) For all a â‚¬ A and b G B, a ^ g(f{a)) and b X f(g(b)).
Below, basic properties of Galois connections are listed. Symmetry in the definition
allows symmetry in proofs. For any true statement about Galois connections, then the
statement obtained by switching the roles of / and g, along with A and B is also a true
statement about Galois connections.
Note that / and g respect ~. The notation / ~ g means â€œfor all a â‚¬ A, f(a) ~ g(b)."
1 â– 9  gfgâ– Suppose a Â£ A. By (gc2), a r< f(g{a)) and g(a) â– < g(f(g{a))). Using
a ^ f(g{a)) and (gel), g(f(g(a))) < g(a).
15
2. f(A/ ~) is dually order isomorphic to {6 G B : b ~ fg{b)}: This follows from the
preceding statement. Since / ~ fgf, any member f(a) G f(A/ ~) is equivalent to
f{g{f(a))). Moreover, if b G B and b ~ f(g(b)), then b G f(A/ ~) because b ~ f(g(b)).
3. a < g{b) b â– < /(a): if a ^ g(b), then b â– < f(g(b)) < f{a). The converse is proved
similarly.
4. For any S C A, Bf(S) = ^_1B(5). Calculate
x G B(/(5)) Vsg5,iS/(s)
<=^ Vs G S, s ^ 5(1)
<=* #)eI(S)
â€¢<=>â€¢ a; G p1Â®(5)
5. If A and B are posets, the previous item implies /\/(5) = /(V 5).
It is helpful to rephrase condition 3 for posets: ug(b) is the largest a with b < /(a).â€
In symbols,
g(b) = \/{a E A:b < /(a)}.
Basic information about Galois connections has been wellknown sincethe 1940s;
Raney [28] contains a bibliography of this early literature. The basic properties and defÂ¬
inition are listed in Herrlich and Strecker [13, Exercise 27Q]. The particular summary here
is by the author.
The concept of a Galois connection is symmetric, and allows one to transfer a great
deal of information between preordered sets. However, the fact that the functions involved
are order reversing is sometimes inconvenient. The concept of an adjoint connection between
posets is obtained by formally reversing one of the posets involved.
16
Definition and Remarks 2.2.4. Suppose A and B are preordered sets. A pair / : A â€”> B,
g : B â€”> A of order preserving functions is an adjoint connection between A and B if
(adl) For all a 6 A, a â– < g(f(a)).
(ad2) For all b e B, f(g(b)) < b.
Each basic property for Galois connections corresponds to a basic property of adjoint
connections. The basic properties of adjoint connections are listed below; proofs are omitted.
1 â€¢ /  fgf and g ~ gfg
2. f(A/ ~) = {b G B : b ~ g(f{b))} and g{B/ ~) = {a 6 A : a ~ f(g(a))}.
3. /(a) ^ b â– <=>â€¢ a ^ g(b)
4. For any SC A, Bg(S) â€” /_1B(5) and B/(S) = g_1B(5)
5. If A and B are posets, /(/\ S) = f\ f(S) and g(\J S) = \/ g(S).
Because of the asymmetry between / and g, and property 3, / is called the left adjoint
and g the right adjoint. Again there is an interpretation of 3 in words, â€œ/(a) is the smallest
b such that a â– < g{b)\ g(b) is the largest a such that /(a) ^ 6.â€
The definition and basic properties of adjoint connections are â€œfolklore.â€ For another
discussion of them see Johnstone [16, Chapter I, Paragraph 3].
Definition 2.2.5. A poset A is complete if each subset SC A (including the empty set)
has a supremum. Since the supremum of set of lower bounds for S is a lower bound for 5,
A is complete if and only if each subset S C A (including the empty set) has an infimum.
There is a criterion for testing when a given order preserving (resp. reversing) map
between posets is part of an adjoint (resp. Galois) connection.
Theorem 2.2.6. Adjoint Existence Suppose A and B are posets.
17
1. Suppose f : A â€”> B is order preserving and A is complete. Then f is a left adjoint if
and only if f(\J S) = \J f(S) for all S C A.
2. Suppose g : A â€”Â» B is order preserving and A is complete. Then g is a right adjoint if
and only if g(f\ S) = /\ g(S) for all S C A.
3. Suppose f : A â€”> B is order reversing and A is complete. Then f is part of a Galois
connection if and only if f(\J S) = f\ f(S) for all SC A.
This theorem is the poset version of the adjoint functor theorem. Folklore: see Johnstone
[16, Chapter I, Section 4, Paragraph 2] or Joyal and Tierney [17, Chapter 1, Section 1].
Proof. A proof for 1 follows; the other items are similar. Define g : B â€”> A by
9ÃP) = \J{aeA: f(a) < b}.
Since / preserves suprema, f(g(b)) < b. Thus, g(b) is the largest a such that /(a)
In adjoint (resp. Galois) connections, /(a) and g(b) can be defined as suprema or
intima. There is a sort of converse to this; one can view the supremum as an adjoint to a
particular map. See Lemma 5.1.1.
2.3 Natural Transformations
Maps compare objects in a category, functors compare categories, and natural transÂ¬
formations compare functors. This section contains no new results; results and some expoÂ¬
sition are paraphrased from Herrlich and Strecker [13, Section 13].
Definition 2.3.1. Let F, G : A â€”> 3 be functors. A natural transformation a : F â€”> G is
a rule which assigns a map olA : FA â€”> GA to each A E A such that if / : A â€”> B is an
Amap the following diagram commutes.
Ff
FA â€”>FB
r\A
V B
GA
Gf
GB
18
Construction 2.3.2. Let F, G, H : X â€”Â» A be functors, a : F â€”> G and f3 : G â€”> H natural
transformations. (The situation is drawn in the diagram below.)
F
G
Q
H
' J
0
Then the assignment (/3a)A = (/3A)(aA) is a natural transformation.
Proof. Let / : A â€”> B be a map. Since a and (3 are natural, each square below commutes.
FA
Ff
â– *FB
GA
Gf
GB
HA
as
HB
Therefore, the outside rectangle commutes; hence, (3a is natural. â–¡
Call /3a the vertical composition of a and (3.
Construction 2.3.3. If F : A â€”> B and G,H : 3 â€”> C are functors and a : G â€”> H is a
natural transformation, then (aF)A := a(FA) is a natural transformation aF : GF â€”> HF.
Construction 2.3.4. If F : â€œB â€”> C and G,H : A â€”â–º B are functors and a : G â€”> H is a
natural transformation, then (Fa)A := F(aA) is a natural transformation Fa : FG â€”> FH.
Construction 2.3.5. Suppose F, G : A â€”> Ãœ3 and H,J:T>â€”*G are functors, and a : F â€”>
G, (3 : H â€”> J are natural transformations. (The situation is drawn below.)
F H
0 e
19
Then for each object A â‚¬ Obj(.A) the following square commutes.
FHA F0A > FJA
aHA
aJA
GHAâ€”G0A >GJA
The assignment (/3 â– a)A := (aJA)(F/3A) = (G(3A)(aHA) is a natural transformation.
Proof. The square commutes because a is a natural transformation; to see this, one applies
the natural property at the map (3A : HA â€”+ JA. To prove the natural property of Pa,
we use the squares which define Pa A and 0aB. Compare the corners of the squares using
maps obtained from / by application of the functors FH, GH, FJ, and GJ. The resulting
commutative cube shows Pa is a natural transformation. â–¡
Call P â€¢ a the horizontal composition of a and p.
Definition 2.3.6. A natural equivalence a : F â€”* G is a natural transformation such that
each component aA is an isomorphism. Given functors F and G are naturally equivalent if
there is a natural equivalence a : F â€”> G.
Categories A and B are equivalent if there are functors F : A â€”> B and G : B â€”> A
such that FG is naturally equivalent to ids and GF is naturally equivalent to id^.
A and B are dual if there exist contravariant functors F : A â€”> B and G : B â€”> A
such that FG is naturally equivalent to ids and GF is naturally equivalent to id^.
2.4 (Co)Limits
For intuition, it is useful to view the objects of a category as a preordered class, with
the preorder A < B if and only if there is a map f : A â€”* B. Categories are more complex
than preordered classes, because there may be many maps / which manifest A < B.
Definition and Remarks 2.4.1. Let A be a category. A diagram D in A (a.k.a. a small
subcategory of A) is a set of objects Obj(Â£>) C Obj(A) and maps Map(D) C Map(A)
20
between them; for technical reasons one requires that if f,g G Map(D) and fg is defined
then fg G Map(D), and that for all A G Obj(D), idA â‚¬ Map(D). This definition of a diagram
is equivalent to, but differs from the one in the majority of the literature; see MacLane [21],
Borceux [6], or Herrlich and Strecker [13] for the standard definition.
Suppose D is a diagram. A source (S, {s^ : S â€”Â» A : A G Obj(Â£))}) for D consists of
S G Obj(A) and maps sA : S > A such that if / : A > A' is a map in D, fsA = sA>. A source
for D is a â€œlower boundâ€ compatible with all maps in D. A source (L, {tA : A G Obj(Z))})
for D is the limit of D if whenever (5, : A G Obj(D)}) is a source, there is a unique map
c : S â€”â–º L such that for each A G Obj(D), sA â€” iAc.
Dually, define a sink, or cosource for D to be an object S together with maps iA :
A â€”> S (for A G Obj(D)) such that for each f : Aâ€”> A' in Map(D), iA = iA>f.
A colimit (C, {Â¡jla : A G Obj(D)}) for D is a sink for D, such that if (5, {iA : A G
Obj(D)}) is any sink there is a unique map c : C â€”> S such that for each A G Obj(D),
Ha = ciA.
Note that limits and colimits of D are unique up to compatible isomorphism. To
prove this (for limits), suppose (L,Â£A) and (L',Â£'A) are both limits for D. The limit property
guarantees that there are unique compatible maps c : L â€”> L' and d : L' â€”> L. But
cd : L' â€”> L' and dc : L > L are both compatible maps. By uniqueness, cd = id'Â£ and
dc = idL. From now on, we shall use the definite article when writing about limits and
colimits.
Example 2.4.2. Let us consider Set. If X is a set of sets, form a diagram containing
all members of X and no functions. The limit for this diagram is the Cartesian product
(nx,7rx).
Recall that Â¡Q X contains all functions / : X â€”* (J X,with the feature that f(X) G X
for all X G X. Such / are called choice functions, because they choose one member of each
X G X. The functions nx 'WX X are defined by nx(f) = f(X).
21
If there is a source (5, {s* : S > X : X â‚¬ X}), there is a unique function c : S Â» f] ^
making
commute for each X. It is defined by c(a)(X) â€” sx(a).
Motivated by this example, one defines the categorytheoretic product of a set X of
objects (in any category) as the limit of the diagram containing all members of X and no
maps. In most categories of â€œsets with structureâ€ products (exist and) look like products in
Set, with suitable structure added.
In Set, the colimit of the diagram containing all members of X and no functions,
is the disjoint union ]J X, with inclusion maps px 'â– X â€”iâ–º JJ X. The colimit property is
satisfied by (JJ X, px) because if there is a sink (C, ix), the function c : JJ X â€”â–º C  defined
by c(x) = ix{x) for the unique X â‚¬ X containing x  is the unique compatible map.
In categories other than Set, coproducts are defined identically. Usually the coprodÂ¬
uct of a set X of ^.objects is â€œthe ^1object freely generated by JJ X.â€
Example 2.4.3. The [co]equalizer of /, g : A â€”> B is the [co]limit of the diagram containing
objects A and B along with maps / and g. The notation for equalizers is customarily
simplified by omitting the source map to B. Previously, i : E â€”> A was defined to be the
equalizer if fi = gi and i factors through any other map which right identifies / and g. i
is the source map to A. The source map to B is redundant: it must be fi = gi : E â€”* B.
Similar notational economy is applied to coequalizers.
22
Example 2.4.4. Let us consider the diagram
A
f
B  *C.
(The diagram also contains identities for all objects, but for brevity these are omitted.) The
limit (B Xc A, nA, ttb) of this diagram is called the pullback of / along g, or the pullback of g
along /. To be explicit, fnA = 7rBg and if (Q,qA : Q â€”* A,qs â– Q â€”â–º B) satisfies fqA = qsg,
then there is a unique map i : Q â€”> B Xp A.
In Set,
B xc A = {(b, c) e B x A: g(b) = /(c)},
and the projection maps n are the restrictions of the projections from the cartesian product.
Pullbacks in any concrete category A  equipped with a limit preserving faithful functor
U : A â€”* Set  are computed identically. Two particular instances of pullbacks deserve
special attention.
First, let / : A â€”> C be any map and g : S â€”> C be a subset inclusion of S C C.
Note that (s, a) â‚¬ S xc A if and only if g(s) = /(a); suppressing mention of g, this reads
S xc A â€” {(s,6) : f(b) = s}. Thus, this pullback is canonically isomorphic to the preimage
of S under /. This example partially motivates the name â€œpullback.â€
Second, let / : A â€”> B be any map and consider the pullback of the diagram,
A
f
A *B
i.e., the pullback of / along itself. Using the computation for pullbacks in Set, given above,
A xB A = {(a, a') e A x A : f(a) = /(a')}.
23
This relation on A is often called the kernel of /. Thus, one calls (Axb A, nie/t, irnght) the
kernel pair of /. More indepth discussion of pullbacks is given in Borceux [6, Volume 1,
Section 2.5] and Herrlich and Strecker [13, Section 21].
Definition and Remarks 2.4.5. A category is said to be complete if each diagram has a
limit; it is said to be cocomplete if each diagram has a colimit.
Unlike the situation for posets, a category may be complete without being cocomplete
and vice versa. (See Herrlich and Strecher [13, Section 23, pl61ff] for a detailed discussion
of this and related issues.)
Consider a functor F : A â€”> B. If D is a diagram in A, there is a diagram FD with
Obj (FD) = {FA : A e Obj(D)} and Map (FD) = {Ff : f 6 Map(D)}.
Let D be a diagram in A. A functor F : A â€”> B preserves the limit of D, if whenever
the limit (L,Â£a) exists in A, (DL, D(Â¿a)) is the limit of FD. F preserves limits if for any
diagram D, F preserves the limit of D.
If G : A â€”Â» B is contravariant, D is a diagram in A, and (L, {Â£a â– A â‚¬ Obj(D)}) is
the limit of D, then (GL, {GiA : GA â€”> GL}) is a sink for GD. If (GL, {GÂ¿a 'â– GA â€”â–º GL})
is the colimit for GD, then G takes limits to colimits. Analogously, may G take colimits to
limits.
2.5 Adjoint Functors
There are several useful concepts of adjoint connections between categories. There is
a strong analogy between preordered sets and categories; any category may be preordered
by
A^B <=> 3/ : A > B.
Categories are more complex, because many maps could manifest A A B. We begin with
the concept analogous to 2.2.3.
Definition and Remarks 2.5.1. A (functorial) Galois connection between categories A
and 'B consists of contravariant functors F : A â€”> â€œB, G : B A, together with natural
24
transformations
77 : id* â€”Â» GF
and
e:W3 â€”> FG
such that (Fr]A)(eFA) = id^ and (GeB)(r]GB) = ides for each A 6 Obj(A) and B â‚¬
Obj(33); these equations are the socalled triangle identities. (Alternate terminology: if
(F,G,rj,e) is a functorial Galois connection, then F and G are adjoint on the right.)
1. The categories fix(rÂ¡) and fix(e)  containing all objects such that r]A (resp eB) is an
isomorphism  are dual. [3, Section 4, Lemma 1]
2. There is a natural bijection A{A, GB) â€”* Â®(5, FA) given by
(f : A â€”> GB) H> F(f)(eB) : B â€”> FA.
The inverse map is
(g : B â€”> FA) >> G(f)(vA) : A > GB.
There are two (identical) calculations required to check that the functions are mutually
inverse. One is summarized by the diagram below.
GFA â€”(GFf) > GFGB â€”â€” * GB
The square commutes because 77 is natural. The triangle commutes because of the
identity ides = (GeB)(rÂ¡GB). The reader may formulate and check what is meant by
â€œnaturalityâ€ of the bijection.
3.F and G both take colimits to limits.
25
4. Each rjA has the following universal property: if / : A â€”â–º GB, there is a unique map
/ = F(f)(eB) : B â€”â–º EM which makes the following diagram commute:
A â€”â€”*â– GFA
Each eB has the analogous universal property.
Much of the literature just deals with adjoint connections, where both functors are covariant.
By duality, any such result can be translated in terms of Galois connections. Banaschewski
and Bruns [3] includes a reasonably thorough expository section on functors which are adjoint
on the right.
The â€œ(functorial) Galois connectionâ€ concept is symmetric, but the contravariance of
the functors involved is sometimes awkward. An analogous asymmetric concept, with the
functors both covariant is described below.
Definition and Remarks 2.5.2. A (functorial) adjoint connection between categories A
and 3 consists of functors F : A â€”> 3 and G : 3 â€”â–º A and natural transformations
77 : idyi â€”> GF and e : FG â€”> idÂ® such that (GeB)(rÂ¡GB) = id^s and (eFA)(FrÂ¡A) = id^;
these equations are the socalled triangle identities. In this situation, one also says â€œF and
G are adjoint functors,â€ â€œF is the left adjointâ€ and â€œG is the right adjoint.â€ The basic
properties of functorial adjoint connections closely correspond to the basic properties of
functorial Galois connections.
1. The categories fix(r/) and fix(e)  containing all objects such that 7/.4 (resp eB) is an
isomorphism  are equivalent.
2. There is a natural bijection A(A, GB) â€”> 3(FA, B) given by
(f : Aâ€”* GB) t> (eB)F(f) : FA â€”> B.
26
The inverse map is
(g:B+FA)~ G(f)(rÂ¡A) : A â€”> GB.
3. F preserves colimits; G preserves limits. Borceux [6, Volume 1, 3.2.2]
4. Each rjA has the following universal property: if / : A â€”> GB, there is a unique map
/ = (eB)F(f) : FA â€”> B which makes the following diagram commute:
A â€”â€”>GFA
Each eB has the analogous universal property: if / : FA â€”â–º B, there is a unique
/ = G(f)(r)A) : A â€”* GB such that
commutes. Any functor for which there is a natural transformation q with the above
universal property is part of an adjoint connection.
For discussions of adjoint functors, see MacLane [21, Chapter IV], Herrlich and Strecker [13,
Sections 26, 27, 28], or Borceux [6, Volume 1, Chapter 3].
There is a criterion for determining when functors are adjoints, which corresponds to
Theorem 2.2.6.
Theorem 2.5.3. (Adjoint Functor Theorem, [21, V.6.2],) Let A be a complete category.
A functor G : A â€”> B has a left adjoint if and only if
1. G preserves limits, and
27
2. (solution set condition) for each B â‚¬ 3 there is a set I and an Iindexed family of
maps fi'.Bâ€”> GAi such that any map f : B â€”> GA can be written as h = (Gt)fi for
some i 6 / and t : AÂ¡ â€”> A.
CHAPTER 3
ALGEBRAS OF A MONAD
Monads (a.k.a. triples, a.k.a. standard constructions) and their (EilenbergMoore)
algebras provide a concise formulation of many important categorical aspects of universal
algebra. The category of algebras for a monad has special properties, which are summarized
in Section 3.1; a category of algebras is often complete and cocomplete, and always has
â€œfree objects.â€ Each (functorial) adjoint connection induces a monad; the correspondence
between adjunctions and monads is discussed in Section 3.2. In Section 3.3, the question of
when an adjoint connection connects A to a category of algebras is addressed. In Section
3.4, the question â€œwhen does the composite of two monadic adjunctions yield a monadic
adjunction?â€ is discussed.
The results in this chapter are reasonably well known. MacLane [21, Chapter VI],
Borceux [6, Volume 2, Chapter 4], Barr and Wells [4, Chapters 3 and 9], Manes [23] and
the introduction to the seminar notes [1] contain good expositions of monads from various
perspectives.
3.1 Categories of Algebras
Definition 3.1.1. A monad T = (T, rj, /i) on A consists of a functor
T:A^A,
a natural transformation
V : idA > T,
and a natural transformation
/i: T2 â€”> T,
28
29
such that the following identities (expressed by commutative diagrams) hold: the unit laws
(In these diagrams, Tn denotes the nfold composite of T with itself.)
Intuitively, TA is the free object on A; r]A is the â€œinsertion of variablesâ€ map; fjA
is the â€œsemantic composition,â€ i.e., a map which allows one to view a polynomial with
polynomial variables as a polynomial. See Example 3.2.2 for a concrete illustration of the
roles of T, 77 and Â¡1.
Definition and Remarks 3.1.2. Let T be a monad on A. A Talgebra (A, a) consists of
A â‚¬ Obj(A) and a : TA â€”> A (the socalled structure map) such that a(rjA) â€” kU (unit
law) and
T2A >TA
M a
TA
30
(associative law) commutes. A homomorphism / : (A, a) â€”> (B, b) of Talgebras is an Amap
/ such that
Tf
TA â€”>TB
a b
commutes. The category of all Talgebras and Talgebra homomorphisms is denoted A1.
There is a forgetful functor UT : Ar â€”â™¦ A; it is defined by UT(A, a) = A and
UT(f) = f. UT has a left adjoint FT : A â€”â–º AT defined by
Ft{A) = (TA,iiA)
and
FT(f) = Tf.
The associated natural transformations are
7?r = 77: kU  UtFt = T
and
eT : FtUt â€”> id^r : eT(A, a) a.
[21, VI.2.Theorem 1]
Limits in Ar are â€œcomputed in Aâ€ in the following sense.
Proposition 3.1.3. [4, 3.3.4], [6, Volume 2, 4.3.1] Suppose (T,rj,p) is a monad on A. If D
is a diagram in AT such that UT(D) has a limit (L,pu(A)), then there is a unique structure
map I :TL â€”> L such that each pu(A) a Talgebra homomorphism.
Proof. Let D be a diagram in AT, such that UTD has a limit (L, IA). For each A G Obj(D),
name the structure map sÂ¿ : TA A. The goal is to produce a structure map s : TL â€”> L
such that each : L â€”> A is a Talgebra map.
31
The requirement that each is Talgebra map amounts to: for each A, the diagram
below commutes.
FL
SA
Since (L,Â£a) is a source for UD, (FL, sa(FIa)) is a source for UD. Thus, there is a unique
A map s : FL â€”> L making each diagram above commute.
Checking the unit law, s(iL) = idÂ¿: the diagram
commutes using the definition of s, and that i is a natural transformation. Thus, for each
A, Â£a â€” Â£as(ÃL). Prom the uniqueness of the map from the limit of D to any other source
for D, it follows that id/, = s(iL).
A similar â€œcomparision of squares diagramâ€ can be used to verify that the algebra
associative law holds. â–¡
Corollary 3.1.4. If A is complete and T is a monad on A, then AT is complete.
If A is cocomplete, AT is often also cocomplete. The following theorem was originally
proved with fewer hypotheses in Linton [20], Other expositions are given in Borceux [6,
Volume 2, 4.3.4] and Barr and Wells [4, Section 9.3].
Theorem 3.1.5. Let A be cocomplete and T be a monad on T. AT is cocomplete if and
only if AJ has coequalizers.
32
3.2 Adjoint Connections induce Monads
The following proposition gives a correspondence between monads and adjoint conÂ¬
nections. It should be emphasized that the correspondence is not bijective. Each monad
gives rise to a unique adjoint connection, but in general many adjunctions induce the same
monad.
Proposition 3.2.1. Correspondence between monads and adjoint connections
1. Let F : A â€”> 3, G : "B â€”> A, rÂ¡ : kU â€”> GF, e : FG â€”> idÂ® be an adjoint connection.
Then (GF,rÂ¡,GeF) is a monad on A.
2. If T = (T, T], /i) is a monad, then
GtFt = T,
T
V =v,
and
GTeTFT = fi.
The proof of the preceding Proposition consists of verifying identities: the triangle
identities for adjoint connections imply the unit laws; the associative law holds because
the square defining horizontal compositions commutes. Details are given in MacLane [21,
VI. 2. Theorem 1],
This correspondence allows construction of many examples of monads. Often, but not
always, â€œa naturally occurringâ€ adjoint connection corresponds to the category of algebras
over the induced monads.
Example 3.2.2. Consider the category Grp of groups. The forgetful functor Ucrp â– Grp â€”Â»
Set and the free group functor FgtP : Grp â€”Â» Set form an adjoint connection between Grp
and Set. Recall that ForpiX) is the set of all reduced words
33
where xÂ¿ 6 X, sÂ¿ 6 Z; a word is reduced if for all i with 2 < i < n, x,_i / xÂ¿. The operation
on Fcrp(X) is concatenation of words, followed by reduction. T = Ucrp^Grp is the functor
part of the induced monad; 77 : idset â€”> T is the natural transformation whose component at
X sends x â‚¬ X to word x1; if wâ€œl â€¢ â€¢ â€¢ iuâ€œn 6 T2(X) and for each i, = x^1 â€¢ â€¢ â€¢ then
â€¢â€¢â€¢<â– ) = (*5y^a{rtÃ³Ã*Ã3!3r.
6l.m(U'l
Each group G is an Talgebra; the group multiplication and inversion give a map from the free
group on the underlying set of G to G. Conversely, each Talgebra structure gives a group
multiplication and inversion. Grp is equivalent to the category of Setr. (This example is
typical in the sense that any category of finitary algebras  in the sense of universal algebra
 is also a category of monad algebras. The parts T, 77, and /r of a monad generally have the
same roles as in this example.)
Example 3.2.3. Some (nontrivial) adjunctions involving Set induce the trivial monad.
For example, consider the category Top of topological spaces and continuous maps. The
forgetful functor
U'Xop â– Top â€”* Set
has a left adjoint
FtbP : Set >â€¢ Top,
which sends a set X to the discrete topological space with underlying set X. One easily
checks that T = GTopFTop is the identity functor on Set, and that all associated natural
transformations are identities.
Definition 3.2.4. Suppose F : A â€”G:Bâ€”>Fl, 77: id^ 7 GF, and e : FG â€”> idÂ®
forms an adjoint connection. By Proposition 3.2.1, the adjoint connection induces a monad
(T = GF, 77, GfiF). If B is equivalent to AT, then we say the adjoint connection (sometimes
just the right adjoint G) is monadic. Often a concrete category B has a â€œcanonicalâ€ forgetful
functor G : B â€”> A] in this case, one may even say B is monadic over A  omitting mention
of G.
34
These examples illustrate the qualitatively different behaviors of monadic and non
monadic adjoint connections. Monadic categories are determined by the combinatorial strucÂ¬
ture of the free algebra functor FT. Most â€œrelationalâ€ categories  like Top, P, the category
of graphs, etc.  have free functors, but these free functors do not add any structure to the
underlying set; they merely attach the â€œmost discreteâ€ possible relation to the given set.
3.3 Detecting Categories of Algebras
Because of the special properties of AT, the question of when an adjoint connection
is monadic has great practical importance.
Remark 3.3.1. Let (A, a) be a Talgebra. Because of the associative and unit laws, and
because 77 is natural, the following equations hold:
(nat) (Ta){r[TA) â€” (rp4)a,
(unitl) (fiA)(rjTA) = idTA,
(unit2) a(r)A) = idÂ¿, and
(assoc) a(Ta) = a(fiA).
These equations imply that a â€” coeq(Â¡iA,Ta)\ for if b : TA â€”â–º B rightidentifies Â¡1A and Ta,
then b(r]A) : A â€”> B and
b(rÂ¡A)a = b(Ta){T]TA) by (nat)
= b(fiA)(r]TA) since b right identifies
= b by (unitl).
One verifies that / = b{qA) the unique map / : A â€”> B with fa = for if fa = b, then
/ = fa(rjA) = b{rÂ¡A).
These equations give a great deal of information about a. Thus, the key hypothesis
in Beckâ€™s theorem  the criterion for when an adjunction is monadic  is the preservation
and reflection coequalizers obeying the equations described above.
Using the above described equations for motivation we offer the following definitions.
35
Definition and Remarks 3.3.2. Consider maps f,g:Aâ€”>B.
1. If e : B â€”Â» C has the feature that for any functor F, Fe = coeq(Ff, Fg) then we say e
is an absolute coequalizer.
2. If there are maps e : B â€”> C, sc : C â€”> B and sb : B â€”> A such that
fsB
id/i = gsB,
idc = esc,
ef = eg,
then we say e is a split coequalizer.
Note that for every Talgebra {A, a), a is a split coequalizer of Ta and pA. It is also easy to
see that every split coequalizer is absolute.
The following theorem is due to Beck (unpublished). Linton [19] contains a detailed
discussion of variations on the theorem. The theorem was originally phrased in terms of
split coequalizers only; Pare [26] refined the theorem to include the â€œabsolute coequalizerâ€
condition. Many variations on the hypotheses exist; the version stated here is found in
MacLane [21, VI.7.1], Other expositions of the theorem may be found in Barr and Wells [4,
Section 3.3] and Borceux [6, Volume 2, Section 4.4].
Theorem 3.3.3. Let (F : A â€”> 25, G : 3 â€”> A, r?, e) be an adjoint connection, and T = GF,
rj â€” q, and p = GeF be the associated monad. The following conditions are equivalent:
1. The adjunction (F,G,r],e) is monadic.
2. If f,g : A â€”â–º B 6 Map(!B) and the pair (Gf,Gg) has an absolute coequalizer e!, then
e = coeq(f,g) exists and Ge â€” e'.
36
3. If /, g : A â€”* B â‚¬ Map(!B) and the pair (Gf,Gg) has a split coequalizer e', then
e = coeq(f,g) exists and Ge = e'.
Recognizing monadic adjunctions is complex because compositions of monadic funcÂ¬
tors are not generally monadic.
Example 3.3.4. The category Ab  of abelian groups  is monadic over Set, and tfAb (see
Example 2.1.3) is monadic over Ab, because any reflection is monadic. Each free abelian
group is torsion free, so the monad on Set induced by the free abelian group functor is the
same as the monad induced by the â€œfree torsion free abelian group functor.â€ Thus, if tfAb
were monadic over set, then tfAb and Ab would be equivalent categories  both equivalent
to a suitable category of monad algebras. The categories tfAb and Ab are obviously not
equivalent. (The preceding example is summarized from Borceux [6, Volume 2, Example
4.6.4].)
If G : A â€”* Set, the hypotheses of Beckâ€™s theorem can be reformulated to make it
easier to check whether G is monadic.
Theorem 3.3.5. [11, Theorem 4.2] Let G : A â€”> Set be a functor with a left adjoint.
Suppose that A is complete and has coequalizers. The following are equivalent:
1. G is monadic.
2. G satisfies the following conditions
â€¢ G preserves and reflects regular epimorphisms;
â€¢ if f : GA â€”> X is an isomorphism, then there is a unique map g : A â€”> B such
that Gg â€” f;
â€¢ G reflects kernel pairs.
The following lemma provides useful information concerning when a functor between
two categories of algebras is monadic.
37
Lemma 3.3.6. [6, Volume 2, Corollary 4.5.7] Suppose U : 23 â€”> A, V : G â€”> A, and
Q : 23 â€”> C are functors. IfU â€” VQ, U and V are monadic, and 23 has coequalizers, then Q
is monadic. In particular, Q has a left adjoint.
3.4 Distributive Laws
Compositions of monadic adjunctions are not generally monadic; when a composition
of monadic adjunctions is monadic, it indicates a distributive law between the two structures.
The following section summarizes the results later needed from Beckâ€™s [5]; this exposition
follows Beckâ€™s notation, except that composition of functions here will read righttoleft.
For the duration of the section, assume there are two monads T = (T, t]t, pT) and S =
(S, rjs, ps) over some base category A. For any monad, we use F and U, with superscripts for
the name of the monad, to denote the free algebra functor and forgetful functor associated
with a monad, respectively. See 3.1.2 for the definitions of F and U.
This material is rather abstract, so it helps to have an example in mind: after each
definition and theorem, we will illustrate what it means using S  the free monoid monad
over Set  and T  the free abelian group monad over Set; the composite monoid TS gives
the free ring. For a set X, SX consists of all â€œstringsâ€ from X, with concatenation as the
operation: 5 acts on functions by
Sf(xxx2 â€¢ â€¢ â€¢ xn) = f(x1)f(x2) â– â€¢ â€¢ f(xn).
TX consists of all formal (finite) linear combinations of elements of X, with integer coeffiÂ¬
cients. T acts on functions in the expected way.
Despite the concrete illustrations in terms of these monads, all theorems and defintions
will apply to any monads S and T.
Definition and Remarks 3.4.1. A distributive law of S over T is a natural transformation
Â£ : ST>TS
38
satisfying the compatibility conditions:
T S
If there is a distributive law Â£ : ST â€”â–º TS of S over T, then the composite monad is TS =
(TS,r]TS := r]TT]s,pTS := pT ps(TÂ£S)). In the definitions of rÂ¡TS and pTS, juxtaposition
denotes horizontal composition of natural transformations. The verification that TS actually
defines a monad is omitted; for more information see Beck [5],
In the case when â€œS=free monoidâ€ and â€œT=free abelian group,â€ the natural transÂ¬
formation i : ST â€”> TS expresses a productofsums as a sumofproducts in the usual
way:
n 5^ aitkÂ»+ n ac(k),k
keK iei c keK
where c ranges over all choice functions c : K â€”> I. It is a somewhat enlightening exercise
to check that Ã  defined this way  is a natural transformation ST â€”â–º TS.
A distributive law may of T over S also be viewed as a â€œway of lifting T to a monad
over S.
39
Definition 3.4.2. T has a lifting into ./Is, if there is a monad T such that TUS = UST,
UsrjT â€” r]TUs and Usp^ = pTUs. To sketch the situation, T is a lifting of T onto if
commutes.
Theorem 3.4.3. There is a bijective correspondence between liftings of T to /Is and disÂ¬
tributive laws of S over T.
The proof is outlined in Beck [5]. The correspondence is defined as follows: if l is a
distributive law, and (A, a) is an 5algebra, then T is defined as follows:
T(A,a) = (TA,(Ta)(Â£A)),
rjT(A, a) = t]TA : (A, a) > T(A, a),
and
fiT(A, a) = pTA : TT(A, a) Â» T(A, a).
Consider the following diagram:
the square commutes because rÂ¡T is a natural transformation; the triangle commutes because
l is compatible with rf. Since the perimeter of the diagram commutes, rjTA is an 5algebra
40
map, A similar diagram, which uses the compatibility between l and pT, shows pF is an
5algebra map. Since the underlying maps are defined in A and obey the monad laws, tjt
and pT also obey the monad laws. If T is a lifting of T over As, i is defined to be the
following composition:
ST STS = STUsFs = USFSUSTFS usflfs USTFS = ST.
(In the above equation, es is the counit of the adjunction (Fs, Us, r/5, es); es(A. a) = a.)
After some detailed computation, one verfies that this is a distributive law, and that the
correspondences described are mutually inverse.
Corollary 3.4.4. If T has a lifting to As, then there is a composite monad TS.
Theorem 3.4.5. Suppose I: ST â€”> TS is a distributive law. The categories ATS and (AS)T
are equivalent.
Let us consider this in more detail. An object in (AS)T consists of an Aobject,
A, along with an 5structure as : SA â€”> A, Tstructure ar : TA â€”+ A, and 5structure
t: ST A â€”> A for TA such that the following diagram commutes.
SciT
ST A 1â€” SA
t aS
TA â€” >A
Given a T5algbra {A, a), A has a Tstructure aT map defined by
aT = a(Tr]SA) :TA>A
and an 5structure map as defined by
as = a(r]TSA) : SA â€”> A.
41
The following diagram  which expresses the distributivity of the structure maps  commutes.
Thus we have a map ATS â€”> (AS)T. given by,
(A, a) e ATS i * (A, ar, as,Tas(Â£A))
which is functorial, because both rj1' and rf are natural.
The inverse functor (AS)T â€”> ATS, maps (A,ar,as,t) to asS(ar) â€” art. The verifiÂ¬
cations for these assertions is given in Beck [5],
CHAPTER 4
GENERATING SUBMONADS
In this chapter, a technique for creating monads is laid out; before proceeding formally,
let us consider a rough outline of the technique. Suppose (T, r], /r) is a monad on a reasonable
category, and that / : F â€”* T is a subfunctor of T. The goal is to extend F to a submonad
of T.
Intuitively speaking, F is a natural collection of polynomials. One should therefore
require that for all A, (r/A)(A) C F(A)  i.e., that for each A and a G A, FA contains the
â€œconstant polynomial with value aâ€. F is not generally â€œclosed under compositionâ€; i.e., it
is not generally true that (nA)(F2A) C A. To correct this problem, we start with F and
iteratively add polynomials obtained by composing members of FA.
The purpose of the chapter is to formalize the preceding vague outline. The author
abstracted and clarified [25, Proposition 3.4], which may be viewed as a special case of this
result. Thus, the main results of this chapter are called Meseguerâ€™s Lemmas  4.2.1, 4.2.2,
4.2.6, and 4.2.7  to acknowledge the analogy which prompted the technique. The author
believes the general formulation of the technique is new.
Section 4.1 details requirements on a â€œreasonable categoryâ€, explains precisely what
is meant by â€œsubfunctorâ€, and gives methods for constructing subfunctors. Section 4.2 gives
a proof of Meseguerâ€™s Lemmas. Section 4.3 describes an example showing the necessity of a
technical hypothesis of the lemmas.
4.1 Subfunctors
This section describes a general theory of subfunctors. A subfunctor of F : X â€”> Set
is a natural transformation r] : E â€”+ F such that each component is a subset inclusion. The
42
43
concept â€œsubfunctorâ€ is quite useful, but in some categories such as P, Top, and Locale
(explained in Example 4.1.2), there is either
â€¢ no obvious meaning for subset inclusion (in Locale), or
â€¢ more than one possible structure on each subset (in P and Top).
The following section expounds a theory in which â€œsubfunctorâ€ means â€œnatural transformaÂ¬
tion whose components are all extremal monoâ€.
Axioms 4.1.1. Throughout, the base category A is assumed to have the following features:
1. A is complete.
2. A is an (epi, extremal mono)  category. Recall that an (epi, extremal mono)  category
is a category with the features that
â€¢ For each map /, it is possible to factor / as / = me where e is epi and m is
extremal mono. This factorization is unique, in the sense that if / = m'e' is
another way of writing / as an epi followed by an extremal mono, there is an
isomorphism i such that
commutes.
â€¢ if a and b are extremal mono and ab is defined, then ab is extremal mono.
3.A is extremallywellpowered. This means: for each A, there is a set
{fx 'â– S\+ A}
of extremal monomorphisms, such that if m : T â€”â–º A is any extremal monomorphism,
there is an index A and isomorphism i such that f\ â€” mi.
44
If A has these features, then it is an SFcategory. (â€œSFâ€ stands for â€œsubfunctorâ€.)
Example 4.1.1. By Herrlich and Strecker [13, 34.5], any wellpowered category with inÂ¬
tersections and equalizers is an (epi, extremal mono)  category. So if A is complete and
wellpowered then A is an SFcategory.
It follows that P, Set, Top and practically any reasonable category of topological
spaces, is an SFcategory.
Example 4.1.2. The category of frames  i.e., complete lattices in which
a a\J S = ^ {a A s \ s E S}
holds for all elements a and subsets S  is complete and cocomplete. This category also has
(regular epi, mono) factorizations. Thus, Locale, the dual category to the category of frames,
is an SFcategory. Locale is extremallywellpowered, but not wellpowered. The authorâ€™s
interest in locales motivated him to use the given definition (which only requires extremalÂ¬
wellpoweredness) for SFcategory rather than defining SFcategory to mean â€œcomplete and
wellpoweredâ€, so that the theory of subfunctors would apply to localic subfunctors in addition
to the previous examples. (For background on frames and their relation to pointset topology,
see Isbell [15], Johnstone [16], Joyal and Tierney[17] and Madden [22].)
Lemma 4.1.3. Assume that A is an (epi, extremal mono)  category. Then
1. (diagonalization) [13, 33.3] If ge = mf, where e is epi and m is extremal mono, then
there exists k such that mk = g and ke = f.
2. [13, 34.2(2)] Any intersection of extremal subobjects is extremal.
3. [13, 34.2(3)] Pullbacks of extremal monomorphisms are extremal mono.
Definition and Remarks 4.1.4. Defining Sub(A)  the lattice of extremal subobÂ¬
jects: In an SFcategory A, the lattice of (equivalence classes) of extremal subobjects has
particularly nice structural features.
45
An extremal subobject of A is an extremal monomorphism s : S â€”> A. Define a
preorder on the class of extremal subobjects by
(s : S â€”â–º A) C (t : T â€”* A) *==>â– 3c : S â€”* T, s â€” tc.
For brevity, one often only mentions the map or object part of an extremal subobject. To
avoid confusion, the same letter will be used to denote both parts, with the lower case letter
used for the function.
Since t is mono, there is at most one map c which manifests s C t. If s C t and t C s,
then there are cj and c2 such that s â€” tc\ and t â€” sc2. So t = icjc2 and s = sc2cj. Since
c2ci shows s C s, c2cj = ids. Similarly, cic2 = idy. One identifies extremal subobjects s and
t if s C t and f C s, or equivalently, when there is an isomorphism c such that s â€” tc. The
set of equivalance classes under this relation is denoted Sub(A).
Any map c, which exhibits s C t, is extremal mono. For if ca â€” cb, then sa =
tea â€” teb = sb, so (since s is mono) a â€” 6; so c is mono. If c = me, where e is epi, then
s = tc = tme. Since s is extremal mono, e must be an isomorphism. Thus c is extremal
mono.
Recall that the intersection of a set M of monomorphisms with a common codomain
is the limit of the diagram generated by M. Since A is complete and intersections of extremal
monomorphisms are extremal, Sub(A) is a complete lattice, with meet operation  f)  and
join operation  J. In general, joins in Sub(A) are not disjoint unions; in general, (J At is
computed using
(jAi = p{A':Vz, ACA'}.
(See Borceux [6, Volume 1, 4.2.2, 4.2.3, 4.2.4].)
A map f : A â€”* B induces an adjoint connection between Sub(A) and Sub(B). (See
Borceux [6, Volume 1, 4.4.6]; Borceuxâ€™s results are phrased in terms of â€œstrong monomorÂ¬
phismsâ€. Under our assumptions a map is strong mono if and only if it is extremal mono.)
If s : S â€”* A â‚¬ Sub(A), use the (epi, extremal mono)  factorization to obtain a unique
46
s' 6 Sub(Â£?) such that
A
f
B
S
S /+1(S)
commutes. The notation /+1(5) is used for the image of S, to remind the reader of the
analogy to ordinary set theoretic images of subsets under maps. The map (s : S â€”> A) iâ€”â–º
{s' : S' = f+1{S) â€”> B) is the left adjoint. The right adjoint f~l : Sub(S) â€”> Sub(A) is
obtained by taking the pullback of s : S â€”> B along /. Thus,
f+1{S) SCf\T),
/+,(US') = U^1(S<).
and
r1{f]Si) = f]r1{Si).
Definition and Remarks 4.1.5. Let F : X â€”â™¦ A be a functor. A subfunctor E of F is a
rule that selects an extremal subobject eX : EX â€”> FX for each X G Obj(X) such that if
f : X â€”>Y, then
(Ff)+1(EX) C EY.
Any such assignment E of F gives rise to a functor E : X â€”> A. Ef : EX â€”> EY is defined
to be the composition EX â€”â–º (Ff)+1{EX) C EY.
Since eY : EY â€”> FT is mono, F/ is uniquely determined by the condition that
Ff
FX â€”>FY
eX
eY
EX â€”â€”EY
commutes. Thus, (Ff)+1{EX) C EY implies that Ef can be defined to make the above
square commute.
47
The converse also holds: if there is a map Ef such that the square commutes, then
(Ff)+1(EX) C EY. To prove this, suppose there is a map Ef which makes the square
commute. Use the unique factorization Ef = ab where a : A â€”â–º EY is extremal mono and b
is epi. Note that (Ff)(eX) = (eY)(Ef) â€” (eY)ab gives a factorization of (Ff)(eX) into an
epi b followed by an extremal mono (eY)a. Thus,
[(eY)a : A â€”> FT] â€” [(Ff)+1(EX) * FT]
and (Ff)+1(EX) C EY. Thus, a subfunctor is exactly â€œa natural transformation whose
components are all extremal monomorphismsâ€.
Let Subfun(F) denote the class of subfuctors of F : X â€”> A. Define a preorder on
Subfun(F) by
E\ C F2 <=> W1 â‚¬ Obj(.A), Ei(A) C E2(A).
If Ei C E2, then, for each A, there is a unique extremal monomorphism cA such that
e\A â€” (e2A)(cA). One easily verifies that cA are the components of a natural transformation
c : Ex â€”> E2; in fact, c : Ex â€”â–º E2 is a subfunctor.
The following constructions show that the class Subfun(F) behaves very much like
Sub(j4).
Construction 4.1.6. Subfun(F) is complete. If {FÂ¿ : i â‚¬ 1} is any class of sub functors,
there is a supremum (JFÂ¿ and infimum P) Et. The supremum is given by (IJEi)(A) â€”
UFj(j4). The infimum is given by (PFÂ¿)(A) = P Ei(A).
Proof. Since A is extremally wellpowered, for each A, the class {EfA) : i 6 1} has a
representative set. Thus, the objectbyobject definitions for J F, and fj Ex make sense. It
is obvious that if U FÂ¿ and p) FÂ¿ are subfunctors that they are the optimum bounds in the
subfunctor lattice.
Let f : X â€”*Y be any map. A h> PÂ¿ EX(A) is a subfunctor because for each i â‚¬ I,
(F/)+1(P EÃ(A)) C (Ff)+1(Ei(A)) C Et(B),
48
so (F/^a Â£i(Â¿))c a Â£(Â£)â€¢
^ Ei(A) is a subfunctor because,
(F/)+1(LJfÂ¿(A)) = U(Ff)+1(Ei{A)) C U^(5).
i i
a
Construction 4.1.7. Suppose a : F â€”+ G is a natural transformation. Then there is an
(order theoretic) adjoint conection
a+1 : Subfun(F) Â«> Subfun(G) : a1.
The left adjoint a+1 is defined by (a+1E)(A) = (aA)+1(EA). The right adjoint a1 is defined
by (a~1E)(A) = a~1(EA).
Proof The order on Subfun(F) is defined â€œobjectbyobjectâ€, and whenever
f : A B,
f+1 and f~l form an adjoint connection between Sub(.4) and Sub(F). So it suffices to check
that a+1E and a~lE define subfunctors.
49
Let E â‚¬ Subfun(F), / : X â€”> Y be a map, and consider the following diagram.
(â– aA)+1(EA) (aB)+1(EB)
The trapezoids (i) and (iii) are obtained by factoring (aB)(eB) and (aA)(eA), respectively;
in each case 2 is the epi part and e' is the extremal mono part. The square (iv) expresses
the naturality of a. The trapezoid (ii) expresses that e : E â€”* F is a subfunctor.
To show a+1E is a subfunctor of G, it suffices to show there is a map
k : (aA)+\EA) * (aB)+1(EB)
that makes the top trapezoid commute. For this, use the diagonalization property from
Lemma 4.1.3. Define / = zB(Ef), g = (Gf)e'A, e = zA and m = e'B; note that ge  mf
with e epi and m extremal mono. Thus, the diagonalization property guarantees the desired
k exists.
Assume f : A â€”> B. A similar diagram is used to verify that a_1E is a subfunctor
of F whenever E e Subfun(G). The missing map (aA) 1{EA) â€”â–º (aB) X{EB) is obtained
from the pullback property of (aB)~1(EB). â–¡
50
Corollary 4.1.8. Let a : F â€”> G be a natural transformation, and e : E â€”> F be a subfuncÂ¬
tor of F. Let e' : a+1E â€”> F be the subfunctor described in Construction 4.1.7. There is a
natural transformation z : a+1E â€”> F such that for each object A, (aA) â€” (e'A)(zA) is the
(epi, extremal mono)factorization of aA.
Proof. One defines e' and 2 as in the preceding proof. Examining the comparison of squares
diagram used to produce a+1E/, shows that 2 and e! are natural. â–¡
Construction 4.1.9. Assume E,F,G,H are functors A â€”> A. Suppose that
1. a : E â€”* F and (3 : G â€”> H are sub functors,
2. Either E or F preserves extremal monomorphisms.
then the horizontal composition (3a : EG â€”* FH is a subfunctor.
Proof. The horizontal composition of natural transformations is always natural. Hypothesis
2 implies that, for any A, (3aA is an extremal monomorphism, because the class of extremal
monomorphisms is closed under composition and
((3a)A := (aHA){E(3A) = (F(3A)(aGA),
by definition of horizontal composition (see Construction 2.3.5). â–¡
4.2 Meseguerâ€™s Lemmas
Lemma 4.2.1. Suppose A is an SFcategory, (T, p, rj) is a monad on A, and F is a subÂ¬
functor ofT, such that (r]A)+1{A) C FA. There is a smallest subfunctor F of T such that
F C F C T and for all A 6 Obj(^l)
(pA)(F2A) C FA.
If the equation
(fj,A)(G2A) C GA
holds for G, we say G is closed under p.
51
Proof. Let T denote the class of all subfunctors of T, that are larger than F and closed
under p. S' is nonempty because it contains T. By Construction 4.1.6, F = exists.
Since S (pA)(S) is an order preserving map Sub(T2v4) â€”* Sub(2L4), so for each A
{pA)(F2 A) C FA. â–¡
The preceding gives an easy candidate for the functor part of the monad generated
by F. One needs a more complex argument if one wants detailed information about the
natural transformations  related to rÂ¡ and /i  which make F into a monad.
The proofs of the following lemmas require detailed computation. For clarity, the
goal of each paragraph in the proof is written in boldface.
Lemma 4.2.2. (Assume notation and hypotheses from Lemma If.2.1.) There are natural
transformations Ã± : id^ â€”Â» F and m : F2 â€”> F which make (F, n, m) a monad.
Proof. One defines four sequences of natural transformations:
1. f\'.F\â€”>T the subfunctor generated at stage A,
2. n\ : id^ â€”Â» F\  a natural transformation obtained from rj by suitably modifying the
domain and codomain,
3. m\ : Ff â€”> F\+1  a natural transformation obtained from p by suitably modifying
domain and codomain, and
4. c\ : F\ â€”> F\+i  the inclusion.
Define F0 = F; use the notation f0 : F0 â€”> T. m0 and Co are defined according to the same
pattern that defines later â€œmâ€s and â€œcâ€s, which is described below.
52
Defining f\ : F\ â€”> T (A > 0), mx : Fx* * Fx+i, and cx : Fx â€”Â» Fx+1. Assume
that fx has been previously defined. Consider the following diagram:
(M)+1(F\)u F*A
Given fx and (aA, the square is obtained by (epi, extremal mono)factorization of the comÂ¬
posite (fj,A)(fxA). All maps in the triangle are extremal monomorphisms, obtained by
comparing subobjects of TA. Define
FX+1A = (nA)+\FÂ¡A) U FXA
and
mx : FXA â€”> Fx+iA
to be the map shown on the left side of the diagram. It follows from Constructions 4.1.7 and
4.1.6 that fx+1 : Fx+i â€”â–º T is a subfunctor. Note that
{fx+i)(mxA) = (nA)(flA)
and mx is natural. For bookkeeping purposes, let us call
c\: Fx â€”> Fx+1
the subfunctor which exhibits Fx C Fx+i.
If k, is a limit ordinal, define
Fk = : A < k};
53
fK'FKâ€”>T is a subfunctor by 4.1.6.
Defining the sequence of â€œn*â€ s: The following definition of nx is not recursive; n\
can be defined once we know f\, but the definition of f\ does not involve nx at all. Consider
the following commutative diagram of functors and natural transformations.
The natural transformations z and e are defined by condition that rj = ez is the (epi, extremal
mono)factorization of 77, as described in 4.1.8. By assumption and the construction of the
sequence (Fx),
(r)idA)+1 CF0C Fx.
Let i\ denote the natural transformation such that e = fxiX) ix exists because e C i. Define
nx := i\z. Evidently,
V = f\nx
and
^a+i = cxnx.
What happens when the sequence terminates: Since A is extremally wellpow
ered, for each A, the sequence (F\A)\ of subobjects of TA eventually terminates, say when
A = Â«1. Again using extremal wellpoweredness there is an ordinal, say k2, such that the
sequence (F\FK(A))x of subobjects of TFKlA terminates at k2. Define Ã± = nK2, f â€” fK2
and m = mK2. These assignments give natural transformations; to check this, one considers
a map / : A â€”* B, and chooses k large enough that the subobject sequences (described
above) terminate for both A and B. It should be clear that each subobject sequence {FXA)x
terminates at FA.
54
Verifying that (F, n, m) is a monad. To prove the unit laws, it suffices that for
all A,
(mxA){FxnxA) = (mxA)(nxFxA) = cxA
(because, once the sequence terminates, the â€œcaAâ€s become identity maps on FXA). Fix any
ordinal A. Naturality of fx and nx implies
CTnxA)(fxA) = (fxFxA)(FxnxA)
and
(nxTA)(fxA) = (FxfxA)(nxFxA);
Note that rj = fxnx and, by definition of horizontal composition,
flA = (TfxA)(fxFxA) = (fxTA)(FxfxA).
Calculating, one finds
(flA){FxnxA) = (fxTA)(FxfxA)(FxnxA) def. noted above
= (fxTA)(Fxr]A) noted above
= (TrjA){fxA). /a is natural
Thus,
UlA){FxnxA)  (TVA)(fxA)
and (by a similar computation)
(.flA){nxFxA) = (rjTA)(fxA).
To show (mxA)(FxnxA) = cxA, consider the following diagram:
FXA
55
the upper square commutes by an identity proved above; the lower square is the definition
of mx; the upper triangle is trivial; the lower triangle is the unit law for (T, rÂ¡, /i). Reading
the perimeter of the diagram, one finds
fx A = (fx+1A)(mxA){FxnxA).
By uniqueness of maps manifesting inequalities between subobjects,
cxA = (mxA)(FxnxA).
The diagram needed to prove
{m\A)(nxFxA) = cxA
is similar and omitted. This establishes the unit laws.
To prove the associative law for (F,Ã±,m), one may choose sufficently large A, then
draw a diagram comparing the associative squares for T and F, using appropriate powers of
fx to compare the corners. The comparison squares commute because of the definitions of
mx and horizontal composition. â–¡
Definition 4.2.3. Assume that A is cocomplete. In A, unions of chains are colimits, if
whenever D is a diagram, where Obj(D) consists of a chain of extremal subobjects of A and
Map(D) consists of all inclusions (in Sub(A)) that exist among elements of Obj(D), the map
c : colimD â€”> A induced by colimit properties is an extremal monomorphism.
In this case, c : colimD â€”> A is an extremal subobject which contains each element of
Obj(D); thus there is an extremal monomorphism colimD â€”> (J Obj(D). Since (J Obj(D) is
the supremum of Obj(D) in the Sub(A), this map must be an isomorphism.
Remark 4.2.4. The condition that unions of chains are colimits is satisfied in many natÂ¬
urally occuring categories. For example, in â€œrelationalâ€ categories like Top, P, and Set,
extremal subobjects are just subsets with the induced structure, so all unions are colimits.
Assume A is a category of finitary algebras. The (set theoretic)union of a chain
(Ai) of subobjects is a subobject. The (set theoretic)union A' of (A,) lies inside a smallest
56
extremal subobject, which by our notation is (JAÂ¿. Evidently, the coproduct of the chain
(Ai) of subobjects is A'. So the issue about whether â€œunions of chains are coproductsâ€
amounts to checking whether A! = 1J At. This depends on the delicate issue of whether
epimorphisms are surjective. If epis are not necessarily surjective, then 1J A2 will be the
largest subobject in which A! is epi, which will generally be larger than A'.
Definition 4.2.5. As above, assume A is an SFcategory, is a monad, and F is
a subfunctor of T which contains r?+1 id^. A partial algebra (A, a) is an object A equipped
with a map a : FA â€”> A such that o(noA) = id^.
Let pAlg(Fo,no) denote the category of all partial algebras, with maps / : (A, a) â€”>
(B, b) such that / : A â€”> B is an Amap, and fa â€” b(Fof).
Lemma 4.2.6. Continue with situation and hypotheses from f.2.2. Suppose A is cocomplete
and unions of chains are colimits. Suppose (A, a) is a partial algebra. The partial algebra
structure a extends to an Falgebra structure map a : FA â€”> A if and only if for each ordinal
A there is a unique a^+i making the diagram below commute.
H(A) â€”â€”
m\
Fx(A)
FX(A)â€”^ >A
Proof. To prove a map a : FA â€”> A exists, use transfinite induction. The hypothesis gives ax
at sucessor ordinals. To construct ax, when A is a limit ordinal use the coproduct property
of FXA â€” UÂ«a
Inductively one shows that
(ext) for each A, ax â€” aA+i(cAA) (each ax+i extends the preceding ax),
(unit) for each A, ax(nxA) = kU,
57
If (unit) holds for sucessor ordinals, then it holds for all ordinals. Suppose
k, is a limit ordinal. The map aK is epi, because it is a limit of epimorphisms. Note that aK
and aKnKaK are both compatible maps FKA â€”> A; the definition of colimits implies that
because aK is epi, id^ = aKnK.
The algebra associative law holds for (^4, a). If a â€” ax, then ax = aA+1, so by
the hypothesis regarding the existence of ax such that a\+i(rri\A) = ax(Fxax) proves the
associative law.
If (unit) holds for A, then (ext) holds for A. Note that the equation a\+\(rn\A) â€”
a\{F\a\) defines a>+1. By induction hypothesis (unit),
(â– Fxa\)(FxnxA) = idFj,4;
therefore precomposing both sides of the defining equation for aA+i with (F\ri\A) shows
aA = ax+i(mxA)(FxnxA) = ax+i{cxA).
The second equality follows from (mxA)(FxnxA) = (cxA), which was proved in verifying the
unit laws for F.
If (unit) holds at A, then it holds for A + 1. Consider the following diagram.
The left triangle commutes because (unit) holds for A; the bottom trapezoid commutes by
definition of ax+1; the top trapezoid commutes because
(nxA)(FxnxA)(mxA) = (nxA)(cxA) = nx+1A.
58
The first equality holds by proof of the unit law for F; the second equality holds because of
the compatibility between â€œnAâ€s. Thus,
aA+i(nA+i) = ax+1(mxA)(FxnxA)(nxA)
= ax(Fxax)(FxnxA)(nxA)
= ax(nxA)
= hU;
this establishes the unit law for (A, a). â–¡
Use the notation Alg(F0, no) to denote the full subcategory of pAlg(Fo, no) containing
objects satisfying the hypotheses of Lemma 4.2.6.
Lemma 4.2.7. Suppose the preceding lemmas apply. Also suppose each mx is epi. The map
(A, a) iâ€”> (A, a) induces an equivalence of categories; the functors involved are
E : Alg(F0, n0) *â€¢ AT,
defined by extension of structure and
> Alg(F0,n0)
given by restriction.
Proof. E is legitmately defined. Lemma 4.2.6 defines E on objects. Maps in AF are
Amaps compatible with the structure. So it suffices to show that for any Alg(Fo,n0) map
f): A B,
FXA FXB
a\
bx
B
59
commutes for each ordinal A. By definition of Alg(F0, n0), the square commutes for A = 0.
Suppose it commutes for A; note that
ax(Fxax)
def. a^+i
= bx(Fx)(Fxax)
ind. hyp.
= bx(Fxbx)(F^)
ind. hyp.
= bx+l{mxB){Fl4>)
def. bx+1
= bx+l(Fx+i4>)(mxA)
naturality mx.
Because mx is epi, one concludes ). At limit ordinals, properties of
colimits insure that the diagram commutes. By transfinite induction, and the defintion of
the extended structure, any map which preserves (Fo, n0)structure preserves Fstructure.
One also notes that restriction respects maps, because Fo is a subfunctor of F.
E and R form an equivalence. Obviously,
V(A, a0) G Obj(Alg(Fo, n0)), RE(A,ao) â€” (A,a0).
The construction of the extended structure shows the restriction of any Fstructure to Fo
uniquely determines the Fstructure; hence
ER(A, a) = (A, a).
â–¡
4.3 A Partial Algebra Which Does Not Extend
The category Set is an SFcategory and unions are colimits, so Meseguerâ€™s Lemma
applies to Set. We show the necessity of Lemma 4.2.6â€™s hypothesis that a map ax+j such that
ax+imx = ax(Fxax) can be defined. The section discusses a subfunctor of the freemagma
monad, which has an algebra that cannot extend to a monad algebra. A magma is a set
with a binary operation, subject to no equations.
The free magma monad (T, rj, fi) has the following parts:
1. Given a set X, TX consists of all words with variables in A. A magma word is any
expression formed by finitely many applications of the rules:
60
(i) If x G X, then (^)is a magma word.
(ii) If s, t are magma words, then
is a magma word  the product of s and t. For ease in reading, we use binary tree
notation for products.
2. T defines a functor: given / : X â€”> Y, to compute Tf we apply / to all members of
X in a given word, leaving the tree and circle structure unchanged.
3. The â€œinsertion of variablesâ€ map rÂ¡X : X â€”> TX. (rjX)(x) = (^.
4. The â€œsemantic compositionâ€ map [iX : T2X â€”> TX sends a word s G T2X of words
to a word in TX, by removing the circles around each element of TX used in making
s. n is also a natural transformation.
The notation takes a little while to soak in; to expedite the process, we consider a
calculation with TN. Suppose s G TN is the word
and ti = ss G T2(N) is the word
61
then (/Â¿N)(tx) is
Now we define a subfunctor F of T. Define the depth dep(n) of a node n in a binary
tree inductively by: the depth of the dep(root) = 0; if a is immediately below b, then the
dep(a) = dep(b) + 1. Define a leaf to be a node that has nothing below it. Let F consist of
all rooted, labeled, binary trees (i.e., magma words) such that the depth of each leaf is the
same. For example s and (/rN)(ti) are in F(N), but
is not, because dep(57) = 1 and dep(l) = dep(4) = 2. Evidently, F is a subfunctor, and for
any X, (rÂ¡X)+1(X) C F(X). But, for any nonempty X, (pX)(F2(X)) is not contained in
F(X). One readily verifies that F = T.
Now we define an partial algebra structure on N that does not extend to a Talgebra
structure. As in Lemma 4.2.6, a structure map for a pair (F,rj : id^ â€”Â» F) is a map
a : FA â€”> A satisfying idÂ¿ = a(rjA). Define n : F( N) t Non a tree r as follows: if the
depth of each leaf of r is odd then n(r) is the leftmost label, if the depth of each leaf of r
is even (or zero), then n(r) is the rightmost label. For example, n(s) â€” 1 (s defined above)
and n(/iN)(ti) = 4.
62
In order for n : F(N) â€”> N to extend to n : Fi(N)
making
F2( N)
Fn
â– F{ N)
/xN
F( N)
n i
â– *N
commute. No such a: can exist, for (/zN)(ti) is
N there must be a function rii
which equals (/iN)(t2), where t2 is
Finally, note that (Fn)(ti) is
so n(Fn)(ti) = 1. Note that (Fn)(t2) = (/rN)(t2), so n(Fn)(t2) = 4. Thus we have elements
11,12 G F2N, that /xN identifies and n(Fn) does not identify, so there is no function n\ such
that n\(nN) = n(Fn).
CHAPTER 5
FREE ALGEBRAS
This chapter explores a generalization of complete distributivity for P^objects. In
several steps, the free complete distributive lattice monad is constructed, then Meseguerâ€™s
Lemmas are used to construct an appropriate submonad, whose algebras generalize comÂ¬
pletely distributive lattices.
Section 5.1 describes monads for complete semilattices; D is the monad for complete
join semilattices, 11 is the monad for complete meet semilattices.
The gist of section 5.2 is that there is a lifting of 'll over T> (and a lifting of D over 11).
Therefore, UT> and T>U are monads over P. (See results of Beck [5] summarized in Section
3.4.) In Raney [27], it was shown that complete meet distributivity is the same as complete
join distributivity. Hence, the composite monads UT> and Dll have the same category of
algebras. The objects in either category are complete lattices where meets distribute over
joins, and joins distribute over meets.
Completely distributive complete lattices have been thoroughly studied. The basic
structure is described in Raney [27], [28], and [29]. Free objects over Set were initially
described in Markowsky [24]. Tunnicliff [32] discusses properties of the free completely
distributive lattice over a poset. Free objects over P and the relationship between completely
distributive lattices and continuous lattices are described in Hoffman and Mislove [14]. The
approach here is apparently new, but yields obviously equivalent free objects.
Each pair (j,m) of subset systems gives rise to a subfunctor of UT>. Meseguerâ€™s
Lemmas are applied to this subfunctor to produce a monad F. Any P^,objects which
63
64
is P^embeddable in a completely distributive complete lattice is an Falgebra. Any F
algebra has a natural structure. Because of computational difficulties, no exact algebraic
characterization of Falgebras is given here.
A word about notation: the (functor parts) of the monads described below are given
by families of sets. Thus, checking the unit and associative laws requires working with many
levels of the power set tower. Roman letters S,T,  â– â– denote sets. A subscipt designates the
â€œpower set complexityâ€: Si G 7(A), S2 G 72A  S2 is a family of sets, S3 G 73A  S3 is a
family of families of sets, etc.
5.1 Complete semilattices
In this section, we describe the free complete join (resp. meet) semilattice on a poset
A, using the monad (D, d, /u) (resp. ('U,i,/x)). The description of D is well known; for
example, Meseguer [25] uses it. The reader will have noticed that q is used as a name for
two different natural transformations; this would normally be horrible notation, but in this
case, the formula for y is the same. Thus, our notational econony should cause no confusion.
The functors 'll and T> act on a poset A by
11(A) = {SCA:x>y6S =Â£> x G 5}
 the set of increasing subsets of A ordered by reverse inclusion and
T>(A) = {SCA:x x G 5}
the set of decreasing subsets of A ordered by inclusion.
Given monotone f : A â€”> B, we define
Vf : 'D(A) * D(B) : S ^ {b G B : 3s G S, b < f(s)}
and
Uf : U(A) Â» U(B) : S ^ {be B :3s e S,b> f(s)}
The following facts will be of later use.
65
1. tÃ and D are functors: trivially they respect identity arrows. Observe that for any
S â‚¬ A and montone functions / : A â€”> B, g : B â€”> C,
T>(gf)(S) = {câ‚¬C:3seS,c
= {ceC:3f(s)e'Df(S),c
= (%)('D/)
and, similarly, U(gf)(S) = (Ug)(Uf)(S).
2. For each poset A, D(A) is a complete lattice, with supremum operation given by set
theoretic union, and infimum given by intersection. tt(A) is a complete lattice, with
infimum given by union and supremum given by intersection.
3. For each S2 C D(A),
In particular, this holds if 6 is empty. So Vf preserves all suprema. Similarly, if
S2 C 'U(A), then
Uf{\j6)=\j{(Uf)(S):See}
so that Uf preserves all Ãnfima.
Define
dA: A â€”> T)(A) : a {x â‚¬ A : x < a},
iA : A â€”* 'U(A) :aH{iÂ£i:i>a}, and
fiA : T>2{A) V(A) : 52 â‚¬ D2(A) ^ (J S2.
It may be puzzling that D(A) is ordered by subset inclusion and li(.4) is ordered by
reverse subset inclusion. The fact that
x < y dA(x) C dA(y) <=> iA(y) C iA(x)
66
motivates the choice, because one wants iA to be order preserving. Moreover, one wants 1if
to preserve all Ãnfima, which only happens if the infimum operation in UA is union.
The reader may check that i, d and fi are natural transformations. The monad
assocative law
D3 â€”â€”*T>2
[iD
D2
â– V
holds for both D and 'll, because if S3 e IP3 (.A), then
Ul> = U{Usâ€™:S*e4
The monad unit laws
hold for V, because if S G D(A), then
S = \J {{dA)(x) : x e 5} = U [t â‚¬ T>(A) : T C 5}.
Similar computation shows the unit laws hold for It.
Now to describe algebras over these monads:
Lemma 5.1.1. Let T C T>(A), such that for all x â‚¬ A, (dA)(x) â‚¬ T. Then the following
are equivalent:
1.there is an order preserving map a : T â€”> A such that id^ = a(dA);
2.each S e T has a supremum;
3.dA : A â€”* T>A has a right adjoint.
67
The analogous conditions involving T C 1LA are also equivalent.
Proof. We give the proof for T C T>A, leaving the upsidedown argument for T C UA to the
reader.
(2 => 1) Define a(S) = \J S and compute.
(1 => 2) a(S) is an upper bound for S, because if x G S, then (dA)(x) C S,
whence x = a(dA)(x) < a(S). Suppose for all x G S, x < u. Then (dA)(x) C (dA)(u), so
S = J{(eL4)(:r) : x G 5} C (dA)(u), therefore a(S) < u.
(2 3) By definition of V S, j x C S <*=>â€¢ x <\J S. â–¡
Proposition 5.1.2. PÂ°Â° is equivalent to the category of T)algebras. Poo is equivalent to the
category ofUalgebras.
Proof. The first assertion is proved, leaving the second to the reader. By Lemma 5.1.1, any
Dalgebra has a supremum for each decreasing set; the supremum of an arbitrary set S is
equal to the supremum of the decreasing set ]. S. By the definition Dalgebra homomor
phisms coincide with order preserving functions which preserve all suprema of decreasing
sets. It is easy to see that a map preserves suprema of all decreasing sets if and only if it
preserves suprema of all sets.
Conversely, if A G PÂ°Â° it has structure map
a : V{A) â€”* A : S S.
In Lemma 5.1.1 it was noted that a satisfies the unit law for structure maps, a satisfies the
associative law by the order theoretic fact proved in Lemma 5.1.3 below. â–¡
Lemma 5.1.3. For any S2 G T>2(A),
V {x â‚¬ A : 3S1 G S2, x<\JSr] = \J\JS2.
Proof. To verify this equality, one notes that for all x G G S2 such that x â‚¬ Si,
so x < V Si. Thus the left hand side dominates the right. If u is an upper bound for (J S2,
and 3Si G S2, x < \J Si, then x < u. So the left hand side is the least upper bound of [J 6;
this establishes the lemma.
â–¡
68
5.2 Completely Distributive Complete Lattices
In the following arguments, the forgetful functors are supressed from the notation.
The lemmas that follow establish that there is a lifting of It over V (and similar arguments
show that there is a lifting of V over It). Explicitly this means:
1. For any Ualgebra A, there is a ltstructure on V(A).
2. It maps Dalgebra maps to Dalgebra maps, so we may view U as a functor PÂ® â€”> PÂ®.
3. Both natural transformations i and p \ It2 â€”> It preserve all joins; similarly, d and
fi : D2 â€”*â– D preserve all meets.
Lemma 5.2.1. For any poset A, both lt(A) and V{A) are complete lattices. Thus, li(.A)
and V(A) are both 'llâ€” and Dâ€” algebras.
Lemma 5.2.2. For any poset A,
1. dA : A â€”> V(A) prserves all existing Ãnfima.
2. iA : A â€”Â» U(j4) prserves all existing suprema.
Proof. Suppose S C A and u = f\S exists. Since dA is order preserving ,
(dA)(u) C P jcL4(:r) : x G
For the reverse inequality, suppose i G f]{dA(x) : x G S'}, i.e., i is a lower bound of S. Then
l < u, so Â£ G dA(u). The proof for uA is upsidedown, but otherwise identical. â–¡
Lemma 5.2.3. If f : A â€”* B preserves all suprema, then It/ : 11(^4) â€”> U(B) also preÂ¬
serves all suprema. Similarly, if f preserves Ãnfima, so does T>(/). Therefore, for any order
preserving map f, both UT>f and VIIf preserve all Ãnfima and suprema.
Proof. A proof of the first fact is given, the second is very similar, but notationally easier. Use
the fact that lt(A) is a complete sublattice of tP(A), and therefore completely distributive.
69
Let (S\)\eL be an indexed subset of li(A); for each S\, choose a family (x\^K : k E K) such
that
S\ =T {x\,K : k â‚¬ K).
Recall that:
â€¢ joins in UA are intersections;
â€¢ iA preserves suprema, which reads
t \] y\ = (iA)(\J y\) = 0 T y
A
for any indexed family (y\) C A.
One computes as follows:
= u/(niu^")
A K
= U/(U{ n t xa,c(A) : C : L * /i})
= u{ t/(VxV(A)) : c:L>k}
'a
= c:L>k}
= c.LK}
1 A
= n uf^
â–¡
Corollary 5.2.4. For any poset A,
1. p : T>2(A) â€”> D(A) prserves all Ãnfima.
2. p : li2(A) â€”> U(A) prserves all suprema.
70
Proof. Suppose S2,a Â£ X>2(^), for A e L : claim 1 amounts to
unsw=nusw
A A
Clearly, the left hand side is contained in the right. Suppose that x is a member of the right
hand side, that is, for each A there exists S\t\ â‚¬ Â¿2,a with x â‚¬ Si,a. Then
Si = Pi Si,a
A
is in each S2,a because each S2,a is downward closed  relative to the inclusion order in Â©2(.A).
Since a: â‚¬ Si, this proves that x is in the right hand quantity, and the desired equality of
sets holds.
The proof for the It works similarly. â–¡
The preceding lemmas show that there are liftings of 'll over Â© (and vice versa), so
we have:
Corollary 5.2.5. ItÂ© = (ItCD, 77, where
rj : idp â€”â–º ItÂ© : a i> {S â‚¬ VA :aCS}
and
1/: UDltÂ© â€”* ItÂ© : S4 iâ€”> un*.
is a monad over IP whose algebras are completely distributive complete lattices.
Proof. The only thing left to be proved is the formulas for the natural transformations. By
Beckâ€™s reasoning  as summarized in Section 3.4, g is the horizontal composition i â– d. To
compute v : ltÂ©ltÂ© 4 ItÂ© one uses the fact that vA is the structure map for the free algebra
'llÂ©A Consider the situation in light of the discussion following Theorem 3.4.5. (Here It
plays the role of 5, while Â© plays the role of T.) The relation a â€” ar{Tas) applied to the
object UVA implies
Mxs.) = u(Â®n)(*)
= U{reBA:3S2efS,,rcfS2}
= Ui>
71
The assertion about algebras follows from Theorem 3.4.5. â–¡
Remark 5.2.6. IÃD is a monad because there is a lifting of 'll over PÂ®. One could use
the correspondence between liftings and distributive laws outlined in 3.4 to find a natural
transformation DU â€”> UD. This distributive law is not needed for the calculations which
follow, and is somewhat cumbersome, so its explicit description is omitted.
5.3 Some categories of algebras
In this section, Meseguerâ€™s Lemmas are applied to the monads described above. Note
that P  the category of posets and order preserving maps  is a cocomplete SFcategory in
which unions of chains are colimits. Note that extremal monomorphisms in P are inclusions
of subsets with the induced order: each order preserving map / : A â€”> B factors as A â€”>
f(A) â€”y B, where f(A) is the settheoretic image of A, with the order induced from B.
Thus, the results of Section 4.2 do apply to P.
The reader is advised to review the definition of subset system, given in 1.2.1, if
necessary.
Assume all subset systems Z are nontrivial in the sense that for each A, and a â‚¬ A,
{a} â‚¬ Z(A); this does not reduce the generality of the argument, because given any subset
system, one may adjoin all singletons to it without changing which optimum bounds are
preserved.
Theorem 5.3.1. [25] P7 and Pm are monadic.
Proof. Given a subset system j
J0(A) = {iS:Sej(A)}
defines a subfunctor of D. This extends to a subfunctor j of D such that (j, d, Jl) is a monad.
The natural transformations d and JL are obtained from d and /i by suitably modifying the
domains and codomains. The precise definition of these maps is contained in the proof of
Lemma 4.2.6.
72
If (A, a0) has a J0structure, Lemma 5.1.1 shows that
ao{S) = \J S
for all 5 e J0(A). Thus, to define ax one must show some map JXA A makes the following
diagram commute.
JSA JÂ°(V) *J0A
U
JXA
Lemma 5.1.3 shows that \J S is defined for all S â‚¬ J\A and, moreover, that for S2 â‚¬ JqA,
VU S2 = V(JÂ«V )(Â«>)â€¢
Identical arguments show show that for each A, if
ax(S) = \/S
that
ax+1(S) = \/S.
It follows that any poset, A, in which each member of JqA has a supremum, each member of
J(A) has a supremum. Moreover, each map preserving J0suprema also preserves Jsuprema.
Thus, one obtains an equivalence of categories between P and PJ.
An upsidedown version of this argument shows that Pm is a category of algebras. â–¡
Given subset systems m and j, F := M0J0 forms a subfunctor of
T := UV.
By Corollary 5.2.5, (T,u,r]) is a monad on P. By 4.2.1, there exists a smallest F above F
that is closed under v. This monad is used to discuss the categories defined below.
73
Definition and Remarks 5.3.2. Let OP^ denote the category of completely distributive
complete lattices with maps, which preserve order and all optimum bounds. PP^ denotes
the full subcategory of containing objects such that there is a P3m map : A â€”> B, with
4> Pextremal mono, and B â‚¬ Obj(DP^).
denotes the category of Falgebras, where F is the monad described above.
It will be shown that
SpWm c mi c Ml.
To see the first inclusion, one notes that any family T of subsets of a set X which is also a
P^ object is in DP^; the inclusion T C 7(X) is a P^embedding and T(A') is a completely
distributive complete lattice. Proposition 5.3.7 shows the second inclusion.
Let A 6 BP^, and : A â€”+ B be given as in the definition of BP^; let b : TB â€”> B
be the Tstructure for B. The verification of the hypothesis of 4.2.6 proceeds by comparing
the diagram to be completed with the lt!Dalgebra associativity diagram for B. Prom this
point on, the notation of Lemma 4.2.2 is adopted with slight modification; u : T2 â€”> T is
the monad multiplication; v\ : F% â€”â–º F\+\ is defined inductively (and plays the role that
was played by m\ in the proof of 4.2.2); rÂ¡ : idp â€”â–º T is the unit; rÂ¡\ : idjÂ» â€”> F\ are defined
inductively, and play the role that was played by n\ in 4.2.2. The natural transformations
fx : Fx â€”> T and c* : F\ â€”* F\+i play the same roles as in Lemma 4.2.2.
Remark 5.3.3. By Lemma 3.4.5, the structure map b : TB â€”> B is given by:
b^ = A{V5i:5lâ‚¬52}
Since is a P^map, the following diagram commutes.
Fd> f\B
FA â€”>FF >TB
ao
B
B
74
The left square expresses that <Â¡> commutes with the Fstructure; the right square expresses
that the Fstructure on B is the restriction of the Tstructure.
Lemma 5.3.4. Define ao : FA â€”> A by a0(S2) = f\(U \J)(S2); this map satisfies a0T]o = id^.
Proof. Let us begin by more explicitly calculating ao. Using the definition of li on maps we
find that
a0(S2) = A(UV)(*)
= /\{xeA:3S1eS2, \/Si
= /\{\/S1:S1eS2}.
This definition makes sense; each Si G Jq(A) so each \J S\ exists because A G P^. (IX V) is
an order preserving map, and S2 G mojo(A), which implies that
(U\J)(S2) e M0(A).
Thus, the meet defining ao(S2) exist.
Recall rj(x) = {S â‚¬ DA x C 5}. Now one calculates
ao(>?(*)) = /\{\/S:SeDA,lxCS}
= x
â–¡
For later calculations, it is crucial to know that some of the maps are epi or mono.
Lemma 5.3.5. For each ordinal X, u\A is surjective. Therefore v\A is Pepi.
Proof. Because of the structure of P, unions of subposets actually are set theoretic unions.
Let 5 G Fx+i = Fx U (i>\A)+1(FfA). If S â‚¬ (v\A)+1(FfA), then there is an S2 G F%A such
that (v\A)(S2) = S. Since c\A = (nxA)(rixFxA) so for any Si G FxA, T2 := (rjxFxA)(S) is a
member of FfA such that (uxA){T2) â€” S\.
This lemma holds in any category where unions of subobjects actually are settheoretic
unions. â–¡
75
After noting that extremal monomorphisms in P are injections / : A â€”> B, where
f(A) has the order induced as a subset of B, one easily verifies:
Lemma 5.3.6. The functors U, T>, T, and F\ all preserve extremal monomorphisms.
Now one can verify the hypothesis of 4.2.6 is satisfied.
Proposition 5.3.7. Let A â‚¬ Obj(DP^), B e Obj(BP^), : A â€”> B as in the definition of
BP^,. Then
(ex) for each A, there is a unique map
â€œA+i â€¢ F\+iA A,
such that
a\(F\a\) â€” Ã¼a+iI'a;
(com) for each A, fiax = b(f\B)(F\).
Thus, OP^ C
Proof. Remark 5.3.3 established (com) for A = 0. If (com) holds for sucessor ordinals,
properties of unions insure that (com) holds at all ordinals.
76
Suppose (com) holds at stage A; we will establish (com) and (ex) hold for
stage A + 1. Consider the following diagram.
T2B >TB
The diagram commutes: the outer square commutes because of the Talgebra associative law
for B\ the left outer trapezoid commutes because of the definition of vxB\ commutativity of
the right outer trapezoid is obvious; the left inner trapezoid expresses naturality of vX] the
top and right inner trapezoids commute by (com).
Establishing commutativity of the top outer trapezoid requires a bit more. Since
fx:Fx>Tis natural, (Tb)(fxTB) â€” (fxB)(Fxb). By definition of horizontal composition
flB = (fxTB)(FxfxB).
Thus,
CTb){flB) = (Tb)(fxTB)(FxfxB)
= (fxB)(Fxb)(FxfxB)
This proves that the top outer square commutes.
Consider the (epi,extremal mono)factorization of
b(fx+1B)(Fx+1y,
77
say it factors as
Fx+lA ^A'CB.
Since vxA is epi, the (epi, extremal mono)factorization of
q := b(fx+1B)(Fx+1)(vxA)
is
FXA A! C B.
The proof of Lemma 4.2.6 shows that each existing ax is split epi, which implies both ax and
(Fax) are epi. Since
q = 4>ax(Fxax),
the (epi, extremal mono) factorization of q is also:
FXA â€œa(4â€œa) ACB.
Uniqueness of factorization implies there is an isomorphism i : A â€”> A! compatible with the
factorizations. So a^+i = ia is a map making (com) hold for A + 1.
Reading off the diagram one finds that:
ax(Fax)  b(Tb)tf2xB)(F2x)
= b{vB){flB){Fl)
= b(fx+1B)(Fx+1)(vxA)
=
The last step used that (com) holds for A + 1. Since is a monomorphism, one concludes
that (ex) is satisfied for A + 1.
ax+1 is the only map which makes the inner square commute because vxA is epi, by
Lemma 5.3.5. â–¡
78
Corollary 5.3.8. DP^ is a full subcategory ofMPm.
Proof. By Lemma 5.3.3, any P^map preserves the Fstructure of a B)Pjn object. Lemma
4.2.7 implies that any P^map extends to a map. â–¡
Corollary 5.3.9. The forgetful functor DP^ â€”> P has a left adjoint.
Proof. First note that FA G Obj(DP^) for any A, because fA : FA â€”> TA is an embedding
of FA in a completely distributive lattice. Because FA is the free Falgebra on A, if
/ : A â€”â–º B is an order preserving map from A G Obj(P) to Be Obj(DP^) C Obj(PF), then
there is a DP^unique map f* : FA â€”> B. This proves that rj has the universal property
described in 2.5.2.3; thus, F is the left adjoint to the forgetful functor DP^ â€”Â»P. â–¡
Corollary 5.3.10. The forgetful functors DP^ â€”> OP^, and DP^ â€”> have left adjoints.
Proof. Apply Proposition 3.3.6 and Corollary 6.2.8. (Note: the proof of 6.2.8 does not
depend on the arguments in this section, so the result is listed here. There is no circular
argument.) â–¡
There is some, rather limited, information about members of M^.
Proposition 5.3.11. If A G Obj(M^), then A G Obj(P^).
Proof. Define JA, MA, a, Â¡3, 7, and 5, by the requirement that the squares below are
pullbacks.
UA
*WD A
By assumption, j(A) and rn(A) contain singletons, therefore
(â– iT>A)+1{J0A) C MqJqA C FA,
79
and
(UJA)+1(M0A) C Mq Jo A C FA.
It follows that
J0A C JA = (iDA)_1FA
and
Mo A C MA = (ltdA)_1FA
Again using the fact that j and m contain singletons,
(JA)+1(A) C JA
and
(Â¿A)+1(A) C MA,
so corestrictions dA\ : A â€”> JA and iA : A â€”> MA satisfying a(JA) = JA and 7(1 A)  = fA
exist.
If a : FA â€”> A is an Falgebra, then a/3(JA) = id^ and a5(iA\) = id^. Lemma 5.1.1
shows that for any subset 3 C DA containing each principal downsegment [ x, an order
preserving map y : 7 â€”* A satisfies y(JA) = id.4 if, and only if, y(S) = V5 for each 5 â‚¬ J
(and analogously for subsets of ICA). Thus, we conclude a(3 = \J and aS = /\. â–¡
Remark 5.3.12. If j = m = u>  the subset system which selects all finite subsets, then
SpF^ = DPÂ£. For general j and m, the author does not know if equality holds, but he
suspects that the equality does not always hold.
Remark 5.3.13. A fundamental difficulty working with M^n is that one does not immeÂ¬
diately know any order theoretic formula for the Fstructure maps. One might hope that
each Fstructure map is S2 /\(U \f)S2, but the author cannot presently substantiate such
hopes.
80
Remark 5.3.14. The author knows no examples of objects in Obj(M^) \ Obj(DP^). In
light of Corollary 5.3.10, any object is freely embedded in a completely distributive
lattice. In fact, A â‚¬ Obj(M^) is in Obj(DP^) if, and only if, the unit of the adjunction
mentioned in 5.3.10 is Pextremal mono.
Li[18] explicitly constructs a map u : P â€”> ISF(P) (not necessarily mono), with the
following properties:
â€¢ ISF(P) is a complete, completely distributive lattice.
â€¢ u preserves designated meets and joins.
â€¢ If / : P â€”> A is a map preserving designated meets and joins, and A is complete,
completely distributive, then there is a unique map f* : ISF(P) â€”> A such that
/ = f*u.
To explain the relationship between Liâ€™s results and the results here, note:
â€¢ Li uses families SP C TP and IP C TP, which are only required to contain singletons.
Joins of SPsets and meets of IP sets are required to be preserved. In this document,
the choice of distinguished optimum bounds is made for all posets at once, via a subset
system.
â€¢ Comparing universal properties, one sees that Liâ€™s construction applied to P G Obj(M^)
with SP â€” j(P) and IP = m(P) yields the left adjoint mentioned in 5.3.10.
Unfortunately, it is very difficult to see when u is an embedding. Let [5] denote the
smallest downward closed family of A containing S which is closed under SPjoins. Let TP
denote the class of increasing sets, which are closed under /Pmeets. Li defines IS(P) to
be the familly of all decreasing sets in TP, ordered by S â– < T if, and only if, there is an
indexed chain (SÂ¿)jâ‚¬/ , with I = Q D [0,1] such that S0 = S, Si = T and whenever i < j,
Si fl [P \ Sj] â€” 0. The map u : P â€”> ISP is defined by u(x) = {U e 1L4 : U ^ x}.
81
Nonetheless, BP^, it is a fairly nice category. In fact, it is complete.
Proposition 5.3.15. BP^ is complete.
Proof. Note that a poset A is a P^ object if and only if there is an order preserving map
a : J0A x M0A â€”> A such that a(d0A x i0A) = kU.
Suppose (Ai)iâ‚¬i is a family of P^ objects and for each i â‚¬ I,
(j)i â€¢ ^ Pi
is an embedding of Ai into a completely distributive lattice The proof of Lemma 3.1.3
implies that the poset product At is a P^object. Moreover, the productinduced map
: nÂ¿â‚¬/ A* â€”â–º nÂ¿â‚¬/ Bi is an embedding which preserves (j, m)optiinum bounds. (Essentially
what is going on is that meets and joins are computed coordinatebycoordinate.)
If /, g : A â€”> B axe PÂ¿, maps, then the proof of Lemma 3.1.3 implies eq(f,g) is
a (j, m)complete subset of A. If A â‚¬ Obj(DP^), then A can be (j,m)embedded into a
completely distributive lattice, so eq(/, g) may also be so embedded.
Thus, any set of objects in BP^ has a product, and any pair of BP^ maps has an
equalizer. By Borceux [6, Volume 1, 2.8.1], this implies BPfa is complete. â–¡
CHAPTER 6
COEQUALIZERS
6.1 Epis and Equalizers in P
The results of this section characterize epis and regular epis in the category P of posets.
The presentation and proofs (except Construction 6.1.2 which is discussed in Meseguer [25])
are the work of the author, but the author believes it likely that they are not new. In all
statements, A and B are arbitrary posets.
Lemma 6.1.1. Let f : A â€”> B be a monotone map. f is epi if and only if f is onto.
Proof. Since the forgetful functor P â€”> Set is faithful, if / is onto then / is epi. For the
converse, suppose b G B \ f(A). Define
Si =i (f(A)n  b)
and
S2 =1 b.
Evidently, both S\ and S2 have the same intersection with f(A), but b G S2 \ Si. Therefore
characteristic functions of B \ Si and B\ S2 are distinct maps (say
Cj, c2 : B â€”> 2
are respectively the characteristic functions of B \ Si and B \ S2) such that Cif = c2/. Both
Q are order preserving, because Si and S2 are decreasing sets. â–¡
Definition and Remarks 6.1.2. The following construction is paraphrased, following MeseÂ¬
guer [25] pp 7374.
If / : A â€”> B is any order preserving map, we may define a preorder â– < on A by
ai 4 a2 <=> /(ax) < /(a2).
82
83
This relation is obviously reflexive and transitive, but because / may not be injective
may not be antisymmetric.
Suppose (A, <) is a poset and ^ is a preorder strengthening <, i.e.,
Ãœi < Ãœ2 => a\ â– < 02.
Then we have an order preserving map a : A â€”â–º A/ â– <. (Recall that A/ â€¢< is the set of
equivalence classes
{x â‚¬ A : x < a and a â– < x}
partially ordered by X.)
Define maps /x, /2 : A â€”* to be equivalent, if there is an isomorphism, i : Bx â€”> B2
such that
B2
commutes.
One verifies that:
1. The maps (f : A * B) and (<)>+ a \ A A/ < are mutually inverse
correspondences between
â€¢ the set (modulo equivalence) of surjective maps / with domain A.
â€¢ the set of preorders on A which strengthen <.
2. For any / : Aâ€”* B, f = ca, where a : A â€”> A/ ^ and c(a(a)) = /(a).
3. Given surjections fi,h'.Aâ€”* BÂ¿, there is a c : B\ â€”> 52 if and only if
/i h
al ^ a2 => Â«1 ^ Â«2
84
Remark 6.1.3. Note that any intersection of preorders is a preorder. Thus, the class of
preorders strengthening < is a complete lattice. We say a set S of ordered pairs generates
^ if ^ is the smallest preordered containing S. Note that if 5 generates â– <, we can describe
< explicitly: a0 â– < an if
â€¢ clq dm or
â€¢ there is a finite sequence ai,a,2,â€™ , aâ€ž_i, such that for all i with 0 < i < n â€” 1,
(Â®Â¿) Â®t+i) Â£ S.
(The first bulleted condition insures that < is reflexive. The second requirement insures
that â– < is transitive; generally, if S is a relation, the relation obtained by applying the second
bulleted item only is called the transitive closure of S.)
One may verify the following construction using 6.1.2.
Construction 6.1.4. Let f,g : A â€”> B be order preserving maps. The coequalizer of
f and g is the quotient of (B, <) by the smallest preorder containing < and C(g, h) â€”
{(gx,hx),(hx,gx) : x G A).
Lemma 6.1.5. Let f : A â€”> B be a surjective order preserving map. The following are
equivalent.
1. f is the coequalizer of some pair g, h : Ao hâ–º A.
2. There exist a poset Ao, and maps g, h : Ao â€”> A such that the preorder â– < on (A, <) is
generated by < and C(g,h).
3. The preorder X on (A, <) is generated by < and K(f) â€” {(x,y) E BxB : f(x) = f(y)}
f. The preorder â– < on (A, <) is generated by < and some equivalence relation.
Proof. (1 <=> 2) is evident from 6.1.4. (3 => 4) is trivial, because K(f ) is an equivalence
relation. We have
C(g,h)CK(f) Â¿
85
because fg = fh and
(x,y) â‚¬K(f) 4=^ f{x) < f(y) and f{y) < f(x).
f f
Since C(g,h) and generate X, K(f) and < generate â– <.
For (4 => 1), let 4 be generated by < and the equivalence relation E C A x A
with the trivial order. Then the projection maps tt\, tt2 : E â€”> A, given by
ni(xi,x2) = Xi
are order preserving. The coequalizer of 7Ti and 7r2 is /. â–¡
Example 6.1.6. Some readers may be surprised by the fact that surjections are not always
Pregular epi. An example of this situation is any map / : 2/Â¿a( â€”Â» 2 from the (trivially
ordered) two point set onto a twopoint chain. Any such map / fails condition 3 from
Lemma 6.1.5.
6.2 Factorization of Maps Using Preorders
This section modifies the preorder factorization, to give a first approximation to
coequalizers in FJm.
Definition 6.2.1. Let Z be a subset system. Z is said to admit congruences if for all posets
A,
Z(A x A) = {S â‚¬ A x A : tt<(5) â‚¬ Z(A),i= 1,2}.
Lemma 6.2.2. Suppose both j and m are subset systems which admit congruences. Let
(A, <) be a poset and â– < be a preorder strengthening <. Then a : A â€”> A/ â– < preserves
(j, m)optimum bounds if and only if < is a FJmsubobject of Ax A.
Proof. We show that q(VxÂ¿) = Vq(xÂ¿) for any (x, : i G 1} â‚¬ j{A). Clearly a(Vx,) is an
upper bound of (a(xÂ¿) : i â‚¬ I}. Since ^ is a P^subobject of A x A, {x{ : i G 1} G j(A) and
for all i, Xi < a,
V{(xÂ¿, a) : i e 1} = (VXj, a)
86
is a member of X; VxÂ¿ < a. Thus, any upper bound of {a(xÂ¿) : i Â£ 1} dominates a(VxÂ¿).
(The proof for meets is identical, so omitted.)
Conversely, if a preserves jjoins and mmeets, and (for example)
{xi'.i â‚¬ /}, {yt : i Â£ 1} Â£ j{A)
and for all i Â£ /, xÂ¿ â– < yt, then VÂ¿x, X V,yt. So ^ is a P^subobject of ,4 x A. â–¡
Definition and Remarks 6.2.3. Suppose j and m admit congruences. Let (A, <) â‚¬
Obj(P^). If ^ is a preorder strengthening < and simultaneously a P^subobject of A x A,
then we say â– < is a (j, m)pocongruence.
Pocongruences of P^objects behave very much like preorders of posets.
1. The maps (/ : A â€”â™¦ B) and (^) a : A â€”> A/ â– < are mutually inverse
correspondences between
â€¢ the set (modulo equivalence) of surjective maps /, which preserve (j, m)optimum
bounds and have domain A.
â€¢ the set of (j, m)pocongruences on A.
2. For any / : A â€”> B, f = ca, where a : A â€”> A/ â– < and c(a(a)) = /(a). If / preserves
(j, m)optimum bounds, so do c and a.
3. Given (j,m)optimum bound preserving surjections fi,f2'Aâ€”>Bl, there is c : B\ â€”â–º
Z?2 if and only if
A Si
ai di a2 a,i ^ Ã¼2
Consideration of Lemma 6.2.2 shows that if / : A â€”> B is surjective and S Â£ j(.4)
(resp. 5 Â£ rn(A)) then f(S) has a supremum (resp. infimum) in B. Moreover, sets of the
form f(S) for S Â£ j(A) (resp. S Â£ m(A)) are the only sets whose optimum bounds are guarÂ¬
anteed to exist. Thus, we are motivated to offer the following conditions on subset systems
to guarantee surjective (resp. Pregular epi) images of PÂ¿, objects are (j, m)complete.
87
Definition and Remarks 6.2.4. Let Z be a subset system. We say Z preserves surjections
if whenever / : A â€”> B is a poset surjection,
Z(B) = {f(S):SeZ(A)}.
Similarly, we say Z preserves regular epis if whenever / : A â€”* B is a poset regular
epimorphism,
Z(B) = {f(S) : 5 â‚¬ Z(A)}.
Since poset regular epimorphims are surjective, any Z that preserves surjections also
preserves regular epimorphisms.
Any cardinal k, the associated subset system k preserves surjections. For if f : A â€”> B
is surjective, then / is split epi (in Set). Hence, there is s : B â€”> A (not necessarily order
preserving) such that idg = fs. Therefore, any set S C B with cardinality less than n is the
image of some set s(S) C A. The following remark provides an example of a subset system
that does not preserve surjections or regular epis. The author currently does not know of
any subset system that preserves regular epis without preserving surjections, but it seems
likely such a subset system exists.
A more interesting question, which the author also cannot currently answer is: â€œdo
there exist subset systems other than cardinals which preserve regular epimorphisms?â€
Remark 6.2.5. Let j = dir be the subset system which selects all upward directed subsets
of a poset. Let N denote a countable disjoint union of twopoint chains. (See drawing below
for help visualizing, and to fix notation.)
to t\
bo b\
Evidently N is jjoin complete.
88
Consider the function / : N â€”Â» N, where N = {0,1,2, â€¢â€¢â€¢} with the usual order,
defined by /(fcÂ¿) = i (for i > 0) and /(ÃÂ¿_i) = i (for i > 1). Speaking roughly, â€œ/ stacks
the two point chains.â€ / is regular epi as a map of posets. Moreover, ^ is a jsubobject of
N x N, because / preserves all existing joins. But N is not jjoin complete, because N is
upward directed but has no join!
Because of the pocongruence factorization for (j, m)bound preserving maps (outlined
in 6.2.3) , we have the following first approximation to the coequalizer in P^. The only thing
stopping map a : B â€”> B/ < (described below) from actually being a coequalizer in P^ is
that B/ â– < is generally not (j, m)complete.
Construction 6.2.6. Suppose f,g:Aâ€”> (B, <) are Â¥Jm maps, and C(f,g) is defined as in
6.I.4â– Let X be the smallest pocongruence containing < and C(/, g). Then a : B â€”> B/ X
has the following universal property: if h : B â€”> C preserves (j, m)optimum bounds and
hf = hg, then there is a unique map i : B/ â– < â€”* C such that h â€” ia.
Since the quotient map a : B â€”> B/ ^ is a surjection we have 
Corollary 6.2.7. Continue with the notation from Construction 6.2.6. If j and m preserve
surjections, then B/ < is complete. Thus, a : B â€”â–º B/ < is the coequalizer of f and
g in FJm. In particular, if j = k and m â€” A are cardinality subset systems, then a is the
coequalizer.
Corollary 6.2.8. DPâ€œ has coequalizers.
Proof. The preceding shows that any pair f,g:Aâ€”>Be P^ has a coequalizer. Suppose
B is completely distributive. Since a : B â€”> coeqPoo (/, g) is a surjection which preserves all
meets and joins, coeqp* (/, g) is completely distributive. In particular if both A and B are
completely distributive, then coeqpâ„¢(f,g) is also the coequalizer in DPâ€œ. â–¡
6.3 Factorization of Meetsemilattice maps
Several simplifications occur describing coequalizers in PÂ¿, if the objects have a meetÂ¬
semilattice structure, i.e., m> uj.
89
Lemma 6.3.1. Let f : A â€”> B be a meetsemilattice map. f is Fregular epi if and only if f
is surjective.
Proof. Since Pregular epimorphisms are always surjective, one implication is trivial. For
the other, suppose f : A â€”> B is a surjection preserving binary meets. Then
x k y <=> f(xAy) = f(x).
Thus we have the sequence x,x A y, y, where f(x) = f(x A y) and x Ay < y. By Lemma
6.1.5, this proves / is regular epi. â–¡
As noted in the proof, the relation â– < is completely described by K(f). If objects
have a meetsemilattice structure, one may use congruences (equivalence relations that are
simultaneously (j, m)subalgebras) rather than the more complex pocongruences.
Finally we have:
Theorem 6.3.2. Suppose m > to and j and m preserve regular epis. Let f : A â€”> B be a
P]m map. The following are equivalent.
1. f is surjective,
2. f is a quotient by some congruence K,
3. f is P regular epi,
4. f P^regular epi,
5. f is Ff,extremal epi.
Proof. The basic facts about pocongruences show (1 2). The preceding lemma shows
(1 <=> 3). (4 => 5) holds in any category.
To show (2 => 4), assume / : A â€”> A/K. Since K C A x A, we have projection
maps 7Ti,n2 : K â€”> A. f is plainly the coequalizer of 7ri,7r2.
90
(5 => 2) holds because we may factor any map / through its associated pocogruence.
If / is extremal epi, then / must coincide with the quotient by its associated pocongruence.
â–¡
Corollary 6.3.3. Let j and m be subset systems which preserve regular epis. Let m > w.
The forgetful functors Up : FJm â€”> P and USet : PÂ¿, ~1' Set preserve and reflect regular epis.
Lemma 6.3.4. The forgetful functor Um^ : DP^ â€”> Set reflects kernel pairs.
Proof. One verifies that the kernel pair of a map / : A â€”> B (calculated in DPÂ¿J consists of
the projections (from Ax A) restricted to the set
{(x,y) â‚¬ A x A : f(x) = f(y)}.
â–¡
Applying Theorem 3.3.5, one obtains the following.
Corollary 6.3.5. Let j and m be subset systems which preserve regular epis. Let m > uj.
The forgetful functor Uset 'â– HMEâ€, â€”> Set is monadic.
6.4 Coequalizers in DPÂ¿,
Theorem 6.4.1. For any pair g,h : A â€”> B â‚¬ IMP^, the coequalizer coeq(
Proof. The construction of the coequalizer of maps g and h is illustrated by the following
diagram. The maps on the diagram will be defined below.
~D{C)
91
Let a : B â€”> B/ < be the map given by Construction 6.2.6. Let / : B â€”> C be any
BP^ map such that fg = fh. The existence and uniqueness of / is given by 6.2.6.
To construct the leftmost square, we apply the free completely distributive complete
lattice functor D = lÃD; rÂ¡ is the natural transformation which injects a poset H into DH.
To construct the rightoutside square, we note that (oo, oo)pocongruences are closed
under intersection. So there is a smallest (in the sense that it makes the fewest possible
identifications) (oo, oo)quotient
k : D(B/ â– <) â€”> E(B/ X)
such that k(r](B/ X))a preserves all (j, m)optimum bounds. Similarly, define
Â£ : D(C) > E(C)
to be the smallest quotient such that Â£(rÂ¡C)f preserves all (j, m)bounds. Since
Â£D(J)(V(B/ l))a = Â£(VC)fa = Â£(rjC)f,
preserves all (j, m)bounds we have the induced map E(f).
Define Eo to be the smallest BP^subobject of E(B/ â– <) through which k(rj(B/ ^))
factors. (By Proposition 5.3.15, the intersection defining Eo exists.) The maps j : B/ < â€”>
Eq and E0 â€”> E(B/ X) are obtained by factoring (rÂ¡(B/ <))a) through E0.
By construction, the map Â£(rÂ¡C) : C â€”> E(C) has the universal property that any map
(j, m)optimum bound preserving : C â€”* B, with B â‚¬ Obj(BP^) factors as (f) = tÂ£(rÂ¡C) for
a uniquely determined map t. Since C â‚¬ Obj(BP^r[), there is such a (j) which is Pextremal
mono. Thus, Â£(r]C) is Pextremal mono. Therefore, the P(epi, extremal mono) factorization
of Â£(r)C) produces the factorization Â£(jgC) = qm. By construction, m is both Pepi and
Pextremal mono; so m is a Pisomorphism. Moreover, q : C â€”> E(C) is necessarily the
smallest BP^subobject of E(C') through which Â£{gC) factors.
We claim that coeq(g,h) = Eo via j : B â€”> Eo. It suffices to show the existence
of a unique compatible map * : Bo â€”Â» C. The commutativity of the largest rectangle
92
in the diagram above implies that k(rÂ¡(B/ ;<)) factors uniquely through the BP^object
Â£'(/)_1(C') C E(B). Because E0 is the smallest BP^object through which k(r](B/ :<))
factors, Eq C E(f)~1(C). This insures the existence and uniqueness of *. â–¡
CHAPTER 7
(j,m)SPACES
This chapter studies spaces obeying a Tostyle separation axiom. Section 7.1
defines spaces and describes a functorial Galois connection, which specializes to Galois conÂ¬
nections between DP^ and SJm. Section 7.2 develops a convenient description of epimorphisms
in SJm, which generalizes a known characterization of epimorphisms of Tospaces. Section 7.3
gives constructions of limits, similar to those for topological spaces. Section 7.4 describes
quotient maps, and characterizes extremal and regular epis as quotient maps.
The last two sections are related to the problem of finding reflections in SJm] Section 7.5
gives the flat spectrum (co)reflection on spatial objects  which is equivalent to a reflection
on SÂ¿j. This reflection on spaces is a generalization of the Tireflection of topological spaces.
Last, but not least, Section 7.6 partially describes the epicomplete SJrn objects; the
description is complete for T0spaces. Epicomplete Tospaces are chains with the specializaÂ¬
tion order. Products of epicomplete T0spaces are not epicomplete, so there is no functorial
epicompletion in the category of Tospaces.
As mentioned in the introductory chapter, the research leading to this dissertation
began in an attempt to find reflections and coreflections in categories of generalized topoÂ¬
logical spaces. After proving the results of this chapter, and reading Meseguer [25], the
author realized that additional assumptions where required on subset systems to insure that
subspaces could be reasonably defined. This realization prompted much of the thought sumÂ¬
marized in Chapter 6  in particular Section 6.3; the author wanted to find when the theory
of this chapter was valid. These considerations, and construction of free DP^objects, beÂ¬
came the main focus of the dissertation. However, this state of affairs leaves many questions
concerning reflections and coreflections in SJm untouched.
93
94
This chapter assumes slightly more background then the rest of the text. Closure
operators are used without comment. If c is a closure operator,
fix(c) := {x : c(x) = x}.
7.1 Spatial/Sober Functorial Galois Connection
Definition and Remarks 7.1.1. Let APos denote the category of whose objects A consist
of
(obi) an underlying partially ordered set, denoted A,
(ob2) a family of designated subsets 3 A, such that \J 5 exists for all
Se3A,
(ob3) a family of designated subsets MA, such that f\ S exists for all
S â‚¬ MA.
Items (ob2) and (ob3) of the data defining a member of Obj (APos) will be referred to as,
the signature of A. Note that a given poset A may have several possible signatures. A
APosmap 4> 'â– A â€”> B is a function A â€”* B such that:
(mapO) cj) is monotone, i.e., a < b e A => 4>{a) < (f>(b).
(mapl) 4>[3A] := {(5) : 5 G 3A} C 0P, and VS â‚¬ 3A, S) = \J (f>{S)
(map2) (/)[MA\ C MB, and VS G MA, 0(A S) = A 4>{S).
The category AS has objects (X,D(X), E(X), A(X)) where A is a set, O(X) is a family of
subsets of A, and E(A), A(A) are families of subsets of D(A), such that (Â£>(A), E(A), A(A))
is a APosobject, in which the optimum bound operations  meet and join  are the set theÂ¬
oretic operations of intersection and union. AS maps are functions f : X â€”* Y such that
/1 : 0(Y) â€”> Â£>(A) is a APosmap.
Definition and Remarks 7.1.2. We describe the functorial Galois connection between AS
and APos.
95
1. D : AS â€”â™¦ APos is the contravariant functor which sends a space X to the APosobject
(O(X), Â£(JA), A(A)), and a map /, to the associated inverse image map, which sends
U G D(Y) to f~\U) G D(X).
2. A prime filter on A G Obj (APos) is a set P C A such that
(pFill) P is increasing, that is, if a G P and a < b then b G P.
(pFil2) P is 0inaccessible; that is, if S G 0A and \J S G P, then Sf]P is
nonempty.
(pFil3) P is Mclosed; that is, if S G MA and S C P, then f\ S G P.
For A G Obj (APos), there is a natural bijection between prime filters, prime ideals (decreasÂ¬
ing, 0closed and Minaccessible subsets of A) and characters A â€”* 2 G AS: if x : A â€”> 2 is a
map, then x1(0) is a prime ideal and x_1(l) is a prime filter, x can be recovered either from
x1(0) or x1(l). For a â‚¬ A, let coz(a) denote the collection of all prime filters containing a.
1. Let 'LA denote the collection of all prime filters of A. 'LA is a ASobject if we define
D(^(A)) = coz[A] := {coz(a) : a G A},
E(A) := coz[[0A]] := {{coz(a) : a G 5} : S G 0A}, and
A (A) := cozfpVtA]].
By definition of maps in AS, 'L extends to a contravariant functor.
2. In Banaschewski and Bruns [3] it is verified that >L and D are adjoint on the right,
where the unit natural transformations have components
VÃa) = coz(a)
and
ex(x) = {U G 0{X) :xeU}
96
In verifying the adjunction, one shows that
coz (V*) = U{ coz(a) : a G sj
for any S â‚¬ 8(A), and
coz (A5)=n{ coz(a) : a G 5,
for any S G JVt(A).
3. We also follow Banaschewski and Bruns [3] in defining SpAPos to be the full subcateÂ¬
gory of APos containing all A such that rÂ¡,\ is an isomorphism, and So AS to be the full
subcategory of AS containing X all X such that ex is an isomorphism. These categories
are said to consist of spatial posets and sober spaces (respectively). Banaschewski and
Bruns [3] show that SpAPos is ontoreflective in APos and that So AS is reflective in
AS.
In order to apply the results of this section, we identify with the full subcategory
of APos, containing all (j, m)complete objects with appropriate signature, i.e.,
Obj(Fm) = {A G APos : 8A = j(A), JVC A = m(A)}.
Rather than AS, we discuss subcategories that roughly correspond to the Pjn categories.
We use the following notations: given X G Obj(AS) and x G X, ex(x) denotes the class of
all open sets (members of Â£>) that contain x. We refer to members of Obj(AS) as spaces.
We assume all spaces satisfy a separation axiom, which generalizes the usual T0 axiom.
Namely, we require:
(sep) Vx, y G X, e(x) = e(y) =â–º x = y
As in general topology, whenever the axiom (sep) holds for X, the soberification map ex is
injective.
Define SJm by
Si = {X G AS : D(X) G PJm, (sep) holds for X}.
97
Remark 7.1.3. To strengthen the analogy between topological spaces and Â§Jmobjects, we
assume that the subset systems j and m preserve regular epimorphisms.
The functorial Galois connection of Banaschewski and Bruns [3] restricts to a duality
between and SpÂ¥3m. Because of (sep),
e : idÂ¡.j â€”> DÃ'
Â£>m
is a monoreflection on S^.
7.2 The Skula Topology and Extremal Monos
Assume j and m preserve regular epimorphisms. If S C X â‚¬ Obj(S^), we may put a
S^structure on S by defining
D(S) = {U n S : U â‚¬ D(X)}.
By analogy with general topology, we call S with this S3m a (j, m)subspace. Because the
poset map dual to the inclusion S C X is onto, and j and m are subset systems that preserve
regular epis, D(S) is (j, m)complete.
Prompted by the treatment of similar problems in To and sober topological spaces,
we define a topology on (underlying sets) of objects from S^,; see Skula [30], We define the
Skula topology on A â‚¬ Obj(S^) to be the topology with base
D{X) U {X \ U : U â‚¬ D(X)}.
It may seem somewhat confusing to put an actual topology on a generalized topological
space, but it is also useful. Corollaries 7.2.2 and 7.2.3 identify epimorphisms and extremal
monomorphisms using the Skula topology. Without it, the results become more cumbersome
to state and less intuitive.
Lemma 7.2.1. The Skulaclosure of S C X, denoted b(S), is given by the formula
b(S) = {x â‚¬ X : VGX, G2 â‚¬ D(X), G1n5 = G2n5=>x^ GxXG2}
98
Proof. Let b'(S) denote the right hand side of the given formula. The operator b' is order
preserving, and increasing. If S C U, we have a trace map from 0(U) â€”> O(S), given
by tr(A) = A fl S. Note that b'(S) is the largest subspace of X so that the trace map is
injective. Thus, b' is also idempotent, and, in fact, a closure operator. It now suffices to
show fix(h') = fix(6).
We first show fix(b') C fix(6). Suppose b'(A) = A, and let U be the complement of A
and x E U. We will give a Skulaopen neighboorhood of x contained in U. If x Â£ b'(A), then
we have Gi, G2 G O(X) with x E Gi\G2 and Gx fl A = G2 fl A. Evidently, Gi fl (X \ G2)
is a Skula neighboorhood of x contained in U.
For fix(6) C fix(6'), it suffices to show SD(X\T) is b' closed, whenever 5,T E D(X).
Let x $ b'(S n{X\ T)), with S,T â‚¬ O(X). Then we have GUG2 E D(X) with
G1nsn(X\T) = G2nsn(X\T)
and x E Gi\ G2. It follows that x S n (X \ T). â–¡
Corollary 7.2.2. Let j and m be subset systems which preserve regular epis. A map f :
X â€”> Y E is epi if and only ifb(f(X)) = Y.
We say f : X â€”* Y is a sub space embedding, if / corestricted to f(X) 
considered as a (j, m)subspace of Y  is an Â§Jm isomorphism.
Corollary 7.2.3. Let j and m be subset systems which preserve regular epis. A map f :
X â€”Â» Y is extremal mono if and only if f is a (j,m)subspace embedding and b(f(X)) â€”
/(*)â€¢
Proof. Suppose / is extremal mono. We may factor / = X â€”* b(f(X)) â€”â–º Y; the first
factor is epi, ergo an isomorphism. Conversely, consider the map b(f(X)) C Y. If this
factors as gs, where s is epi with codomain Z, we can factor g as Z â€”* b(g(Z)) C Y. The
composition b(f(X)) â€”> Z â€”> b(f(X)) is an epi containment of subspaces of Y. By 7.2.2,
b(f(X)) â€” b(g(Z)), so s is an isomorphism. â–¡
99
Corollary 7.2.4. Let j and m be subset systems which preserve regular epis. Suppose X 6
Obj(SoSÂ¿J is sober. Then S C X is sober if and only ifb(S) = S.
Proof. Since SoSJm is monoreflective in and Skulaclosed subobjects are extremal, if
b(A) = A, then A is sober. For the converse, suppose A is sober. We have the epi conÂ¬
tainment A C b(A) of sober spaces. Since the containment is epi, its dual is injective. Since
it is a containment, the dual is surjective. We now apply the fact that bijective maps in P^
are isomorphisms. â–¡
Lemma 7.2.5. Let f,g:A+Be DP^ and
c = coeqmin(f,g) : B â€”*C.
The SpP3mcoequalizer of (f,g) is obtained by following c with the spatial reflection.
Theorem 7.2.6. Let j and m be subset systems which preserve regular epis. Suppose f :
X â€”Â» Y G SoSf is a map of sober objects, with dual
1. f is extremal mono.
2. (Â¡> is extremal epi.
3. f is a subspace inclusion.
j. f is a quotient map (in P3m) such that the codomain is spatial.
5. f is onto.
6. f is regular epi.
7. / is regular mono.
Proof. (1 o 2) and (6 Â«=> 7) hold because SoSJm and SpF3m are dual. (2 o 3) follows from
7.2.3 and 7.2.4. (3 =>â€¢ 4) holds because the SbS^structure on a subspace is defined by a
P^congruence. (6 4^ 4 Â«=> 5) holds by 6.3.2 and 7.2.5. (7=> 1) holds in any category; see
Lemma 2.1.9. â–¡
100
7.3 Computing Limits
We describe how to calculate limits in Â§Jm. To this end we prove a result analogous
to the fact in general topology that any family of subsets generates a smallest topology.
Lemma 7.3.1. Let X be a set. If $ is a collection of families of subsets of X, such that
each t G $ is closed under junions and mintersections, then a := f'5 is also closed under
junions and mintersections.
Proof. For any r G 5, the inclusion map from a â€”> r preserves subset inclusion. Thus, for
any r 6 J, j(o) C j(r) and m(cr) C m(r), so if S G j{cr) (resp. S G rn(a)) then JS â‚¬ t
(respectively, p 5 G r). Hence a = p 5 is closed under unions of jfamilies and intersections
of mfamilies. â–¡
If a is a family of subsets of X, we consider 5, the collection of all (j, m)complete
families of sets of X that contain a; f] j? is called the (j, m)complete family generated by a.
Definition and Remarks 7.3.2. Defining S^products. Consider a set {XÂ¿} C Obj(S^)
The underlying set of X is the cartesian product nXÂ¿, with canonical projections 7r, :
X â€”* Xi. O(X) is the (j, m)complete family generated by sets of the form R~X{U) where
U C O(Xi).
1. The object X thus defined satisfies (sep): if x, y G X and x ^ y, there exists i such that
7ri(x) 7Ti(y). Since satisfies (sep), there is U G O(Xi) with U fl {/xix,'Kiy}\ = 1.
Thus there is V G O(X) (namely V = n~l(U)) with V fl {x,y} = 1.
2. X is the categorical product of (XJ: since the underlying set of X is the cartesian
product of {Xj}, whenever we have an object Y G Obj(S^) and functions :Y â€”> XÂ¿,
there is a unique function / : V â€”Â» X so that for all i, Rif = fi. If each /Â¿ is an
S^jmap, then because of our definition of D(X), / is also an S^rnap.
Definition and Remarks 7.3.3. S^equalizers. Let j and m be subset systems which
preserve regular epis. Let f,g : X â€”> Y be a pair of maps in S^. The set E := {x G X :
101
f(x) = g(x)} has a natural S^structure 0(E) = {U fl E : U â‚¬ D(X)}. The relation R on
D(X) given by
urv u n E = v n E
is a P^congruence because
f)uinE= (fÃ¼i)nÂ£7
i i
and
(J^nÂ£=((Jt/Â¿)nÂ£
i i
for any set E and family of subsets (t/j). Since O(E) is the quotient of O(X) by a P^
congruence, and j and m preserve regular epis, O(E) is (j,m)complete. E satisfies (sep),
because X does. If h : Z â€”> X equalizes /, g, then the image of h must be contained in E.
Since h is an S^map, the inclusion of the image of h in E is also an S^map.
7.4 Quotients, Extremal and Regular Epis
Definition and Remarks 7.4.1. Let X G Obj(S^), Eqv(X) denote the set of equivalence
relations on X, and denote the set of subfamilies of O(X). We consider equivalence
relations to be sets of ordered pairs. If R â‚¬ Eqv(X) and x â‚¬ X, we use the notation
R(x) = {y Â£ X : (x, y) G R} for the equivalence class of x.
1. If R e Eqv(X), we say U â‚¬ O(X) is Rsaturated if (x, y) â‚¬ R and x G U imply that
y â‚¬ U\ let R* denote the family of all R saturated members of O(X). If R\ C R2, then
r*2 c r*v
2. If a â‚¬ ?P, we define an equivalence relation a* on X by (x, y) G a* if and only if for
each U G (J, we have x G U y G U. Note that cq C &2 ^ av
3. If a G *}3. then each member of a is cr*saturated. In symbols, a C a**.
4. Suppose R G Eqv(X). U G D(X) is Rsaturated if and only if for all (x,y) G R
x G U <=$â– y G 17. Thus, R C R**.
102
5. To summarize, * gives a Galois connection between Eqv(X) and *}3.
Given R â‚¬ Eqv(X), we define an S^structure on the set X/R of equivalence classes, by
defining 0(X/R) to be the family {{R(x) : x G U} : U G R*}. Since arbitrary unions and
intersections of saturated sets are saturated and since the inclusion of R* in D(X) is order
preserving, 0(X/R) is (j, m)complete.
We omit the proofs of the following lemmas, which are straightforward verifications.
Lemma 7.4.2. If f : X â€”*Y G Map(S^), R G Eqv(X), and f is constant on each equivaÂ¬
lence class R(x), then f factors as X â€”> X/R â€”> Y, where X â€”Â» X/R is the quotient map,
and X/R â€”> Y is defined by R(x) f{x). For all U G D(T), f~l(U) is Rsaturated and in
0(X).
Lemma 7.4.3. Let R\,R2 â‚¬ Eqv(X) with quotient maps : X â€”> X/Ri There is a
quotient map q : X/R\ â€”* X/R2 such that q2 = qq\ if and only if R\ C R2.
Lemma 7.4.4. For R G Eqv(X) the following are equivalent:
1. R** = R,
2. there is a Â£ p such that R = a*,
3. X/R satisfies (sep).
Proof. The equivalence of the first two conditions is trivial. If R = a* and R(x) ^ R(y),
then there exists U G a such that U fl {a;,y} = 1. So (2 =>â€¢ 3).
(3 =$â– 1) Since R C R**, we have a quotient map q : X/R â€”â–º X/R** as described,
in the preceding lemma. It suffices to show that q is injective. If R(x) / R(y) and X/R
satisfies (sep), then there is U G R* with JC7 n {x,y}\ â€” 1. Thus, R**(x) Â± R**{y) â–¡
Corollary 7.4.5. If X is a To topological space and R G Eqv(X), then X/R is To if and
only if R â€” R**.
103
Lemma 7.4.6. If f : X â€”> Y sS^ and R G Eqv(X) is defined by
R = {(*>2/) : fix) = f{y)}
then R** = R
Proof. We use Lemma 7.4.2. Since Y satisfies (sep), and for all U G 0(Y), f~x{U) is R
saturated, X/R satisfies (sep): if x, y G Y and x ^ y, then we have U G 0(F) containing
exactly one member of {x, y}. â–¡
Definition and Remarks 7.4.7. Coequalizers in Â§Jm. If /, g : X â€”> Y, let a â€” {U G
D{X):f\U) = g\U)}.
The coequalizer of /, g is the quotient map c : Y â€”â–º Y/o*. First, one shows that
cf â€” eg: if c(/(x)) ^ c(g(x)) there is U G a which contains exactly one of {f{x),g(x)}.
But f~l{U) = 5_1(f/), since U G cr; this contradicts the fact that x is in exactly one of
/1(t/), <71(i7). This contradiction shows that cf = eg.
Suppose h : Y â€”> Z and hf = hg. We claim that h â€” we for a unique w. Note that
h factors as Y â€”* Y/R â€”* Z, where R = {(x,y) : h(x) â€” h(y)}. By Lemma 7.4.3, it suffices
to show that a* C R. If h(x) Â± h(y), then we have U G 0{Z), which contains exactly one
member of {h(x),h(y)}. Since hf â€” hg, h~l{U) G a, so (x, y) & a*.
We say / : X â€”â–º Y is a quotient map, if there is an isomorphism i : Y â€”> X/R, and
an equivalence relation R = R** such that R(x) â€” if{x), for all x.
Proposition 7.4.8. For f : X â€”+ Y G Â§Â£,, the following are equivalent:
1. f is a quotient map,
2. f is regular epi,
3. f is extremal epi.
Proof. (1 => 2). Suppose that / is a quotient map, with associated equivalence relation
R = R** and family of Rsaturated sets, R*. Define a S^object Z to have underlying set
104
jJ{C x C : C 6 X/Rj  a disjoint union  and define O(Z) = 7(Z). D(Z) is complete
(hence (j, m)complete) so Z G Obj(S^); moreover any set map from Z â€”> X is a S^map.
We now define two maps g,h : Z â€”> X by g(x, y) = x and h(x, y) = y. By construction, / is
the coequalizer of g, h as computed in Set. Since / is also a S^quotient map, / is also the
coequalizer of g, h in Sjn.
(2 => 3) holds in any category  [13, 17.11].
To show (3 =Â£â– 1) suppose / is extremal epi. We may factor / into a quotient map
q followed by an injection i, using Lemma 7.4.2. Since / is epi, then i is also epi; it now
follows, by the extremal property of / that i is an isomorphism. So / is a quotient map. â–¡
7.5 Flat Spectra
Let j and m be subset systems which preserve regular epis. We now examine a
monocoreflective category of SpÂ¥3m. To define the category, we need to put a partial order
on A, for A â‚¬ Â¥Jm. (Recall that 7A is the set of functions A â€”> 2 which preserve order and
(j, m)optimum bounds.) We think of x, y G Ã'A as characters x, y : A â€”* 2. We define x < y
to mean Va 6 A.x(a) < y(a). Clearly, x < y if and only if e(x) = {a 6 A : x(a) â€” 1} C e(y).
The relation < is antisymmetric by virtue of the fact that our spaces satisfy (sep).
Now we define to be the full subcategory of 5 G containing all A with the
property that for all x, y G 4/(^4), x < y implies x â€” y. To show FPm monocoreflective, we
will use the category theoretic dual of the following theorem.
Theorem 7.5.1. [13, 37.1] Suppose A is a cowellpowered complete category, and 3 is a full,
isomorphism closed subcategory. B is epireflective if and only if it is closed under products
and extremal subobjects.
Proposition 7.5.2. is monocoreflective in SpÂ¥3m.
Proof. SpFJm is cowellpowered and cocomplete, because it is wellpowered, complete and has
a coseparator 2. [13, 23.14] We show that F^ is closed under coproducts. Suppose the
105
coproduct JJ Ai has characters x < y. Since x ^ y, there is an index i such that xpi Â± ypi.
But x\ii,yiii are characters of Ai. Thus, some At has characters x < y.
We show that is closed under extremal epis. Suppose A has flat spectrum and
A â€”> A/K is a quotient. If A is nontrival, then it cannot have constant characters. Thus, if
A has flat spectrum, all characters are onto. Now we use 6.3.2 to conclude that characters
of A/K are precisely characters of A whose kernel contains K. It follows that A/K has flat
spectrum. â–¡
Remark 7.5.3. The theorem would remain true if we redefined to be a full subcategory
of containing flat spectrum objects. The rationale for not using this definition of FÂ¿, is
that nonspatial flat spectrum objects behave rather differently than spatial objects. Without
assuming A spatial, the spectrum gives incomplete information. For instance, there are A
with no characters whatsoever! These objects trivially have flat spectrum. We wish to avoid
this problem.
7.6 Epicomplete objects in SJm
Assume throught this section that j and m preserve regular epimorphisms.
Definition 7.6.1. An object A in any category is epicomplete if any map f : A â€”> B that
is both epi and mono is an isomorphism.
Note: any monoreflection map is epi; see Herrlich and Strecker [13, Section 36],
Soberification is a monoreflection on SJm. Therefore, any epicomplete object in SJm must be
sober. We now examine which sober spaces are epicomplete.
We recall the partial order on characters A â€”* 2 that was used in the preceding
subsection. For a sober space, the points x â‚¬ X are in bijective correspondence to S^maps
D(X) â€”> 2. If A is a space, x < y means e(x) C e(y). Note that if U 6 O(X), x G U and
x < y, then y â‚¬ U] in other words, each U 6 O(X) is a increasing subset of X. Also note
that any continuous map preserves this order.
Lemma 7.6.2. Suppose X â‚¬ Obj(S^) with to C j fl m.
106
1. X is a chain if and only if D(X) is a chain.
2. If X is not a chain, then there is a (j,m)structure r on X, such that r C D(X) and
(X,t) satisfies (sep).
Proof. 1. If X is a chain, then the family 3 of all increasing subsets of X is a chain (ordered
by inclusion). Since Â£>(X) C 3, Â£>(X) is a chain. Conversely, if Â£)(X) is a chain, then the
set of prime ideals of D(X) is a chain, so X is a chain.
2. Suppose D(X) is not a chain. Then we have incomparable U,V â‚¬ D(X). Let
r = {W E D(X) : W C U or U C W}. One may easily verify that r is closed under all
unions and intersections which exist in O(X); since j(r) C j(Q(X)) and m(r) C m(D(X)),
this shows that r is a (j, m)complete.
We now verify that (X, r) satifies (sep). Suppose x, y â‚¬ X are distinct points; without
loss of generality, there is G E O(X) such that x â‚¬ G and y & G. If x E U then G fl U E r
separates x and y. If x $ U and y $ U, then G U U separates x and y. Lastly, if x Â£ U, and
y G U, then U separates x and y. â–¡
Theorem 7.6.3. Let X G Obj(5oS^), ui C j n m. We have the implications (i) => (ii) =>
(in) =Â» (iv). If j â€” oo, then (iv) => (i).
(i) X is epicomplete
(ii) If f : X â€”>Y is SoPmmonic, then f(X), b f(X), and X are all homeomorphic.
(in) There is no (j,m)structure, r on X, such that r C O(X) and (X,t) satisfies (sep).
(iv) X is a chain satisfying (Hi).
Proof. (i=> ii). Suppose / : X â€”> Y is monic. By corestriction of / we obtain a S^map
X â€”> f(X). Composing with the inclusion map of f(X) in bf(X) we obtain a monic epic
SoPm map with domain X\ (i) forces this map to be an isomorphism.
(ii => iii). Suppose r is a (j,m)structure on X and r C D(X). The identity map on
X is a continuous bijection (X, O(X)) â€”> (X, r). Following this map by e(xÂ» : (X, r) â€”Â»
107
Ã'T, we obtain a mono epi map (X,D(X)) â€”Â» i'r. (ii) implies that all maps involved are
isomorphisms.
(iii => iv). This follows from Lemma 7.6.2.
(iv => i). When j = oo, we have more information about which (j, restructures
correspond to a given partially ordered set. Suppose (X, <) is a poset and r is a topology
on X such that x < y if and only if each open set containing x also contains y. The union
of the family {U â‚¬ t : x Â£ U} is open and equals {y 6 X : y x}. Thus, r must contain
the weak topology generated by the sets {y â‚¬ X : y x} for x â‚¬ X.
Suppose X is a chain, and / : X â€”> Y is an epic monomorphism. Since / is injective
and preserves order, f(X) is a chain order isomorphic to X. Ergo, the f(X) has a topology
finer than the weak topology induced by the partial ordering. Since the corestriction / : X â€”â–º
f(X) is continuous, the topology on X is finer than the topology on f(X). Thus, X and
f(X) are homeomorphic. Since X is sober, f(X) is sober, which implies b(f(X)) â€” f(X).
Since / is epi, b(f(X)) = Y, so X and Y are homeomorphic. â–¡
REFERENCES
[1] Seminar on Triples and Categorical Homology Theory, Edited by B. Eckmann. Lecture
Notes in Mathematics 80, SpringerVerlag, BerlinNew York (1969).
[2] R. Balbes and P. Dwinger, Distributive Lattices, University of Missouri Press, Columbia,
Mo. (1974).
[3] B. Banaschewski and G. Bruns, The fundamental duality of partially ordered sets, Order,
5 (1988) no. 1, 6174.
[4] M. Barr and C. Wells, Toposes, Triples, and Theories, SpringerVerlag, BerlinNew York
(1985).
[5] J. Beck, Distributive laws, in: Seminar on Triples and Categorical Homology Theory,
Lecture Notes in Mathematics 80, SpringerVerlag, BerlinNew York (1969) 119140.
[6] F. Borceux, Handbook of Categorical Algebra, (3 volumes), Cambridge University Press,
Cambridge (1994).
[7] M. Erne,Zcontinuous posets and their topological manifestation, Applied Categorical
Structures 7 (1999) 3170.
[8] H. Gaifman, Infinite Boolean polynomials, Fund. Math. , 54 (1964) 230250.
[9] A. W. Hales, On the nonexistence of free complete Boolean algebras, Fund. Math., 54
(1964) 4566.
[10] P. R. Halmos, Naive Set Theory, Undergraduate Texts in Mathematics, SpringerVerlag,
BerlinNew York (1974).
[11] H. Herrlich, Regular categories and regular functors, CaÃ±ad. J. Math., 26 (1974) 709
720.
[12] H. Herrlich, Topologische Reflexionen und Coreflexionen, Lecture Notes in Mathematics
78, SpringerVerlag, BerlinNew York (1968).
[13] H. Herrlich & G. Strecker, Category Theory, HeldermannVerlag, Berlin (1979).
[14] K. H. Hoffman, M. Mislove, Free Objects in the category of completely distributive
lattices, in: Continuous Lattices and their Applications, Lecture Notes in Pure and Applied
Mathematics 101, Marcel Dekker, New York (1985).
[15] J. R. Isbell, Atomless parts of spaces, Math. Scand. 31 (1972) 532.
[16] P. Johnstone, Stone Spaces, Cambridge Studies in Advanced Mathematics 3, Cambridge
University Press, Cambridge (1982).
108
109
[17]A. Joyal, M. Tierney, An Extension of the Galois Theory of Grothendieck, Mem. Amer.
Math. Soc. 51 (1984) no. 309
[18]Y. Li, Generalized (S,Incomplete free complete completely distributive lattices generÂ¬
ated by posets, Semigroup Forum, 57 (1998) 240248.
[19]L. E. J. Linton, Applied functorial semantics II, in: Seminar on Triples and Categorical
Homology Theory, Lecture Notes in Mathematics 80, SpringerVer lag, BerlinNew York
(1969) 5374.
[20] L. E. J. Linton, Coequalizers in categories of algebras, in: Seminar on Triples and
Categorical Homology Theory, Lecture Notes in Mathematics 80, SpringerVerlag, Berlin
New York (1969) 7590.
[21] S. MacLane, Categories for the Working Mathematician, Graduate Texts in Mathematics
5, SpringerVerlag, BerlinNew York (1971).
[22] J. Madden, Kframes, Journal of Pure and Applied Algebra, 70 (1991) 107127.
E. G. Manes, Algebraic Theories, Graduate Texts in Mathematics 26, SpringerVerlag,
erlinNew York (1976).
[24]G. Markowsky, Free completely distributive lattices, Proc. Amer. Math. Soc., 2 (1979)
227228.
[25] J. Meseguer, Order completion monads, Algebra Universalis, 16 (1983) 6382.
[26] J. Pare, Absolute coequalizers, in: Category Theory, Homology Theory and their ApÂ¬
plications, Lecture Notes in Mathematics 86, SpringerVerlag, BerlinNew York (1969)
132145.
[27] G. N. Raney, Completely distributive lattices, Proc. Amer. Math. Soc. 3 (1952) 677680.
[28] G. N. Raney, Tight Galois connections and complete distributivity, Trans. Amer. Math.
Soc. 97 (I960) 418426.
[29] G. N. Raney, A sub directunion representation for completely distributive complete
lattices, Proc. Amer. Math. Soc. vol 4 (1953) 518522.
[30] L. Skula, On a reflective subcategory of the category of all topological spaces, Trans.
Amer. Math. Soc., 142 (1969) 3741.
[31] J. W. Thatcher, J. B. Wright & E. G. Wagner, A uniform approach to inductive posets
and inductive closure, Theoretical Computer Science, 7 (1978) 5777.
[32] W. R. Tunnicliff, The free completely distributive lattice over a poset, Algebra UniverÂ¬
salis, 21 (1985) 133135.
BIOGRAPHICAL SKETCH
Eric Richard Zenk was born on February 2, 1977, in Fridley, MN. He lived in AnÂ¬
dover, MN, for the following eighteen years. Then, Eric attended St. Johns University in
Collegeville, MN, where he majored in mathematics and physics. Upon graduating in 1999,
Eric moved to Florida where he studied mathematics at the graduate level at the University
of Florida.
110
I certify
able standards
dissertation for
I certify
able standards
dissertation for
I certify
able standards
dissertation for
I certify
able standards
dissertation for
I certify
able standards
dissertation for
that I have read this study and that in my opinion it conforms to accept
of scholarly presentation and is fully adequate, in scope and quality, as a
the degree of Doctor of Philosophy.
Jorge Martinez, Chairman
Professor of Mathematics
that I have read this study and that in my opinion it conforms to accept
of scholarly presentation and is fully adequate, in scope and quality, as a
the degree of Doctor of Philosophy.
Pham Tiep
Associate Pr
essor of Mathematics
that I have read this study and that in my opinion it conforms to accept
of scholarly presentation and is fully adequate, in scope and quality, as a
the degree of Doctor of Philosophy.
Alexander Dranymnikov
Professor of Mathematics
that I have read this study and that in my opinion it conforms to accept
of scholarly presentation and is fully adequate, in scope and quality, as a
the degree of Doctor of Philosophy.
/L U'A'
Scott McCullough
Professor of Mathematics
that I have read this study and that in my opinion it conforms to accept
of scholarly presentation and is fully adequate, in scope and quality, as a
the degree of Doctor of Philosophy.
ReyÃialdÃ³ Jimenez
Associate Professor of Romance Languages and
Literatures
This dissertation was submitted to the Graduate Faculty of the Department of MathÂ¬
ematics in the College of Liberal Arts and Sciences and to the Graduate School and was
accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy.
August 2004
Dean, Graduate School
LD
1780
20 OA

