1: \documentstyle[11pt]{article}
2:
3: \raggedbottom
4: \setlength{\textwidth}{6.0in}
5: \setlength{\textheight}{8.5in}
6: \setlength{\oddsidemargin}{0.2in}
7: \setlength{\topmargin}{-0.5in}
8: \title{THE GALAXIES OF NONSTANDARD ENLARGEMENTS OF INFINITE AND TRANSFINITE GRAPHS: II}
9: \author{A. H. Zemanian}
10: \date{}
11:
12: \begin{document}
13: \newcommand{\N} {I \kern -4.5pt N}
14: \newcommand{\R} {I \kern -4.5pt R}
15: \newcommand{\Z} {Z \kern -7.5pt Z}
16: \maketitle
17: \baselineskip21pt
18:
19: {\ Abstract --- This report is an improvement of a prior report (Report 813).
20: It sharpens the principal theorems (Theorems 4.2 and 11.2 of Report 813)
21: while simplifying their proofs. There are also several minor
22: changes involving
23: clarifications and corrections of misprints.
24:
25: The Abstract of the prior report remains the same as follows:
26: The galaxies of the nonstandard enlargements of connected,
27: conventionally infinite graphs as well as of connected transfinite graphs are
28: defined, analyzed, and illustrated by some examples.
29: It is then shown that any
30: such enlargement either has exactly one galaxy, its principal one, or
31: it has infinitely many galaxies. In the latter case, the galaxies are
32: partially ordered by their ``closeness'' to the principal galaxy.
33: If an enlargement has a galaxy different from its principal galaxy,
34: then it has a two-way infinite sequence of galaxies that are
35: totally ordered according to that ``closeness'' property.
36: There may be many such totally ordered seqences.
37:
38: Key Words: Nonstandard graphs, enlargements of graphs, transfinite graphs,
39: galaxies in nonstandard graphs, graphical galaxies.
40: }
41:
42: \section{Introduction}
43:
44: In this work we extend the idea of galaxies in the hyperreal line
45: $^{*}\!\R$ to nonstandard enlargements of conventionally infinite
46: graphs and also of transfinite graphs. We stipulate henceforth that
47: every graph considered herein is connected.
48: Since graphs have structures
49: much different from that of the real line $\R$, the enlargements
50: of graphs have properties not possessed by $^{*}\! \R$.
51: The graphical galaxies of those enlargements comprise one
52: aspect of that distinctive complexity. We will show that
53: that any such enlargement has either one galaxy or infinitely many of them.
54: Moreover, just as $^{*}\! \R$ contains images of the real numbers,
55: called the standard hyperreals, as well as hyperreals that are nonstandard,
56: so too may the enlargement $^{*}\!G$ of
57: a graph $G$ contain ``hypernodes,'' some of which are images
58: of nodes of $G$ and others of which are nonstandard hypernodes.
59: In addition, there are ``hyperbranches'' incident to pairs of hypernodes;
60: some of these hyperbranches are images of branches of $G$, but there may be
61: others that are not.
62:
63: The galaxies graphically partition $^{*}\!G$ in the sense the every
64: hypernode belongs to exactly one galaxy, and so too does every hyperbranch.
65: There is a unique galaxy, which we refer to as the ``principal
66: galaxy,'' that contains the standard hypernodes and possibly
67: nonstandard hypernodes as well. In the event that there are
68: infinitely many galaxies, those galaxies are partially
69: ordered according to how ``close'' they are
70: to the principal galaxy. In fact,
71: if there is a galaxy different from the principal galaxy,
72: then there is a two-way infinite sequence of galaxies
73: that are totally ordered according to their ``closeness'' to
74: the principal galaxy. There may be many such totally ordered sequences, but
75: a galaxy in one such sequence may not be comparable to a galaxy in
76: another sequence according to that ``closeness'' property.
77:
78: We speak of ``conventionally infinite'' graphs to
79: distinguish them from transfinite graphs of ranks 1 or higher
80: \cite[Chapter 2]{tgen}, \cite[Chapter 2]{gn}. Sections 2 through 4
81: herein are devoted to the enlargements of conventionally infinite graphs.
82: The results for such enlargements extend to enlargements of
83: transfinite graphs, but in more complicated ways. We show this in
84: Sections 5 through 11, but only for transfinite graphs of rank 1.
85: Results for transfinite graphs of still higher ranks
86: are obtained similarly but in still more
87: complicated ways and with additional complexity in the symbols.
88: For the sake of brevity, the latter results are not included herein,
89: but they may be found in \cite{gal2} as well as in the archive
90: www.arxiv.org in the category ``mathematics'' under ``Zemanian.''
91:
92: Our notations and terminology follow the usual conventions of nonstandard
93: analysis. $\N= \{0,1,2,\ldots\}$ is the set of
94: natural numbers, and $^{*}\! \N$ is the set of hypernaturals.
95: The standard hypernaturals are (i.e., can be identified
96: with) the natural numbers. Also, $\langle a_{n}\rangle$ or
97: $\langle a_{n}\!: n\in\N\rangle$ or $\langle a_{0},a_{1},
98: a_{2},\ldots\rangle$ denotes a sequence whose elements can be
99: members of any set, such as the set $X$ of nodes in a
100: conventional graph $G=\{X,B\}$, where $B$ is the set of branches,
101: a branch being a two-element set of nodes. On the other hand,
102: $[a_{n}]$ denotes an equivalence class of sequences, where
103: two sequences
104: $\langle a_{n}\rangle$ and $\langle b_{n}\rangle$
105: are taken to be equivalent if $\{n\!:a_{n}=b_{n}\}\in {\cal F}$,
106: where ${\cal F}$ is any chosen and fixed
107: free ultrafilter.\footnote{Also called
108: a nonprincipal ultrafilter.} ${\cal F}$ will be so fixed throughout this work.
109: The $a_{n}$ appearing in $[a_{n}]$ are understood
110: to be the elements of any one of the sequences in the equivalence class.
111: At times, we will use the more specific notation
112: $[\langle a_{0},a_{1},a_{2},\ldots\rangle]$. More generally,
113: we adhere to the notations and terminology appearing in \cite{go}.
114:
115: The ordinals are denoted in the usual way: $\omega$ is the
116: first transfinite ordinal. With $\tau\in\N$,
117: the product $\omega\cdot \tau$ is the sum of $\tau$ terms,
118: each being $\omega$.
119:
120: \section{The Nonstandard Enlargement of a Graph}
121:
122: Throughout Sections 2 to 4, we assume that the conventionally
123: infinite graph $G$ is connected and has infinitely many nodes.
124: The definition of a nonstandard graph that we use herein is given in
125: \cite[Section 8.1]{gn}, a special case of which is the ``enlargement''
126: of a graph $G$.
127:
128: Let us define the {\em enlargement} $^{*}\!G$ of $G$
129: here as well in order to remove
130: any need for referring to \cite{gn}. $G=\{X,B\}$ is
131: now taken to be a conventional connected graph having an infinite set $X$
132: of nodes and therefore an infinite set of branches as well, each
133: branch being a two-element set of nodes. Thus, there are no
134: parallel branches (i.e., multiple branches). ${\cal F}$ will denote
135: a chosen and fixed free ultrafilter.
136: ${\bf x}=[x_{n}]$ denotes an equivalence class of sequences
137: of nodes as stated in the Introduction. Specifically,
138: $[x_{n}]$ and $[x'_{n}]$ are in the same equivalence class $\bf x$
139: if $\{n\!: x_{n}=x'_{n}\}\in{\cal F}$.
140: ${\bf x}$ will be called a {\em
141: hypernode}.\footnote{Our terminology should not be confused with that of
142: a hypergraph---an entirely different concept \cite{be}.}
143: Thus, the set of all sequences of nodes from $G$ is partitioned
144: into hypernodes. $^{*}\!X$ denotes the set of hypernodes.
145: If all the elements of one of the representative sequences
146: $\langle x_{n}\rangle$ for a hypernode ${\bf x}=[x_{n}]$ are the same node
147: (i.e., $x_{n}=x$ for all $n$), then ${\bf x}=[x]$ can be
148: identified with $x$; in this case, ${\bf x}$ is called a {\em standard
149: hypernode}. Otherwise,
150: ${\bf x}=[x_{n}]$ is called a {\em nonstandard hypernode}.
151:
152: We turn now to the definition of a ``hyperbranch.'' Let ${\bf x}=[x_{n}]$ and
153: ${\bf y}=[y_{n}]$ be two hypernodes. Also, let ${\bf b}=[\{x_{n},y_{n}\}]$,
154: where $\langle \{x_{n},y_{n}\}\rangle$ is a sequence of pairs of nodes from
155: $G$ such that, for almost all $n$, $\{ x_{n},y_{n}\}$
156: is a branch in $G$; that is, $\{n\!: \{x_{n},y_{n}\}\in B\}\in {\cal F}$.
157: It can be shown \cite[page 155]{gn} that this definition is independent
158: of the representative sequences $\langle x_{n}\rangle$ and
159: $\langle y_{n}\rangle$ chosen of ${\bf x}$ and ${\bf y}$ respectively and
160: that we truly have an equivalence relation for the set of all
161: sequences of branches from $G$. We let ${\bf b}=[\{x_{n},y_{n}\}]$ denote such
162: an equivalence class and will call it a {\em hyperbranch};
163: we write ${\bf b}=\{{\bf x},{\bf y}\}$. Also, $^{*}\!B$ will denote the set of all
164: hyperbranches. If ${\bf x}=[x_{n}]$ and ${\bf y}=[y_{n}]$ are
165: standard hypernodes, then ${\bf b}=[\{x,y\}]$ is called a
166: {\em standard hyperbranch}. Otherwise, ${\bf b}$ is called a
167: {\em nonstandard hyperbranch}.
168:
169: Finally, the pair $^{*}\!G=\{^{*}\!X,\,^{*}\!B\}$ denotes the
170: {\em enlargement} of $G$. It is a special case of a nonstandard graph,
171: as defined in \cite[page 155]{gn}.\footnote{If $G$ were a finite graph,
172: then every hypernode (resp. hyperbranch) could
173: be identified with a node (resp. branch)in $G$, and $^{*}\!G$
174: would be identified with $G$.}
175:
176: \section{Distances and Galaxies in Enlarged Graphs}
177:
178: The {\em length} $|P_{x,y}|$ of any path $P_{x,y}$ connecting
179: two nodes $x$ and $y$ in a graph $G$ is the number of branches in
180: $P_{x,y}$. The {\em distance} $d(x,y)$ between $x$ and $y$
181: is $d(x,y)=\min\{|P_{x,y}|\}$, where the minimum is taken over all
182: paths terminating at $x$ and $y$. In the trivial case, $d(x,x)=0$.
183: $d$ satisfies the triangle inequality, namely, for any three nodes
184: $x$, $y$, and $z$ in $G$, $d(x,y)\leq d(x,z)+d(z,y)$. In fact, $d$
185: satisfies the other metric axioms, too,
186: and the set $X$ of nodes in $G$ along with $d$
187: is a metric space.
188:
189: The metric $d$ can be extended into an internal function $\bf d$
190: mapping the Cartesian product $^{*}\! X\,\times \,^{*}\! X$ into the set of
191: hypernaturals $^{*}\! \N$ as follows: For any ${\bf x}=[ x_{n}]$ and
192: ${\bf y}=[y_{n}]$ in $^{*}\!X$, ${\bf d}$ is defined by
193: \[ {\bf d}({\bf x},{\bf y})\;=\;[d(x_{n},y_{n}]\,\in\,^{*}\!\N. \]
194: By the transfer principle, we have, for any three hypernodes
195: ${\bf x}$, ${\bf y}$, and ${\bf z}$,
196: \begin{equation}
197: {\bf d}({\bf x},{\bf z})\;\leq\; {\bf d}({\bf x},{\bf y})\,+\,{\bf d}(({\bf y},{\bf z}). \label{3.1}
198: \end{equation}
199: From the point of view of an ultrapower construction, this means that
200: \[ \{n\!: d(x_{n},z_{n})\;\leq\;d(x_{n},y_{n})\,
201: +\,d(y_{n},z_{n}\}\;\in\;{\cal F}. \]
202: The other metric axioms, such as ${\bf d}({\bf x},{\bf x})=0$, are obviously
203: satisfied by ${\bf d}$.
204:
205: We define the ``galaxies'' of $^{*}\!G$ as nonstandard subgraphs
206: of $^{*}\!G$ by first defining the ``nodal galaxies.'' Two hypernodes
207: ${\bf x}=[x_{n}]$ and ${\bf y}=[y_{n}]$ are taken to be in the same
208: {\em nodal galaxy} $\dot{\Gamma}$ of $^{*}\!G$
209: if ${\bf d}({\bf x},{\bf y})$ is no
210: greater that a standard hypernatural $\bf k$, that is, if there exists a natural
211: number $k\in\N$ such that $\{n\!: d(x_{n},y_{n})\,\leq\,k\}\;\in\; {\cal F}$.
212: In this case, we say that ${\bf x}$ and ${\bf y}$ are {\em limitedly distant},
213: and we write ${\bf d}({\bf x},{\bf y})\leq {\bf k}$.
214:
215: Let $N_{{\bf x},{\bf y}}$ be the set of all standard hypernaturals that are no
216: less than ${\bf d}({\bf x},{\bf y})$. $N_{{\bf x},{\bf y}}$ is a
217: well-ordered set, and
218: therefore it has a minimum ${\bf k}_{{\bf x},{\bf y}}$.
219: So, we can say that ${\bf x}$ and ${\bf y}$ are in the same nodal galaxy
220: $\dot{\Gamma}$ if ${\bf d}({\bf x},{\bf y})={\bf k}_{{\bf x},{\bf y}}$.
221:
222: {\bf Lemma 3.1.} {\em The nodal galaxies partition the set $^{*}\!X$
223: of all hypernodes in $^{*}\!G$.}
224:
225: {\bf Proof.} The property of two hypernodes being limitedly distant
226: is a binary relation on $^{*}\!X$ that is obviously reflexive and symmetric.
227: Its transitivity follows directly from (\ref{3.1}). Alternatively,
228: we can use an ultrapower argument. Assume that ${\bf x}=[x_{n}]$ and
229: ${\bf y}=[z_{n}]$ are in some nodal galaxy and that ${\bf y}$ and
230: ${\bf z}=[z_{n}]$ are in some nodal galaxy; we want
231: to show that those galaxies are the same. There exist
232: two standard natural numbers $k_{1}$ and $k_{2}$ such that
233: $N_{{\bf x},{\bf y}}=\{n\!:d(x_{n},y_{n})\leq k_{1}\}\in{\cal F}$
234: and $N_{{\bf y},{\bf z}}=\{n\!:d(y_{n},z_{n})\leq k_{2}\}\in{\cal F}$.
235: Since $d(x_{n},z_{n})\,\leq\,d(x_{n},y_{n})+d(y_{n},z_{n})$,
236: \[ \{n\!: d(x_{n},z_{n})\leq k_{1}+k_{2}\}\;\supseteq\;
237: N_{{\bf x},{\bf y}}\cap N_{{\bf y},{\bf z}}\;\in\;{\cal F}. \]
238: So, the left-hand side is a set in ${\cal F}$. Thus, ${\bf x}$ and ${\bf z}$
239: are limitedly distant, too, and ${\bf x}$, ${\bf y}$, and ${\bf z}$ are
240: all in the same nodal galaxy. $\Box$
241:
242: We define a {\em galaxy} $\Gamma$ of $^{*}\!G$ as a maximal
243: nonstandard subgraph of
244: $^{*}\!G$ whose hypernodes are all in the same nodal galaxy
245: $\dot{\Gamma}$; that is, the hyperbranches of $\Gamma$
246: corresponding to $\dot{\Gamma}$ are all those pairs $\{{\bf x},{\bf y}\}$
247: such that ${\bf x},{\bf y} \in \dot{\Gamma}$. We will say that a
248: hypernode ${\bf x}$ is {\em in} $\Gamma$ when ${\bf x}\in\dot{\Gamma}$
249: and that a hyperbranch $\{{\bf x},{\bf y}\}$ is {\em in} $\Gamma$ when
250: ${\bf x},{\bf y}\in\dot{\Gamma}$.
251: It follows from Lemma 3.1 that the galaxies of
252: $^{*}\!G$ partition $^{*}\!G$ in the sense of graphical partitioning
253: (i.e., each hyperbranch is in one and only one galaxy).
254:
255: The {\em principal galaxy} $\Gamma_{0}$ of $^{*}\!G$ is that unique
256: galaxy, each of whose hypernodes is limitedly distant from some
257: standard hypernode
258: (and therefore from all standard hypernodes).
259: All the nodes in $G$ will be (i.e., can be identified with)
260: standard hypernodes in $\Gamma_{0}$, but there may be nonstandard
261: hypernodes in $\Gamma_{0}$ as well. The following examples
262: illustrate this point.
263:
264: {\bf Example 3.2.} Consider the endless (i.e., two-way infinite)
265: path:
266: \[ P\;=\;\langle\ldots,x_{-1},b_{-1},x_{0},b_{0},x_{1},b_{1}\ldots\rangle \]
267: with nodes $x_{k}$ and branches $b_{k}$, $k\in\Z$, $\Z$
268: being the set of integers. The enlargement $^{*}\!P$ of $P$ has
269: hypernodes, each being represented by $[x_{k_{n}}]$ where
270: $\langle k_{n}\rangle$ is some sequence of integers.
271: Each hyperbranch is represented by
272: $[\{x_{k_{n}},x_{k_{n}+1}]$. The nodal galaxies are infinitely many
273: because they correspond bijectively with the galaxies of the
274: enlargement $^{*}\!\Z$ of $\Z$. Moreover, the principal galaxy
275: $\Gamma_{0}$ of $^{*}\!P$ has only standard hypernodes and in fact is
276: (i.e., can be identified with) $P$ itself. Also, every galaxy
277: is graphically isomorphic to $\Gamma_{0}$ and therefore to
278: every other galaxy. $\Box$
279:
280: {\bf Example 3.3.} Now, consider a one-ended path:
281: \[ T\;=\;\langle x_{0},b_{0},x_{1},b_{1},x_{2},b_{2},\ldots\rangle \]
282: Each hypernode in the enlargement $^{*}\!T$ of $T$
283: is represented by $[x_{k_{n}}]$, where $\langle k_{n}\rangle$
284: is some sequence of natural numbers. Thus, $^{*}\!T$ has a hypernode set
285: $^{*}\!X$ that can be identified with the set $^{*}\!\N$ of
286: hypernaturals. Hence, $^{*}\!T$ has an infinitely of galaxies, too.
287: The principal galaxy $\Gamma_{0}$ of $^{*}\!T$ is the one-ended
288: path $T$. However, any hypernode ${\bf x}=[x_{k_{n}}]$
289: in a galaxy $\Gamma$ different from $\Gamma_{0}$ will be such that,
290: for every $m\in\N$,
291: $\{n\!: k_{n}>m\}\in{\cal F}$. Such a hypernode is adjacent both to
292: $[x_{k_{n}+1}]$ and to $[x_{k_{n}-1}]$, where we are free to replace
293: $x_{k_{n}-1}$ by, say, $x_{0}$ whenever $k_{n}=0$.
294: (The set $\{n\!: k_{n}=0\}$ will not be a member of ${\cal F}$ when
295: ${\bf x}=[x_{k_{n}}]$ is in $\Gamma$.) Thus, ${\bf x}=[x_{k_{n}}]\in\Gamma$
296: has both a predecessor and a successor, which implies
297: that $\Gamma$ is graphically isomorphic to an endless path. In fact,
298: all the galaxies other than $\Gamma_{0}$ are isomorphic to each
299: other, being identifiable with an endless path. $\Box$
300:
301: {\bf Example 3.4.} Consider next the grounded, one-way infinite
302: ladder $L$ of Figure 1. Now, for every $k\in\N$,
303: $d(x_{k},x_{g})=d(x_{k},x_{k+1})=1$, and, for every
304: $k,l\in\N$ with $|k-l|>1$, $d(x_{k},x_{l})=2$. In this case, for
305: every two hypernodes ${\bf x}$ and ${\bf y}$,
306: ${\bf d}({\bf x},{\bf y})\leq[2]=2$. Thus, every two hypernodes
307: are limitedly distant
308: from each other, which means that $^{*}\!L$ has only one galaxy,
309: its principal galaxy $\Gamma_{0}$. Now, $\Gamma_{0}$ has both
310: standard and nonstandard hypernodes. $\Box$
311:
312: {\bf Example 3.5.} Furthermore, consider the graph $G$ obtained
313: from $L$ by appending a one-ended path $P$ starting at $x_{g}$, but
314: otherwise isolated from $L$, as shown in Figure 2.
315: In this case, we again have an infinity of galaxies
316: by virtue of the isolation of $P$ from $L$.
317: The principal galaxy $\Gamma_{0}$ has both standard and nonstandard
318: hypernodes, its nonstandard hypernodes being due to $L$.
319: All the other galaxies are graphically isomorphic to
320: an endless path (as in Example 3.3) and thus to each other,
321: but not to $G$ and not to $\Gamma_{0}$. $\Box$
322:
323: A subgraph $G_{s}$ of $G$ with the property that there exists a
324: natural number $k$ such that $d(x,y)\leq k$ for all pairs of
325: nodes $x,y$ in $G_{s}$ will be called a {\em finitely dispersed}
326: subgraph of $G$. Example 3.5 suggests that the structures of the galaxies
327: other than $\Gamma_{0}$ do not depend upon any finitely dispersed
328: subgraph of $G$. This is true in general because the nodes
329: $x_{n}$ in any representative $\langle x_{n}\rangle$ of any hypernode
330: in a galaxy other than $\Gamma_{0}$ must lie outside any finitely
331: dispersed subgraph of $G$ for almost all $n$ whatever be the choice of that
332: finitely dispersed subgraph.
333:
334: For instance, consider
335:
336: {\bf Example 3.6.} Let $D_{2}$ be the 2-dimensional grid; that is,
337: we can represent $D_{2}$ by having its nodes at the lattice points
338: $(k,l)$ of the 2-dimensional plane, where $k,l\in \Z$ and with its
339: branches being $\{(k,l),(k+1,l)\}$ and $\{(k,l),(k,l+1)\}$.
340: So, the hypernodes of $^{*}\!D_{2}$ occur at $^{*}\!\Z\,\times\,^{*}\!\Z$.
341: Under this representation, the principal nodal galaxy of
342: $^{*}\!D_{2}$ will have its nodes at the lattice points of
343: $\Z\times\Z$.
344:
345: Next, let $G$ be a connected graph obtained from $D_{2}$ by deleting
346: or appending finitely many branches to $D_{2}$. So,
347: outside a finitely dispersed subgraph of $G$, $G$ is identical
348: to $D_{2}$. Then the principal galaxy $\Gamma_{0}$ of
349: $^{*}\!G$ is the same as (i.e., is graphically isomorphic to)
350: $G$, but every other galaxy is the same as $D_{2}$. $\Box$
351:
352: In view of Examples 3.3 and 3.4, the following theorem is pertinent.
353: As always, we assume that $G$ is connected and has an infinite node set $X$.
354:
355: {\bf Theorem 3.7.} {\em Let $G$ be locally finite. Then, $^{*}\!G$
356: has at least one hypernode not in its principal galaxy $\Gamma_{0}$ and thus
357: at least one galaxy $\Gamma_{1}$ different from $\Gamma_{0}$.}
358:
359: {\bf Proof.} Choose any $x_{0}\in X$. By
360: connectedness and local finiteness,
361: for each $n\in\N$, the set $X_{n}$ of nodes that are at a distance of
362: $n$ from $x_{0}$ is nonempty and finite. Also, $\cup X_{n}\,=\,X$
363: by the connectedness of $G$. By K\"{o}nig's Lemma \cite[page 40]{wi},
364: there is a one-ended path $P$ starting at $x_{0}$. $P$ must
365: pass through every $X_{n}$. Thus, there is a subsequence
366: $\langle x_{0},x_{1},x_{2},\ldots\rangle$ of the sequence of nodes of
367: $P$ such that $x_{n}\in X_{n}$; that is, $d(x_{n},x_{0})=n$ for every $n$.
368: Set ${\bf x}=[x_{n}]$. Then, ${\bf x}$ must be in a galaxy
369: $\Gamma_{1}$ that is different from the principal galaxy $\Gamma_{0}$.
370: $\Box$
371:
372: \section{When $^{*}\!G$ Has a Hypernode Not in Its Principal Galaxy}
373:
374: In this section, $G$ is connected and infinite but not necessarily
375: locally finite. Let $\Gamma_{a}$ and $\Gamma_{b}$ be two galaxies that are
376: different from the principal galaxy $\Gamma_{0}$ of $^{*}\!G$.
377: We shall say that $\Gamma_{a}$ {\em is closer to $\Gamma_{0}$
378: than is} $\Gamma_{b}$ and that $\Gamma_{b}$ {\em is further away from
379: $\Gamma_{0}$ than is} $\Gamma_{a}$
380: if there are a ${\bf y}=[y_{n}]$ in $\Gamma_{a}$ and a
381: ${\bf z}=[z_{n}]$ in $\Gamma_{b}$ such that,
382: for some ${\bf x}=[x_{n}]$ in $\Gamma_{0}$
383: and for every $m\in\N$, we have
384: \[ N_{0}(m)\;=\;\{n\!:d(z_{n},x_{n})-d(y_{n},x_{n})
385: \,\geq\, m\}\;\in\;{\cal F}. \]
386: Any set of galaxies for which every two of them, say,
387: $\Gamma_{a}$ and $\Gamma_{b}$ satisfy this condition will be said to be
388: {\em totally ordered according to their closeness to}
389: $\Gamma_{0}$. With Lemma 3.1 in hand, the conditions for a total ordering
390: (reflexivity, antisymmetry, transitivity, and
391: connectedness) are readily shown. For instance, the proof of Theorem 4.3
392: below establishes transitivity.
393:
394: {\bf Lemma 4.1.} {\em These definitions are independent of the
395: representative sequences $\langle x_{n}\rangle$, $\langle y_{n}\rangle$,
396: and $\langle z_{n}\rangle$ chosen for ${\bf x}$, ${\bf y}$,
397: and ${\bf z}$.}
398:
399: {\bf Proof.} Let $\langle x_{n}'\rangle$, $\langle y_{n}'\rangle$,
400: and $\langle z_{n}'\rangle$ be any other such representative sequences.
401: Then,
402: \[ d(z_{n},x_{n})\;\leq\;d(z_{n},z_{n}')+d(z_{n}',x_{n}')+d(x_{n}',x_{n}). \]
403: So,
404: \[ d(z_{n}',x_{n}')\;\geq\;d(z_{n},x_{n})-d(z_{n},z_{n}')-d(x_{n}',x_{n})
405: \;=\; d(z_{n},x_{n}) \]
406: for all $n$ in some $N_{1}\in{\cal F}$.
407: Also,
408: \[ d(y_{n}',x_{n}')\;\leq\;
409: d(y_{n}',y_{n})+d(y_{n},x_{n})+d(x_{n},x_{n}')\;=\; d(y_{n},x_{n}) \]
410: for all $n$ in some $N_{2}\in{\cal F}$.
411: Therefore,
412: \[ d(z_{n}',x_{n}')-d(y_{n}',x_{n}')\;\geq\; d(z_{n},x_{n})-d(y_{n},x_{n}) \]
413: for all $n$ in $N_{1}\cap N_{2}\,\in\, {\cal F}$.
414: So, for $N_{0}(m)$ as defined above and for each $m$
415: no matter how large,
416: \[ \{n\!: d(z_{n}',x_{n}')-d(y_{n}',x_{n}')\;\geq\; m\}
417: \;\supseteq\; N_{0}(m)\cap N_{1}\cap N_{2}\;\in\; {\cal F}, \]
418: which implies that the left-hand side is also a set in $\cal F$.
419: This proves Lemma 4.1. $\Box$
420:
421: We will say that a set $A$ is a {\em totally ordered, two-way infinite
422: sequence} if there is a bijection from the set $\Z$ of integers to the set
423: $A$ that preserves the total ordering of $\Z$.
424:
425: {\bf Theorem 4.2.} {\em If $^{*}\!G$ has a hypernode
426: ${\bf v}=[v_{n}]$ that is not in
427: its principal galaxy $\Gamma_{0}$, then there exists a two-way
428: infinite sequence of galaxies totally ordered according to their
429: closeness to $\Gamma_{0}$ and with $\bf v$ being in one of those galaxies.}
430:
431: {\bf Note.} There may be many such sequences, and a galaxy in one
432: sequence and a galaxy in another sequence
433: may not be comparable according to their
434: closeness to $\Gamma_{0}$. Also, a somewhat
435: different version of this theorem with a rather longer proof
436: can be found in the archival website, www.arxiv.org,
437: under Mathematics, Zemanian.
438:
439: {\bf Proof.} Let ${\bf x}=[\langle x,x,x,\ldots\rangle ]$ be a
440: standard hypernode in $\Gamma_{0}$. Also, let ${\bf v}=[v_{n}]$
441: be the asserted hypernode not in $\Gamma_{0}$. Thus, for each
442: $m\in\N$,
443: \begin{equation}
444: \{n\!: d(x,v_{n})>m\}\,\in\,{\cal F}. \label{4.1}
445: \end{equation}
446: Between every two nodes of a connected, conventionally infinite
447: graph there is a geodesic path whose length is equal to the
448: distance between those nodes. For each $n\in\N$, choose a
449: geodesic path $P$ in $G$ terminating at $x$ and $v_{n}$.
450: If the natural number $d(x,v_{n})$ is even (resp. odd), there is a
451: unique node $u_{n}$ in $P$ such that $d(x,u_{n})=d(u_{n},v_{n})
452: =d(x,v_{n})/2$ (resp. $d(x,u_{n})=d(u_{n},v_{n})-1=(d(x,v_{n})-1)/2)$.
453: It follows from this that, if there is a $k\in\N$ such that
454: $\{n\!: d(x,u_{n})\leq k\}\in{\cal F}$, then there is a
455: $k'\in\N$ with $\{n\!: d(x,v_{n})\leq k'\}\in{\cal F}$,
456: in violation of (\ref{4.1}). Consequently, for each
457: $m\in\N$, $\{n\!: d(x,u_{n})>m\}\in{\cal F}$. This implies that
458: ${\bf u}=[u_{n}]$ is in a galaxy $\Gamma_{a}$ different
459: from the principal galaxy $\Gamma_{0}$.
460:
461: Furthermore, with $d(x,v_{n})$ being even (resp. odd) again and
462: with $u_{n}$ being chosen as before,
463: \[ d(x,v_{n})\,-\, d(x,u_{n})\;=\; d(x,v_{n})/2 \]
464: (resp.
465: \[ d(x,v_{n})\,-\,d(x,u_{n})\;=\;(d(x,v_{n})+1)/2). \]
466: Now, if there is a $k\in\N$ such that
467: \[ \{n\!: d(x,v_{n})-d(x,u_{n})\leq k\}\;\in\;{\cal F}, \]
468: then there is a $k'\in\N$ such that
469: \[ \{n\!: d(x,v_{n})\leq k'\}\;\in\;{\cal F}, \]
470: again in violation of (\ref{4.1}). Thus, for each $m\in\N$,
471: \[ \{n\!: d(x,v_{n})\,-\,d(x,u_{n})\,>\,m\}\;\in\;{\cal F}.\]
472: This implies that ${\bf u}=[u_{n}]$ and ${\bf v}=[v_{n}]$
473: are in different galaxies, $\Gamma_{a}$ and $\Gamma_{b}$
474: respectively, with $\Gamma_{a}$ being closer to $\Gamma_{0}$ than
475: is $\Gamma_{b}$.
476:
477: We can now repeat this argument with $\Gamma_{b}$ replaced
478: by $\Gamma_{a}$ and with ${\bf u}=[u_{n}]$ playing the role that
479: ${\bf v}=[v_{n}]$ played
480: to find still another galaxy $\Gamma_{a}'$ different from
481: $\Gamma_{0}$ and closer to $\Gamma_{0}$ than is $\Gamma_{a}$.
482: Continual repetitions yield an infinite sequence of galaxies
483: indexed by, say, the negative integers and totally ordered
484: by their closeness to $\Gamma_{0}$.
485:
486: The conclusion that there is an infinite sequence of galaxies
487: progressively further away from $\Gamma_{0}$ than is
488: $\Gamma_{b}$ is easier to prove. With ${\bf v}\in \Gamma_{b}$ as before,
489: we have that, for every $m\in\N$, $\{n\!: d(x,v_{n})>m\}\in {\cal F}$.
490: Therefore, for each $n\in\N$, we can choose $w_{n}$ as an element of
491: $\langle v_{n}\rangle$ such that
492: \begin{equation}
493: d(x,w_{n})\geq d(x,v_{n})+n. \label{4.2}
494: \end{equation}
495: Hence, for each $m\in\N$,
496: \[ \{n\!: d(x,w_{n})>m\}\;\supseteq\; \{n\!: d(x, v_{n}) > m\}. \]
497: Since the right-hand side is a member of $\cal F$,
498: so too is the left-hand side. Thus, ${\bf w}=[w_{n}]$ is in a
499: galaxy different from $\Gamma_{0}$. Moreover,
500: from (\ref{4.2}) we have that, for each $m\in\N$,
501: $\{n\!: d(x,w_{n})-d(x,v_{n})>m\}$ is a cofinite set and therefore
502: is a member of $\cal F$.
503: Consequently, $\bf w$ is in a galaxy $\Gamma_{c}$ that is further away from
504: $\Gamma_{0}$ than is $\Gamma_{b}$.
505:
506: We can repeat the argument of the last paragraph with
507: $\Gamma_{c}$ in place of $\Gamma_{b}$ to find still another galaxy
508: $\Gamma_{c}'$ further away from $\Gamma_{0}$ than is $\Gamma_{c}$.
509: Repetitions of this argument show that there is an
510: infinite sequence of galaxies indexed by, say, the
511: positive integers and totally
512: ordered by their closeness to $\Gamma_{0}$.
513:
514: The conjunction of the two infinite sequences along with $\Gamma_{b}$
515: yields the conclusion of
516: the theorem. $\Box$
517:
518: By virtue of Theorem 3.7, the
519: conclusion of Theorem 4.2 holds whenever $G$ is locally finite.
520:
521: In general, the hypothesis of Theorem 4.2 may or may not hold. Thus,
522: $^{*}\!G$ either has exactly one galaxy, its principal one $\Gamma_{0}$,
523: or has infinitely many galaxies.
524:
525: Instead of the idea of ``totally ordered according to closeness
526: to $\Gamma_{0}$,'' we can define the idea of ``partially ordered
527: according to closeness to $\Gamma_{0}$'' in much the same way. Just drop the
528: connectedness axiom for a total ordering.
529:
530: {\bf Theorem 4.3.} {\em Under the hypothesis of Theorem 4.2,
531: the set of galaxies of $^{*}\!G$ is
532: partially ordered according to the closeness of the galaxies
533: to the principal galaxy $\Gamma_{0}$.}
534:
535: {\bf Proof.} Reflexivity and antisymmetry are obvious.
536: Consider transitivity: Let $\Gamma_{a}$, $\Gamma_{b}$, and $\Gamma_{c}$
537: be galaxies different from $\Gamma_{0}$. (The case where
538: $\Gamma_{a}=\Gamma_{0}$ can be argued similarly.)
539: Assume that $\Gamma_{a}$ is closer to $\Gamma_{0}$ than is
540: $\Gamma_{b}$ and that $\Gamma_{b}$ is closer to $\Gamma_{0}$
541: than is $\Gamma_{c}$. Thus, for any ${\bf x}$ in $\Gamma_{0}$,
542: ${\bf u}$ in $\Gamma_{a}$, ${\bf v}$ in $\Gamma_{b}$, and
543: ${\bf w}$ in $\Gamma_{c}$ and for every $m\in\N$, we have
544: \[ N_{uv}\;=\;\{n\!: d(v_{n},x_{n})-d(u_{n},x_{n})\geq m\}\,\in\,{\cal F} \]
545: and
546: \[ N_{vw}\;=\;\{n\!: d(w_{n},x_{n})-d(v_{n},x_{n})\geq m\}\,\in\,{\cal F}. \]
547: We also have
548: \[ d(w_{n},x_{n})-d(u_{n},x_{n})\;=\;d(w_{n},x_{n})-d(v_{n},x_{n})
549: +d(v_{n},x_{n})-d(u_{n},x_{n}). \]
550: So,
551: \[ N_{uw}\;=\;\{n\!: d(w_{n},x_{n})-d(u_{n},x_{n})\geq 2m\}\,
552: \supseteq\,N_{uv}\cap N_{vw}\;\in\;{\cal F}. \]
553: Thus, $N_{uw}\in {\cal F}$. Since $m$ can be chosen arbitrarily,
554: we can conclude that $\Gamma_{a}$ is closer to $\Gamma_{0}$
555: than is $\Gamma_{c}$. $\Box$
556:
557: \section{The Hyperordinals}
558:
559: In the following sections, we shall extend the results obtained so far
560: to enlargements of transfinite graphs of rank 1,
561: that is, to enlargements of 1-graphs. For this purpose,
562: we need to replace the set $^{*}\!\N$ of hypernaturals by
563: a set of ``hyperordinals''; these are defined as follows.
564: A hyperordinal $\underline{\alpha}$ is an equivalence class
565: of sequences of ordinals where two such sequences
566: $\langle \alpha_{n}\rangle$ and $\langle \beta_{n}\rangle$ are taken
567: to be equivalent if $\{n\!: \alpha_{n}=\beta_{n}\}\in {\cal F}$.
568: We denote $\underline{\alpha}$ also by $[\alpha_{n}]$
569: where again the $\alpha_{n}$
570: are the elements of one (any one) of the sequences in the equivalent class.
571: Any set of hyperordinals is totally ordered by the inequality relation.
572: That is, given any hyperordinals $\underline{\alpha}=[\alpha_{n}]$
573: and $\underline{\beta}=[\beta_{n}]$,
574: exactly one of the sets:
575: \[ \{n\!: \alpha_{n}<\beta_{n}\},\;\; \{n\!:\alpha_{n}=\beta_{n}\},\;\; \{n\!:\alpha_{n}>\beta_{n}\} \]
576: will be in ${\cal F}$. So, exactly one of the expressions:
577: \[ \underline{\alpha}<\underline{\beta},\;\;\underline{\alpha}=\underline{\beta},\;\;\underline{\alpha}>\underline{\beta} \]
578: holds.
579:
580: \section{Walks in 1-Graphs}
581:
582: 1-graphs arise when conventionally infinite graphs
583: are connected at their infinite extremities
584: through 1-nodes, the latter being a generalization of the
585: idea of a node. Such 1-nodes and the resulting 1-graphs are defined in
586: \cite[Section 2.1]{tgen} and also in \cite[Section 2.3]{gn}.
587: Let us restate the needed definitions
588: concisely.
589:
590: We will be dealing with two kinds of nodes and two kinds of graphs.
591: A conventionally infinite graph $G^{0}$
592: will now be called a 0-{\em graph} and the nodes in
593: $G^{0}$ will be called 0-{\em nodes} in order to distinguish these ideas
594: from those pertaining to transfinite graphs of rank 1.
595: Similarly, what we called a ``hypernode'' previously will henceforth be called
596: a 0-{\em hypernode}, and what we called a ``galaxy'' in the enlargement of
597: a 0-graph will now be called a 0-{\em galaxy}.
598:
599: An {\em infinite extremity} of a 0-graph $G^{0}$ is defined as an
600: equivalence class of one-ended paths in $G^{0}$, where two such paths
601: are considered to be {\em equivalent} if they are eventually identical.
602: Such an equivalence class is called a 0-{\em tip} of $G^{0}$.
603: $G^{0}$ may have one or more 0-tips (or possibly none at all).
604: To obtain the ``1-nodes,'' the set of 0-tips is partitioned
605: in some fashion into subsets, and to each subset a single 0-node
606: may (or may not) be added under the proviso that, if a 0-node
607: is added to one subset, it is not added to any other subset.
608: Then, each subset (possibly augmented with a 0-node) is called
609: a 1-{\em node}. With $X^{1}$ denoting the set of 1-nodes and $X^{0}$
610: the set of 0-nodes of $G^{0}$, the 1-{\em graph} $G^{1}$
611: is defined as the triplet:
612: \[G^{1}\;=\;\{X^{0},B,X^{1}\}, \]
613: and $G^{0}=\{X^{0},B\}$ is now called the 0-{\em graph of}
614: $G^{1}$. Furthermore, a path in $G^{0}$ is now called a 0-{\em path},
615: and connectedness in $G^{0}$ is now called 0-{\em connectedness}.
616: We will consistently append the superscript 0 to the symbols and
617: the prefix 0- to the
618: terminology for concepts from Sections 2 through 4 regarding 0-graphs.
619:
620: In order to define the ``1-galaxies,'' we need the idea
621: of distances in a 1-graph $G^{1}$. But now, we must make
622: a significant choice. The distances between two nodes
623: (0-nodes or 1-nodes) can be defined as the minimum length of
624: all paths---or, alternatively, of all walks---connecting the two nodes.
625: It turns out that a path need not exist between two nodes
626: in a 1-graph $G^{1}$, but a walk always will exist between them.
627: To ensure the existence of at least one path between every two nodes,
628: additional conditions must be imposed on $G^{1}$ (see
629: \cite[Conditions 3.2-1 and 3.5-1]{tgen} or \cite[Condition 3.1-2]{gn}),
630: and this leads to a more restrictive and yet more complicated
631: theory involving distances. Such can be done,
632: but it is more general and simpler to use walk-based distance ideas.
633: This we now do.
634:
635: A {\em nontrivial 0-walk} $W^{0}$ in a 0-graph is the conventional concept.
636: It is a (finite or one-way infinite or two-way infinite)
637: alternating sequence:
638: \begin{equation}
639: W^{0}\;=\;\langle \ldots,x_{-1}^{0},b_{-1},x_{0}^{0},b_{0},x_{1}^{0},b_{1},\ldots\rangle \label{6.1}
640: \end{equation}
641: of 0-nodes $x_{m}^{0}$ and branches $b_{m}$,
642: where each branch $b_{m}$ is incident to the two 0-nodes $x_{m}^{0}$
643: and $x_{m+1}^{0}$ adjacent to it in the sequence.
644: If the sequence terminates at either side, it is required to terminate
645: at a 0-node. The 0-walk is called {\em two-ended} or
646: {\em finite} if it terminates on both sides, {\em one-ended} if it
647: terminates on just one side, and {\em endless}
648: if it terminates on neither side.
649:
650: A {\em trivial 0-walk} is a singleton set whose sole element is a 0-node.
651:
652: A one-ended 0-walk $W^{0}$ will be called {\em extended} if its 0-nodes
653: are eventually distinct, that is, if it is eventually identical
654: to a one-ended path. We say that $W^{0}$ {\em traverses} a 0-tip
655: if it is extended and eventually identical to
656: a representative of that 0-tip. Finally, $W^{0}$ is said
657: to {\em reach} a 1-node $x^{1}$ if $W^{0}$ traverses
658: a 0-tip contained in $x^{1}$. In the same way,
659: an endless 0-walk can {\em reach} two 1-nodes
660: (or possibly reach the same 1-node) by traversing two
661: 0-tips, one toward the left and the other toward the right.
662: When this is so, we say that the endless 0-walk is {\em extended}.
663: On the other hand, if a 0-walk terminates at a 0-node contained in
664: a 1-node, we again say that the 0-walk {\em reaches}
665: both of those nodes and does so {\em through} a branch incident to that
666: 0-node.
667:
668: Every two-ended 0-walk contains a 0-path that terminates at the two
669: 0-nodes at which the 0-walk terminates, so there is no need to employ
670: 0-walks when defining distances in a 0-graph. On the other hand,
671: such a need arises for 1-graphs. To meet this need, we
672: first define a 0-{\em section} $S^{0}$ in a 1-graph $G^{1}$
673: as a subgraph $S^{0}$ of the 0-graph $G^{0}$ of $G^{1}$
674: induced by a maximal set of branches that are pairwise 0-connected in
675: $G^{0}$. A 1-node $x^{1}$ is said to be {\em incident to}
676: $S^{0}$ if either it contains a 0-node incident to a branch
677: of $S^{0}$ or it contains a 0-tip having a representative one-ended path
678: lying entirely within $S^{0}$. In this case, we also say
679: that that 0-tip {\em belongs to} $S^{0}$.
680: Given two 1-nodes $x^{1}$ and $y^{1}$ incident to $S^{0}$, there
681: will be a 0-walk $W^{0}$ in $S^{0}$ that reaches each of
682: $x^{1}$ and $y^{1}$ through a 0-tip belonging to $S^{0}$ or
683: through a branch in $S^{0}$.\footnote{For examples of when a 0-walk is needed
684: because a 0-path won't do, see Figures 3.1 and 3.2 of \cite{tgen}
685: and Figures 4.1, 5.1, 5.2, and 5.3 of \cite{gn}.}
686: Moreover, there may also be a 0-walk $W^{0}$ in $S^{0}$ that reaches the
687: same 1-node at both extremities of $W^{0}$. To be more specific, let
688: us state
689:
690: {\bf Lemma 6.1.} {\em Let $S^{0}$ be a 0-section in $G^{1}$, and let
691: $x^{1}$ and $y^{1}$ be two 1-nodes incident to $S^{0}$.
692: Then, there exists a 0-walk in $S^{0}$ that reaches
693: $x^{1}$ and $y^{1}$.}
694:
695: {\bf Proof.} That $x^{1}$ is incident to $S^{0}$ means that
696: there is a 0-path $P_{x}^{0}$ in $S^{0}$ that either reaches $x^{1}$
697: through a 0-tip of $x^{1}$ or reaches $x^{1}$ through a branch.
698: Similarly, there is such a 0-path $P_{y}^{0}$ reaching $y^{1}$. Let
699: $u^{0}$ be a node of $P_{x}^{0}$, and let $v^{0}$ be a node
700: of $P_{y}^{0}$. Since $S^{0}$ is 0-connected,
701: there is a 0-path $P_{uv}^{0}$ in $S^{0}$ terminating at $u^{0}$ and $v^{0}$
702: (possibly a trivial 0-path if $u^{0}=v^{0}$).
703: Then, $P_{x}^{0}\cup P_{uv}^{0}\cup P_{y}^{0}$ as a 0-walk
704: in $S^{0}$ as asserted. $\Box$
705:
706: A {\em nontrivial, two-ended 1-walk} $W^{1}$ is a finite sequence:
707: \begin{equation}
708: W^{1}\;=\;\langle x_{0},W^{0}_{0},x_{1}^{1},W_{1}^{0},\ldots,x_{m-1}^{1},W_{m-1}^{0},x_{m}\rangle \label{6.2}
709: \end{equation}
710: with $m\geq 1$ that satisfies the following conditions.
711: \begin{description}
712: \item{1.} $x_{1}^{1},\ldots,x_{m-1}^{1}$ are 1-nodes, while $x_{0}$
713: and $x_{m}$ may be either 0-nodes or 1-nodes.
714: \item{2.} For each $k=0,\ldots,m-1$, $W_{k}^{0}$ is a nontrivial 0-walk
715: that reaches the two nodes adjacent to it in the sequence.
716: \item{3.} For each $k=1,\ldots,m-1$, at least one of
717: $W_{k-1}^{0}$ and $W_{k}^{0}$ reaches $x_{k}^{1}$
718: through a 0-tip, not through a branch. Also, if $m=1$, $W_{0}^{0}$
719: reaches at least one of $x_{0}$ and $x_{1}$ through a 0-tip.
720: \end{description}
721: A {\em one-ended} 1-walk is a sequence like (\ref{6.2})
722: except that it extends infinitely to the right. An {\em endless} 1-walk
723: extends infinitely on both sides.
724: A {\em trivial 1-walk} is a singleton set whose sole element
725: is either a 0-node or a 1-node.
726:
727: We now define a more general kind of connectedness (called ``1-wconnectedness''
728: to distinguish it from path-based 1-connectedness).
729: Two branches (resp. two nodes---either 0-nodes or 1-nodes)
730: will be said to be 1-{\em wconnected} if there exists a 0-walk
731: or 1-walk that terminates at a 0-node of each branch (resp. that
732: terminates at those two nodes). If a terminal node of a walk is
733: the same as, or contains, or is contained in the terminal node of
734: another walk, the two walks taken together
735: form another walk. We call this the {\em conjunction}
736: of the two walks.
737: It follows that 1-wconnectedness
738: is a transitive binary relation for the branch set $B$
739: of the 1-graph $G^{1}$ and is in fact an equivalence relation.
740: If every two branches of $G^{1}$ are 1-wconnected,
741: we will say that $G^{1}$ is 1-{\em wconnected}.
742:
743: \section{Walk-Based Distances in a 1-Graph}
744:
745: The length $|W^{0}|$ of a 0-walk $W^{0}$ is defined as follows:
746: If $W^{0}$ is two-ended, $|W^{0}|$ is the number $\tau_{0}$
747: of branch traversals in it; that is, each branch is counted as
748: many times as it appears in $W^{0}$.
749: If $W^{0}$ is one-ended and extended, we set $|W^{0}|=\omega$, the
750: first transfinite ordinal. If $W^{0}$ is endless and extended in
751: both directions, we set $|W^{0}|=\omega\cdot 2$.
752:
753: As for a nontrivial
754: two-ended 1-walk $W^{1}$, its length $|W^{1}|$ is taken to be
755: $|W^{1}|=\sum_{k=0}^{m}|W_{k}^{0}|$, where the sum is
756: over the finitely many 0-walks $W_{k}^{0}$ in (\ref{6.2}). Thus,
757: \begin{equation}
758: |W^{1}|\;=\;\omega\cdot \tau_{1}+\tau_{0} \label{7.1}
759: \end{equation}
760: where $\tau_{1}$ is the number of traversals of 0-tips performed by $W^{1}$
761: and $\tau_{0}$ is the number of traversals of branches in all the
762: two-ended (i.e., finite) 0-walks appearing as terms in (\ref{6.2}).
763: We take $\sum_{k=0}^{m} |W^{0}_{k}|$ to be the natural sum of ordinals;
764: this yields a normal expansion of an ordinal \cite[pages 354-355]{ab}.
765: $\tau_{1}$ is not 0 because $W^{1}$ is a nontrivial, two-sided 1-walk.
766: However, $\tau_{0}$ may be 0, this occurring when every $W_{k}^{0}$
767: in (\ref{6.2}) is one-ended or endless.
768:
769: A 0-node is called {\em maximal} if it is not contained in a 1-node,
770: and {\em nonmaximal} otherwise. A distance measured from a
771: nonmaximal 0-node is the same as that measured from the 1-node containing it.
772: Given two nodes $x$ and $y$ (of ranks 0 or 1), we define
773: the {\em wdistance}\footnote{We write ``wdistance'' to distinguish this
774: walk-based idea from a distance based on paths.} $d(x,y)$ between
775: them as
776: \begin{equation}
777: d(x,y)\;=\;\min|W_{x,y}| \label{7.2}
778: \end{equation}
779: where the minimum is taken over all two-ended walks (0-walks or 1-walks)
780: terminating at $x$ and $y$. That minimum exists because any set of
781: ordinals is a well-ordered set. In view of (\ref{7.1}), $d(x,y)<\omega^{2}$.
782: If $x=y$, we set $d(x,x)=0$.
783:
784: Clearly, if $x\neq y$, $d(x,y)>0$ and $d(x,y)=d(y,x)$.
785: Furthermore, the conjunction of two two-ended
786: walks is again a two-ended walk, whose length is the natural sum
787: of the ordinal lengths
788: of the two walks. So, by taking minimums appropriately, we obtain the
789: triangle inequality:
790: \begin{equation}
791: d(x,z)\;\leq\;d(x,y)\,+\,d(y,z) \label{7.3}
792: \end{equation}
793: where again the natural sum of ordinals is understood.
794: Altogether then, we have
795:
796: {\bf Lemma 7.1.} {\em The ordinal-valued wdistances between the maximal
797: nodes of a 1-graph satisfy the metric axioms.}
798:
799: \section{Enlargements of 1-Graphs and Hyperdistances in Them}
800:
801: In \cite[pages 163-164]{gn}, a nonstandard 1-node was defined
802: as an equivalence class of sequences of sets of tips shorted together,
803: with the tips taken from sequences of possibly differing 1-graphs.
804: But, since each set of tips shorted together is a 1-node,
805: that definition of a nonstandard 1-node can also be stated as
806: an equivalence class of sequences of 1-nodes.
807: Specializing to the case where all
808: the 1-graphs are the same, we have the following definition of a
809: nonstandard 1-node, which we now call a ``1-hypernode.''
810:
811: Consider a given 1-graph
812: along with a chosen free ultrafilter ${\cal F}$.
813: Two sequences $\langle x_{n}^{1}\rangle$ and
814: $\langle y_{n}^{1}\rangle$ of 1-nodes in $G^{1}$ are taken to be
815: {\em equivalent} if $\{n\!: x_{n}^{1}=y_{n}^{1}\}\in{\cal F}$.
816: It is easy to show that this is truly an equivalence relation.
817: Then, ${\bf x}^{1}=[x_{n}^{1}]$ denotes one such equivalence class,
818: where the $x_{n}^{1}$ are the elements of any one of the
819: sequences in that class. ${\bf x}^{1}$ will be called
820: a 1-{\em hypernode}.
821:
822: The {\em enlargement} of the 1-graph $G^{1}=\{X^{0},B,X^{1}\}$
823: is the nonstandard 1-graph
824: \[ ^{*}\!G^{1}\;=\;\{\,^{*}\!X^{0},\,^{*}\!B,\,^{*}\!X^{1}\,\} \]
825: where $^{*}\!X^{0}$ and $^{*}\!B$ are respectively the set of
826: 0-hypernodes and branches in the enlargement of the 0-graph
827: $G^{0}=\{X^{0},B\}$ of $G^{1}$ and $^{*}\!X^{1}$
828: is the set of 1-hypernodes defined above, that is, the set of all
829: equivalence classes of sequences of 1-nodes taken from $X^{1}$.
830:
831: We define the {\em hyperdistance} ${\bf d}$ between any two hypernodes
832: ${\bf x}$ and ${\bf y}$ of $^{*}\!G^{1}$ (of ranks 0 and/or 1)
833: to be the internal function
834: \begin{equation}
835: {\bf d}({\bf x},{\bf y})\;=\;[d(x_{n},y_{n})]. \label{8.1}
836: \end{equation}
837: Since distances in $G^{1}$ are less than $\omega^{2}$,
838: ${\bf d}({\bf x},{\bf y})$ is a hyperordinal
839: less than $\underline{\omega}^{2}$.
840: We say that a 0-hypernode ${\bf x}^{0}=[x_{n}^{0}]$ is {\em maximal}
841: if the set of $n$ for which $x_{n}^{0}$ is not contained
842: in a 1-node is a member of $\cal F$.
843: All the 1-nodes in this work are perforce
844: maximal because there are no nodes of higher rank.
845: ${\bf d}$, when restricted to the maximal hypernodes,
846: also satisfies the metric axioms, in particular, the
847: triangle inequality:
848: \begin{equation}
849: {\bf d}({\bf x},{\bf z})\;\leq\;{\bf d}({\bf x},{\bf y})\,+\,{\bf d}({\bf y},{\bf z}) \label{8.2}
850: \end{equation}
851: But, now $\bf d$ is hyperordinal-valued.
852:
853: \section{The Galaxies of $^{*}\!G^{1}$}
854:
855: The 0-{\em galaxies} of $^{*}\!G^{1}$ are defined just as they are for
856: the enlargement $^{*}\!G^{0}$ of a 0-graph; see Section 3.
857: However, we henceforth write ``0-galaxy'' in place of ``galaxy''
858: and ``0-limitedly distant'' in place of ``limitedly distant.''
859:
860: As was mentioned above, each 0-section of $G^{1}$ is the subgraph
861: of the 0-graph $G^{0}\{X^{0},B\}$ of $G^{1}$ induced by a
862: maximal set of branches that are 0-connected.
863: A 0-section is a 0-graph by itself.
864: So, within the
865: enlargement $^{*}\!G^{1}$, each 0-section $S^{0}$ enlarges into
866: $^{*}\!S^{0}$ as defined in Section 2. Within each enlarged
867: 0-section there may be one or more 0-galaxies. As a special case,
868: a particular 0-section may have only finitely many
869: 0-nodes, and so its enlargement is itself---all its 0-hypernodes
870: are standard. On the other hand, there may be infinitely
871: many 0-galaxies in some enlarged 0-section. Moreover,
872: the enlarged 0-sections do not, in general, comprise all of the
873: enlarged 0-graph $^{*}\!G^{0}=\{\,^{*}\!X^{0},\,^{*}\!B\,\}$
874: of $^{*}\!G^{1}$. Indeed, there can be a 0-hypernode
875: ${\bf x}^{0}=[x_{n}^{0}]$ where each $x_{n}^{0}$ resides in a different
876: 0-section; in this case ${\bf x}^{0}$ will reside in a 0-galaxy
877: that is not in an enlargement of a 0-section.
878:
879: Something more can happen with regard to the 0-galaxies in
880: $^{*}\!G^{1}$. 0-galaxies can now contain 1-hypernodes.
881: For example, this occurs when a 1-node $x^{1}$ is incident
882: to a 0-section $S^{0}$ through a branch. Then, the standard 1-hypernode
883: ${\bf x}^{1}$ corresponding to $x^{1}$ is 0-limitedly distant from the
884: standard 0-hypernodes in $^{*}\! S^{0}$. So,
885: there is a 0-galaxy containing
886: not only $^{*}\!S^{0}$ but ${\bf x}^{1}$ as well. See Example 9.3 below in
887: this regard. In general, the nodal 0-galaxies partition the set
888: $^{*}\!X^{0}\,\cup\,^{*}\!X^{1}$ of all the hypernodes in
889: $^{*}\!G^{1}$. As we shall see in Examples 9.1 and 9.2 below, there
890: may be a singleton 0-galaxy containing a 1-hypernode only.
891:
892: Let us now turn to the ``1-galaxies'' of $^{*}\!G^{1}$
893: Two hypernodes ${\bf x}=[x_{n}]$ and
894: ${\bf y}=[y_{n}]$ (of ranks 0 and/or 1) in
895: $^{*}\!G^{1}$ will be said to be in the same {\em nodal 1-galaxy}
896: $\dot{\Gamma}^{1}$ if there exists a natural number $k\in\N$
897: depending on the choices of $\bf x$ and $\bf y$
898: such that $\{n\!:d(x_{n},y_{n})\leq \omega\cdot k\}\in{\cal F}$.
899: In this case, we say that ${\bf x}$ and ${\bf y}$ are
900: {\em 1-limitedly distant}, and we write ${\bf d}({\bf x},{\bf y})
901: \leq [\omega\cdot k]$ where $[\omega\cdot k]$ denotes the standard
902: hyperordinal corresponding to $\omega\cdot k$.
903: This defines an equivalence relation on the set
904: $^{*}\!X^{0}\,\cup\, ^{*}\!X^{1}$ of all the hypernodes in $^{*}\!G^{1}$.
905: Indeed, reflexivity and symmetry are obvious. For transitivity,
906: assume that ${\bf x}$ and ${\bf y}$ are 1-limitedly distant and that
907: ${\bf y}$ and ${\bf z}$ are 1-limitedly distant, too. Then,
908: there are natural numbers $k_{1}$ and $k_{2}$ such that
909: \[ N_{xy}\;=\;\{n\!: d(x_{n},y_{n})\leq \omega\cdot k_{1}\}\,\in\,{\cal F} \]
910: and
911: \[ N_{yz}\;=\;\{n\!: d(y_{n},z_{n})\leq \omega\cdot k_{2}\}\,\in\,{\cal F}. \]
912: By the triangle inequality (\ref{7.3}),
913: \[ N_{xz}\;=\;\{n\!:d(x_{n},z_{n})\leq\omega\cdot (k_{1}+k_{2})\}\;\supseteq
914: \;N_{xy}\cap N_{yz}\;\in\;{\cal F}. \]
915: So, $N_{xz}\in {\cal F}$ and therefore ${\bf x}$ and ${\bf z}$ are
916: 1-limitedly distant. We can conclude that the set
917: $^{*}\!X^{0}\,\cup\,^{*}\!X^{1}$
918: of all hypernodes in $^{*}\! G^{1}$
919: is partitioned into nodal 1-galaxies by this equivalence relation.
920:
921: Corresponding to each nodal 1-galaxy $\dot{\Gamma}^{1}$, we
922: define a 1-{\em galaxy} $\Gamma^{1}$ as a nonstandard subgraph of
923: $^{*}\!G^{1}$ consisting of all the hypernodes in
924: $\dot{\Gamma}^{1}$ along with all the hyperbranches both of whose
925: 0-hypernodes are in $\dot{\Gamma}^{1}$.
926:
927: No hyperbranch can have its two incident 0-hypernodes in two different
928: 0-galaxies or two different 1-galaxies because the distance between their
929: 0-hypernodes is 1. Thus, the hyperbranch set $^{*}\!B$ is also partitioned
930: by the 0-galaxies and more coarsely
931: by the 1-galaxies.
932:
933: The {\em principal 1-galaxy} $\Gamma_{0}^{1}$ of $^{*}\!G^{1}$
934: is the 1-galaxy whose hypernodes are 1-limitedly distant from a standard
935: hypernode in $^{*}\!G^{1}$ (i.e., from a node of $G^{1}$).
936:
937: Note that the enlargement $^{*}\!S^{0}$ of each 0-section
938: $S^{0}$ of $G^{1}$ has its own principal 0-galaxy
939: $\Gamma_{0}^{0}(S^{0})$. Moreover, every $^{*}\!S^{0}$ lies
940: within the principal 1-galaxy $\Gamma_{0}^{1}$.
941: Indeed, any standard hypernode ${\bf x}$ by which
942: $\Gamma_{0}^{1}$ may be defined and any standard 0-hypernode ${\bf y}^{0}$
943: by which $\Gamma_{0}^{0}(S^{0})$ may be defined are 1-limitedly
944: distant. Also, the hyperdistance ${\bf d}({\bf y}^{0},{\bf z}^{0})$
945: between any two 0-hypernodes ${\bf y}^{0}$ and ${\bf z}^{0}$ of
946: $^{*}\!S^{0}$ is no larger than a hypernatural ${\bf k}$.
947: So, by the triangle inequality (\ref{8.2}), every 0-hypernode of
948: $^{*}\!S^{0}$ is 1-limitedly distant from ${\bf x}$. Whence our assertion.
949:
950: {\bf Example 9.1.} Consider an endless 1-path $P^{1}$ having an
951: endless 0-path between every consecutive pair of 1-nodes in $P^{1}$.
952: The 0-sections of $P^{1}$ are those endless 0-paths, and each of their
953: enlargements have an infinity of
954: 0-galaxies in $^{*}\!P^{1}$. However, there are other
955: 0-galaxies in $^{*}\!P^{1}$, infinitely many of them. Indeed,
956: consider a 0-hypernode ${\bf x}^{0}=[x_{n}^{0}]$,
957: where each 0-node $x_{n}^{0}$ lies in a different
958: 0-section of $P^{1}$; ${\bf x}^{0}$ will lie in a 0-galaxy $\Gamma_{1}^{0}$
959: different from all the 0-galaxies in any
960: enlargement of a 0-section of $P^{1}$. The 0-hypernodes of
961: $\Gamma_{1}^{0}$ will be the 0-hypernodes that are 0-limitedly
962: distant from ${\bf x}^{0}$. Furthermore, there are still other
963: 0-galaxies now. Each 1-hypernode ${\bf x}^{1}=[x_{n}^{1}]$ is the
964: sole member of a 0-galaxy. In fact, the nodal 0-galaxies partition
965: the set of all the 0-hypernodes and 1-hypernodes.
966:
967: On the other hand, the principal 1-galaxy of $^{*}\!P^{1}$ consists of all
968: the standard 1-hypernodes corresponding to the
969: 1-nodes of $P^{1}$ along with the enlargements of the
970: 0-sections of $P^{1}$.
971: Also, there will be infinitely many 1-galaxies, each of which
972: contains infinitely many 0-galaxies along with 1-hypernodes. In
973: this particular case, each of the 1-galaxies is graphically
974: isomorphic to the principal 1-galaxy, but this is not true in general. $\Box$
975:
976: {\bf Example 9.2.} An example of a nonstandard 1-graph
977: $^{*}\!G^{1}$ having exactly one 1-galaxy
978: (its principal one) and infinitely many 0-galaxies is
979: provided by the enlargement of the 1-graph $G^{1}$
980: obtained from the 0-graph of
981: Figure 1 by replacing each branch by an endless 0-path,
982: thereby converting each 0-node into a 1-node. Again each endless path
983: of that 1-graph $G^{1}$ is a 0-section, and its enlargement is like that
984: of Example 3.2. There are
985: infinitely many such 0-galaxies in each enlargement of a 0-section.
986: Also, there are infinitely many 0-galaxies,
987: each consisting of a single 1-hypernode.
988: With regard to the 1-galaxies, the enlargement $^{*}\!G^{1}$
989: of $G^{1}$
990: mimics that of Example 3.4, except that now
991: the rank 0 is replaced by the rank 1. The hyperdistance between
992: every two 1-hypernodes (resp. 0-hypernodes) is no larger than
993: $\omega\cdot 4$ (resp. $\omega\cdot 6$). Hence, $^{*}\!G^{1}$ has only
994: one 1-galaxy, its principal one. $\Box$
995:
996: {\bf Example 9.3.} Here is an example where each of the 1-hypernodes
997: is not isolated as the sole member
998: within a 0-galaxy. Replace each of the
999: horizontal branches in Figure 1 by an endless 0-path, but do not alter the
1000: branches incident to $x_{g}$. Now, the nodes $x_{k}$ $(k=0,1,2,\ldots)$
1001: become 1-nodes $x_{k}^{1}$, each containing a 0-node of the branch incident to
1002: $x_{k}^{1}$ and $x_{g}$. Each 1-hypernode is 0-limitedly distant from the
1003: standard 0-hypernode ${\bf x}_{g}$ and thus is not so isolated.
1004: The 1-hypernodes
1005: (whether standard or nonstandard) along
1006: with the standard 0-hypernode for $x_{g}$ and the
1007: (standard or nonstandard) hyperbranches
1008: incident to ${\bf x}_{g}$
1009: all comprise a single 0-galaxy. Moreover, there will
1010: be other 0-galaxies obtained through equivalence classes of sequences of
1011: these nodes and branches. In fact, each of the
1012: endless paths that replace the horizontal
1013: branches yield infinitely many 0-galaxies. Again, the nodal
1014: 0-galaxies partition the set of all the hypernodes in $^{*}\!G^{1}$.
1015:
1016: On the other hand, there is again only one
1017: 1-galaxy for $^{*}\!G^{1}$. $\Box$
1018:
1019: {\bf Example 9.4.} The distances in the three preceding examples
1020: can be fully defined by paths.
1021: So, let us now present an example where walks are needed. The 1-graph
1022: $G^{1}$ of Figure 3 illustrates one such case. It consists of an
1023: infinite sequence of 0-subgraphs, each of which is an
1024: infinite series connections of four-branch subgraphs, each
1025: in a diamond configuration, as shown. To save words,
1026: we shall refer to such an infinite series connection as a ``chain.''
1027: The chain starting at the 0-node $x_{k}^{0}$ will be denoted by $C_{k}$
1028: $(k=0,1,2,\ldots)$. Each $C_{k}$ is a 0-graph; it does not contain
1029: any 1-node. Each $C_{k}$ has uncountably many 0-tips. One 0-tip
1030: has a representative 0-path starting at $x_{k}^{0}$,
1031: proceeding along the left-hand sides of
1032: the diamond configurations, and reaching the 1-node $x_{k}^{1}$.
1033: Another 0-tip has a representative 0-path
1034: that proceeds along the right-hand sides
1035: and reaches the 1-node $x_{k+1}^{1}$. Still other 0-tips
1036: of $C_{k}$ (uncountably
1037: many of them) have representatives that pass back and forth
1038: between the two sides infinitely often
1039: to reach singleton 1-nodes; these are not shown
1040: in that figure. The chain $C_{k}$ is connected to $C_{k+1}$ through the
1041: 1-node $x_{k+1}^{1}$, as shown. Note that there is no path connecting, say,
1042: $x_{k}^{0}$ to $x_{m}^{0}$ when $m-k\geq 2$, but there is such a walk.
1043:
1044: Each $C_{k}$ is a 0-section, and its enlargement $^{*}\!C_{k}$
1045: has infinitely many 0-galaxies. Also, the 1-nodes
1046: $x_{k}^{1}$ together produce infinitely many 0-galaxies,
1047: each being a single 1-hypernode. As before, the nodal
1048: 0-galaxies comprise a partition of $^{*}\!X^{0}\,\cup\,^{*}\!X^{1}$.
1049:
1050: On the other hand, the enlargement $^{*}\!G^{1}$ of the 1-graph
1051: $G^{1}$ of Figure 3 has infinitely many 1-galaxies. Its principal
1052: one is a copy of $G^{1}$. Each of the other 1-galaxies is
1053: also a copy of $G^{1}$ except that it extends infinitely in both
1054: directions---infinitely to the left and infinitely to the right.
1055: Here, too, the nodal 1-galaxies comprise a partitioning of
1056: $^{*}\!X^{0}\,\cup\,^{*}\!X^{1}$, but a coarser one. $\Box$
1057:
1058: These examples indicate that the enlargements of 1-graphs
1059: can have rather complicated structures.
1060:
1061: \section{Locally 1-Finite 1-Graphs and a Property of Their Enlargements}
1062:
1063: In general, $^{*}\!G^{1}$ has 1-galaxies other than its principal 1-galaxy.
1064: One circumstance where this occurs is when $^{*}\!G^{1}$
1065: is locally finite in certain way, which we will explicate
1066: below.
1067:
1068: We need some
1069: more definitions. Two 1-nodes of $G^{1}$ are
1070: said to be 1-{\em wadjacent} if they are incident to the same 0-section.
1071: A 1-node will be called a {\em boundary 1-node} if it is incident to
1072: two or more 0-sections. $G^{1}$ will be called {\em locally
1073: 1-finite} if each of its 0-sections has only finitely many incident
1074: boundary 1-nodes.\footnote{Note that a 0-section in a locally 1-finite 1-graph
1075: may have infinitely many incident 1-nodes that are not boundary 1-nodes.
1076: Also, this definition of locally 1-finiteness does not prohibit
1077: 0-nodes of infinite degree.}
1078:
1079: {\bf Lemma 10.1.} {\em Let $x^{1}$ be a boundary 1-node. Then,
1080: any 1-walk that passes through $x^{1}$ from any 0-section $S^{0}_{1}$
1081: incident to $x^{1}$ to any other 0-section $S^{0}_{2}$
1082: incident to $x^{1}$ must have a length no less than $\omega$.}
1083:
1084: {\bf Proof.} The only way such a walk can have a length less than $\omega$
1085: (i.e., a length equal to a natural number) is if it avoids traversing a 0-tip
1086: in $x^{1}$. But, this means that it passes through two branches incident
1087: to a 0-node in $x^{1}$. But, that in turn means that $S^{0}_{1}$ and
1088: $S^{0}_{2}$ cannot be different 0-sections. $\Box$
1089:
1090: Remember that $G^{1}$ is called
1091: 1-wconnected if, for every two nodes of $G^{1}$, there is a 0-walk
1092: or 1-walk that reaches those two nodes.
1093:
1094: {\bf Lemma 10.2.} {\em Any two 1-nodes $x^{1}$ and $y^{1}$
1095: that are 1-wconnected but are not 1-wadjacent must satisfy
1096: $d(x^{1},y^{1})\geq \omega$.}
1097:
1098: {\bf Proof.} Any walk 1-wconnecting $x^{1}$ and $y^{1}$ must
1099: pass through at least one boundary 1-node different from $x^{1}$ and $y^{1}$
1100: while passing from one 0-section to another 0-section.
1101: Therefore, that walk must be a 1-walk. By Lemma 10.1,
1102: its length is no less than $\omega$. Since this is true
1103: for every such walk, our conclusion follows. $\Box$
1104:
1105: The next theorem mimics Theorem 3.7 but at the rank 1.
1106:
1107: {\bf Theorem 10.3.} {\em Let $G^{1}$ be locally 1-finite and
1108: 1-wconnected and have infinitely many boundary 1-nodes.
1109: Then, given any 1-node $x_{0}^{1}$ of $G^{1}$, there is
1110: a one-ended 1-walk $W^{1}$ starting at $x_{0}^{1}$:
1111: \[W^{1}\;=\;\langle x_{0}^{1},W_{0}^{0},x_{1}^{1},W_{1}^{0},
1112: \ldots,x_{m}^{1},W_{m}^{0},\ldots\rangle \]
1113: such that there is a subsequence of 1-nodes $x_{m_{k}}^{1}$,
1114: $k=1,2,3,\ldots$, satisfying
1115: $d(x_{0}^{1},x_{m_{k}}^{1})\,\geq\,\omega\cdot k$.}
1116:
1117: {\bf Proof.} $x^{1}_{0}$ need not be a boundary 1-node, but it will be
1118: 1-wadjacent to only finitely many boundary 1-nodes because of
1119: local 1-finiteness and 1-wconnectedness. Let $X_{0}$ be the
1120: nonempty finite set of those boundary 1-nodes.
1121: For the same reasons, there is a nonempty finite
1122: set $X_{1}$ of boundary 1-nodes,
1123: each being 1-wadjacent to some 1-node in $X_{0}$ but not 1-wadjacent
1124: to $x_{0}^{1}$. By Lemma 10.2, for each $x^{1}\in X_{2}$, we have
1125: $d(x_{0}^{1},x^{1})\geq\omega$.
1126: In general, for each $k\in\N$, $k\geq 2$,
1127: there is a nonempty finite set $X_{k}$ of boundary 1-nodes, each
1128: being 1-wadjacent to some 1-node in $X_{k-1}$ but not 1-wadjacent to any
1129: of the 1-nodes in $\cup_{l=0}^{k-2} X_{l}$.
1130: By Lemma 10.2 again, for any such $x^{1}\in X_{k}$, we have
1131: $d(x_{0}^{1},x^{1})\geq \omega\cdot k$.
1132:
1133: We now adapt the proof of K\"{o}nig's lemma: From each of the
1134: infinitely any boundary 1-nodes in $G^{1}$, there is a 1-walk reaching that
1135: boundary 1-node and also reaching $x_{0}^{1}$.
1136: Thus, there are infinitely many 1-walks starting at $x_{0}^{1}$
1137: and passing through one of the 1-nodes in $X_{0}$, say, $x_{m_{0}}^{1}$.
1138: Among those 1-walks, there are again infinitely many 1-walks
1139: passing through one of the 1-nodes in $X_{1}$, say, $x_{m_{1}}^{1}$.
1140: Continuing in this say, we find an infinite sequence
1141: $\langle x_{m_{1}}^{1}, x_{m_{2}}^{1},x_{m_{3}}^{1},\ldots\rangle$
1142: of 1-nodes occurring in a one-ended 1-walk starting at
1143: $x_{0}^{1}$ and such that $d(x_{0}^{1},x_{m_{k}}^{1})\geq\omega
1144: \cdot k$. $\Box$
1145:
1146: {\bf Corollary 10.4.} {\em Under the hypothesis of Theorem 10.3, the enlargement
1147: $^{*}\!G^{1}$ of $G^{1}$ has at least one
1148: 1-hypernode not in its principal galaxy $\Gamma_{0}^{1}$ and
1149: thus at least one
1150: 1-galaxy $\Gamma^{1}$
1151: different from its principal 1-galaxy $\Gamma_{0}^{1}$.}
1152:
1153: {\bf Proof.} Set ${\bf x}^{1} =[\langle x_{0}^{1},x_{m_{0}}^{1},
1154: x_{m_{1}}^{1},\ldots\rangle]= [x_{m_{n}}^{1}]$,
1155: where the $x_{m_{n}}^{1}$ are the 1-nodes
1156: specified in the preceding proof (replace $k$ by $n$).
1157: With ${\bf x}_{0}^{1}$
1158: being the standard 1-hypernode corresponding to $x_{0}^{1}$, we have by
1159: Theorem 10.3 that ${\bf d}({\bf x}_{0}^{1},{\bf x}^{1})\geq
1160: [\omega\cdot n]$. Hence, ${\bf x}^{1}$ is not 1-limitedly distant
1161: from ${\bf x}_{0}^{1}$ and thus must reside in a 1-galaxy $\Gamma^{1}$ different
1162: from $\Gamma_{0}^{1}$. $\Box$
1163:
1164: \section{When $^{*}\!G^{1}$ Has a 1-Hypernode Not in Its Principal Galaxy}
1165:
1166: We are at last ready to extend the results of Section 4 to the
1167: rank 1 of transfiniteness. The ideas are the much
1168: same as those of Section 4 except for the fact that the proof
1169: of Theorem 4.2 cannot be extended to the present case.
1170: This is because transfinite ordinals cannot be identified
1171: as being even or odd. We need another proof.
1172:
1173: In this section $G^{1}$ is 1-wconnected and has an infinity
1174: of boundary 1-nodes, but $G^{1}$ need not be locally finite.
1175: Let $\Gamma^{1}_{a}$ and $\Gamma^{1}_{b}$ be two 1-galaxies of $^{*}\!G^{1}$
1176: that are different from the principal 1-galaxy $\Gamma^{1}_{0}$.
1177: We say that $\Gamma^{1}_{a}$ {\em is closer to $\Gamma^{1}_{0}$
1178: than is $\Gamma^{1}_{b}$} and that $\Gamma^{1}_{b}$ {\em is further away
1179: from $\Gamma^{1}_{0}$ than is $\Gamma^{1}_{a}$} if there are a
1180: ${\bf y}=[y_{n}]$ in $\Gamma^{1}_{a}$ and a ${\bf z}=[z_{n}]$ in
1181: $\Gamma^{1}_{b}$ such that, for some ${\bf x}=[x_{n}]$ in
1182: $\Gamma^{1}_{0}$ and for
1183: every $m\in\N$,
1184: \[ \{n\!: d(z_{n},x_{n})-d(y_{n},x_{n})\;\geq\omega\cdot m\}\;\in\;{\cal F}. \]
1185: (The ranks of ${\bf x}$, ${\bf y}$, and ${\bf z}$ may now be either 0 or 1.)
1186:
1187: Any set of 1-galaxies for which every two of them, say,
1188: $\Gamma^{1}_{a}$ and $\Gamma^{1}_{b}$ satisfy these
1189: conditions will be said to be
1190: {\em totally ordered according to their closeness
1191: to} $\Gamma^{1}_{0}$. Here, too, the conditions for a total
1192: ordering are readily shown.
1193:
1194: {\bf Lemma 11.1.} {\em These definitions are independent of the
1195: representative sequences $\langle x_{n}\rangle$, $\langle y_{n}\rangle$,
1196: and $\langle z_{n}\rangle$ chosen for ${\bf x}$, ${\bf y}$, and
1197: ${\bf z}$.}
1198:
1199: The proof of this lemma is the same as that of Lemma 4.1
1200: except that
1201: the rank 0 is replaced by the transfinite rank 1. For instance,
1202: the natural number $m$ is now replaced by the ordinal
1203: $\omega\cdot m$.
1204:
1205: {\bf Theorem 11.2.} {\em If $^{*}\!G^{1}$ has a hypernode ${\bf v}=[v_{n}]$
1206: (of either rank 0 or rank 1) that is not in its
1207: principal 1-galaxy $\Gamma^{1}_{0}$, then there exists a two-way
1208: infinite sequence of 1-galaxies totally ordered according to
1209: their closeness to $\Gamma^{1}_{0}$ and with $\bf v$
1210: being in one of those galaxies.}
1211:
1212: {\bf Proof.} In this proof, we use the fact that between any two nodes
1213: in a 1-graph there exists a geodesic walk terminating at those nodes;
1214: that is, the length of the walk is equal to the wdistance
1215: between those nodes. This is a consequence of the facts
1216: that the walks terminating at those nodes have ordinal lengths
1217: and that any set of ordinals is well-ordered and thus has a least ordinal.
1218: That least ordinal must be the length of at least one walk terminating
1219: at those nodes, for otherwise the minimum of the
1220: walk-lengths would be larger.
1221:
1222: As before, let ${\bf x}=[\langle x,x,x,\ldots\rangle]$
1223: be a standard hypernode in $\Gamma_{0}^{1}$.
1224: ${\bf x}$ can be of either rank 0 or 1. Since $\bf v$ is not
1225: in $\Gamma_{0}^{1}$, we have that for each $m\in\N$
1226: \begin{equation}
1227: \{n\!: d(x,v_{n})\,>\, \omega\cdot m\}\;\in\;{\cal F} \label{11.1}
1228: \end{equation}
1229: For each $n\in\N$, if $d(x,v_{n})< \omega\cdot 6$, set $u_{n}=x$, but,
1230: if $d(x,v_{n})\geq \omega\cdot 6$, choose $u_{n}$ such that
1231: \begin{equation}
1232: d(x,v_{n})\;\leq\; d(x,u_{n})\cdot 3\;\leq\; d(x,v_{n})\cdot 2 \label{11.2}
1233: \end{equation}
1234: That the latter can be done can be seen as follows.
1235:
1236: Choose a geodesic 1-walk $W^{1}$ terminating at $x$ and $v_{n}$.
1237: Remember that $W^{1}$, as given by (\ref{6.2}),
1238: is incident to each of its nonterminal
1239: 1-nodes through at least one 0-tip, as was
1240: asserted by Condition 3 of the definition of $W^{1}$.
1241: Moreover, the transition through each 0-tip contributes $\omega$
1242: to the length of $W^{1}$. Upon tracing $W^{1}$ from $x$
1243: toward $v_{n}$, we must encounter at least two 1-nodes,
1244: both of which are neither closer to $x$ by one-third of the number
1245: of 0-tips traversed by $W^{1}$ nor
1246: further away from $x$ by two-thirds of the number of 0-tips
1247: traversed by $W^{1}$. A node on $W^{1}$ between those two 1-nodes can be
1248: chosen as $u_{n}$.
1249:
1250: Suppose there is a $k\in\N$ such that $\{n\!: d(x,u_{n})\leq \omega\cdot k\}
1251: \in {\cal F}$. By the left-hand inequality of (\ref{11.2}),
1252: \[ \{n\!: d(x,v_{n})\,\leq\, (\omega\cdot k)\cdot 3\}\;\supseteq\;
1253: \{n\!: d(x,u_{n})\,\leq\, \omega\cdot k\}\;\in\;{\cal F}. \]
1254: Hence, the left-hand set is a member of $\cal F$, in
1255: contradiction to (\ref{11.1}). (These sets cannot both be
1256: in the ultrafilter $\cal F$.) Therefore, ${\bf u}=[u_{n}]$
1257: satisfies (\ref{11.1}) for every $m\in\N$ when $v_{n}$
1258: is replaced by $u_{n}$; that is, $\bf u$ is in a galaxy
1259: different
1260: from the principal 1-galaxy $\Gamma_{0}^{1}$.
1261:
1262: Furthermore, by the right-hand inequality of (\ref{11.2}),
1263: \[ d(x,u_{n})\;\leq\; (d(x,v_{n})\,-\,d(x,u_{n}))\cdot 2. \]
1264: Suppose there exists a $j\in\N$ such that
1265: \[ \{n\!: d(x,v_{n})\,-\, d(x,u_{n})\,\leq\, \omega\cdot j\}\;\in\;{\cal F}. \]
1266: Then,
1267: \[ \{n\!: d(x,u_{n})\,\leq\,(\omega\cdot j)\cdot 2\}\;\supseteq\;
1268: \{n\!: d(x,v_{n})\,-\, d(x,u_{n})\,\leq\, \omega\cdot j\}\;\in\;{\cal F} \]
1269: So, the left-hand set is in $\cal F$, in contradiction to our previous
1270: conclusion that $\bf u$ satisfies (\ref{11.1})
1271: with $v_{n}$ replaced by $u_{n}$. We can conclude that
1272: $\bf u$ and $\bf v$ are in different 1-galaxies $\Gamma_{a}^{1}$
1273: and $\Gamma_{b}^{1}$ respectively, with $\Gamma_{a}^{1}$ closer
1274: to $\Gamma_{0}^{1}$ than is $\Gamma_{b}^{1}$.
1275:
1276: We can now repeat this argument with $\Gamma_{b}^{1}$ replaced
1277: by $\Gamma_{a}^{1}$ and with ${\bf u}=[u_{n}]$ playing the role that
1278: ${\bf v}=[v_{n}]$ played
1279: to find still another galaxy $\Gamma_{a}^{1}\prime$ different from
1280: $\Gamma_{0}^{1}$ and closer to $\Gamma_{0}^{1}$ than is $\Gamma_{a}^{1}$.
1281: Continual repetitions yield an infinite sequence of galaxies
1282: indexed by, say, the negative integers and totally ordered
1283: by their closeness to $\Gamma_{0}^{1}$.
1284:
1285: The rest of the proof continues just like the argument for
1286: Theorem 4.2 that establishes a sequence of 1-galaxies progressively
1287: further away from $\Gamma_{0}^{1}$ than is $\Gamma_{b}^{1}$.
1288: In this case, the natural number $m$ is replaced by the ordinal
1289: $\omega\cdot m$; also, the last
1290: $n$ in (\ref{4.2}) is replaced by $\omega\cdot n$.
1291: $\Box$
1292:
1293: Finally, by mimicking the proof of Theorem 4.3, we can prove
1294:
1295: {\bf Theorem 11.3.} {\em Under the hypothesis of Theorem 11.2, the set of
1296: 1-galaxies of $^{*}\!G^{1}$ is partially ordered according
1297: to the closeness of the 1-galaxies to $\Gamma^{1}_{0}$.}
1298:
1299: \section{Extensions to Higher Ranks of Transfiniteness}
1300:
1301: The extension of these results to the enlargements of
1302: transfinite graphs of any natural-number rank is quite similar
1303: to what we have presented. The ideas are the same, but the
1304: notations and the details of the arguments are
1305: somewhat more complicated.
1306: Moreover, further complications arise with the extension to the
1307: arrow rank $\vec{\omega}$ of transfiniteness.
1308: Extensions to still higher ranks then proceed in much the same way.
1309: All this is explicated in the technical report \cite{gal2},
1310: which can also be found in the archival web site, www.arxiv.org, under
1311: Mathematics, Zemanian.
1312:
1313: \begin{thebibliography}{99}
1314:
1315: \bibitem{ab} A. Abian, {\em The Theory of Sets and Transfinite Arithmetic},
1316: W.B. Saunders Company, Philadelphia, Pennsylvania, 1965.
1317: \bibitem{be} C. Berge, {\em Graphs and Hypergraphs},
1318: North Holland Publishing Co., Amsterdam, 1973.
1319: \bibitem{go} R. Goldblatt, {\em Lectures on the Hyperreals},
1320: Springer, New York, 1998.
1321: \bibitem{wi} R.J. Wilson, {\em Introduction to Graph Theory},
1322: Academic Press, New York, 1972.
1323: \bibitem{tgen} A.H. Zemanian, {\em Transfiniteness for Graphs,
1324: Electrical Networks, and Random Walks}, Birkhauser-Boston,
1325: Cambridge, Massachusetts, 1996.
1326: \bibitem{gn} A.H. Zemanian, {\em Graphs and Networks: Transfinite and
1327: Nonstandard}, Birkhauser-Boston, Cambridge, Massachusetts,
1328: 2004.
1329: \bibitem{gal2} A.H. Zemanian, {\em The Galaxies of Nonstandard Enlargements
1330: of Transfinite Graphs of Higher Ranks: II}, CEAS Technical Report 820,
1331: University at Stony Brook, Stony Brook, NY 11794, April 2005.
1332:
1333: \end{thebibliography}
1334:
1335:
1336: \end{document}
1337: