1: \documentstyle[12pt,leqno,amssymb,graphics]{amsart}
2:
3: %Immersions of the 2-sphere in 3-manifolds
4:
5: \setlength{\topmargin}{-2mm}
6: \setlength{\oddsidemargin}{-2mm}
7: \setlength{\evensidemargin}{-2mm}
8: \setlength{\textwidth}{6.5in}
9: \setlength{\textheight}{9.3in}
10: \renewcommand{\baselinestretch}{1.5}
11:
12: \newtheorem{thm}{Theorem}[section]
13: \newtheorem{lemma}[thm]{Lemma}
14: \newtheorem{prop}[thm]{Proposition}
15: \newtheorem{cor}[thm]{Corollary}
16:
17: \theoremstyle{definition}
18: \newtheorem{dfn}[thm]{Definition}
19: \theoremstyle{remark}
20: \newtheorem{remark}[thm]{Remark}
21:
22: %\renewcommand{\theequation}{\thesection.\arabic{equation}}
23:
24: \begin{document}
25:
26:
27: \newcommand{\ct}{\cite}
28: \newcommand{\pr}{\protect\ref}
29: \newcommand{\su}{\subseteq}
30: \newcommand{\pa}{{\partial}}
31: \newcommand{\im}{{Imm(F,M)}}
32: \newcommand{\lm}{{\lambda}}
33: \newcommand{\hf}{{1 \over 2}}
34: \newcommand{\p}{{\pi_a}}
35: \newcommand{\ep}{{\epsilon}}
36: \newcommand{\tc}{{\mathfrak{T}}}
37: \newcommand{\Q}{{\Bbb Q}}
38: \newcommand{\R}{{\Bbb R}}
39: \newcommand{\Z}{{\Bbb Z}}
40: \newcommand{\E}{{{\Bbb R}^3}}
41: \newcommand{\B}{{\Bbb B}}
42: \newcommand{\G}{{\Bbb G}}
43:
44: \newcommand{\A}{{\mathcal{A}}}
45: \newcommand{\C}{{\mathcal C}}
46: \newcommand{\I}{{\mathrm{Id}}}
47:
48: \newcommand{\hM}{{\widehat{M}}}
49: \newcommand{\gr}{{\Z[\p]}}
50:
51: \newcounter{numb}
52:
53:
54: \title{Immersions of surfaces into aspherical 3-manifolds}
55: \author{Tahl Nowik}
56: \address{Department of Mathematics, Bar-Ilan University,
57: Ramat-Gan 52900, Israel.}
58: \email{tahl@@math.biu.ac.il}
59: \urladdr{http://www.math.biu.ac.il/$\sim$tahl}
60: \date{October 26, 2005}
61: \thanks{Partially supported by the Minerva Foundation}
62:
63:
64: \begin{abstract}
65: We study finite order invariants of null-homotopic immersions of a closed orientable
66: surface into an aspherical orientable 3-manifold.
67: We give the foundational constructions, and classify all order one invariants.
68: \end{abstract}
69:
70:
71: \maketitle
72:
73: \section{Introduction}
74: Finite order invariants of immersions of a closed orientable surface
75: into $\E$ have been introduced in \ct{o},
76: where all order 1 invariants have been classified. Explicit formulae for the majority of order
77: 1 invariants have been given in \ct{a},\ct{f}. All higher order invariants have been classified
78: in \ct{h}, and the analogue of all the above for non-orientable surfaces has appeared in \ct{n}.
79: A first step in the study of finite order invariants of immersions of surfaces into general 3-manifolds
80: appears in \ct{q} where one specific order 1 invariant has been studied.
81:
82: In the present work we study finite order invariants of null-homotopic immersions of a closed orientable
83: surface $F$ into an aspherical orientable 3-manifold $M$.
84: The immersions being null-homotopic
85: will enable us to lift the immersions to the universal covering $\hM$ of $M$, and
86: $M$ being aspherical will guarantee a simple form for $H_2(\hM - A)$ for discrete subset $A \su \hM$.
87: We develop the foundations for the study of finite order invariants
88: in our setting, as has been done in \ct{o} for the case $M=\E$ (and analogous to chord diagrams
89: and the 1-term and 4-term relations in knot theory). We then classify all order 1 invariants.
90:
91: Revisions to this paper will appear at
92: www.math.biu.ac.il/$\sim$tahl/publications.html
93:
94:
95:
96: \section{The universal covering space}\label{lb}
97:
98: Let $M$ be an aspherical orientable 3-manifold, i.e. $\pi_n(M)=0$ for all $n \geq 2$
99: (a sufficient condition being that $\pi_2(M)=\pi_3(M)=0$).
100: Let $r:\hM \to M$ be the universal covering, so $\hM$ is contractible.
101: For $a \in M$ denote $\p=\pi_1(M,a)$.
102: If $\gamma$ is a path in $M$ with initial point $a$ and if $x \in r^{-1}(a)$
103: then we denote the lift of $\gamma$ to $\hM$ with initial point $x$ by $\gamma^x$.
104: If $p \in M$ is another point and $\gamma$ is a path from $a$ to $p$ then we define a
105: bijection $F^x_\gamma : \p \to r^{-1}(p)$ by
106: $F^x_\gamma(\phi) = (\phi * \gamma)^x (1)$, where $*$ denotes concatenation from left to right.
107:
108: We choose $a \in M$, $x \in r^{-1}(a)$ and for each $p \in M$ we choose a path
109: $\gamma_p$ from $a$ to $p$. We then label the points in each $r^{-1}(p)$ by elements
110: of $\p$ via the bijection $F^x_{\gamma_p}$.
111: We now check in what way this labeling depends on our choices.
112: So let $b \in M$, $y \in r^{-1}(b)$ and for each $p \in M$, $\delta_p$ a path from $b$ to $p$,
113: be another such choice.
114: In order to identify the elements of $\p$ with those of $\pi_b$ we need to make one additional
115: choice, the dependence on which will be apparent as well, namely we choose a path $\mu$ from $a$
116: to $b$ and identify $\p$ with $\pi_b$ by $\phi \mapsto \bar{\mu} * \phi * \mu$ where $\bar{\mu}$
117: is the path inverse to $\mu$ (i.e. $\bar{\mu}(t)=\mu(1-t)$).
118: Finally, let $s$ be the unique (up to homotopy) path in $M$ from $a$ to $b$ such that $s^x(1)=y$.
119: Now for each $p \in M$:
120: $F^y_{\delta_p} (\bar{\mu} * \phi * \mu) = (\bar{\mu} * \phi * \mu * \delta_p)^y (1)
121: = (\bar{\mu} * \phi * \mu * \delta_p * \bar{\gamma_p} * \gamma_p)^{s^x(1)} (1)
122: = (s * \bar{\mu} * \phi * \mu * \delta_p * \bar{\gamma_p} * \gamma_p)^x (1)
123: = F^x_{\gamma_p} ( (s * \bar{\mu}) * \phi * (\mu * \delta_p * \bar{\gamma_p}))$.
124: So given the identification between $\p$ and $\pi_b$ determined by $\mu$, the labeling
125: of the points in $r^{-1}(p)$ have changed by left
126: multiplication by $s * \bar{\mu}$ and right multiplication by
127: $\mu * \delta_p * \bar{\gamma_p}$. Note that $s * \bar{\mu}$ depends only
128: on $a,b,x,y,\mu$ and so we get left multiplication by the \emph{same} element for all points $p$.
129: We summarize this in the following:
130:
131: \begin{prop}\label{dep}
132: The labeling of the points of $r^{-1}(p)$ by elements of $\p$ is well defined
133: up to left multiplication by one common element in $\p$ for all $p$,
134: and right multiplication by an element in $\p$ which may depend on $p$.
135: \end{prop}
136:
137: Once the dependence on choices has been established, we fix $a,x, \{\gamma_p\}_{p \in M}$ once and for
138: all, and so the labeling of the points in $r^{-1}(p)$ for each $p \in M$ is from now on fixed.
139: Note that this labeling is necessarily \emph{not} locally constant.
140:
141:
142: There will be two families of bijections that we will have occasion to use, and we will now
143: see how they are expressed in terms of the labeling.
144: First let $D : \hM \to \hM$ be a deck transformation, then the restriction
145: of $D$ to each $r^{-1}(p)$ is a bijection onto itself.
146: Let $\psi \in \p$ be the unique element such that $\psi^x(1)=D(x)$.
147: Then for any $\phi \in \p$ and any $p \in M$,
148: $D(F^x_{\gamma_p}(\phi)) = D((\phi * {\gamma_p})^x (1)) = (\phi * {\gamma_p})^{\psi^x(1)} (1)
149: =(\psi * \phi * {\gamma_p})^x(1) = F^x_{\gamma_p}(\psi * \phi)$, that is, in terms of the labeling,
150: $D$ is given by common left multiplication by $\psi$. (By \emph{common} we mean as above,
151: that it is the \emph{same} $\psi$ for all $p$.)
152: Secondly, given $p,q \in M$ let $\delta$ be a path from $p$ to $q$, then $\delta$
153: defines a bijection $G_\delta : r^{-1}(p) \to r^{-1}(q)$ given by
154: $G_\delta(y) = \delta^y(1)$. We have for any $\phi \in \p$,
155: $G_\delta (F^x_{\gamma_p}(\phi)) = G_\delta((\phi * \gamma_p)^x(1)) =
156: \delta^{(\phi * \gamma_p)^x(1)}(1) = (\phi * \gamma_p * \delta)^x(1) =
157: (\phi * (\gamma_p * \delta * \bar{\gamma_q}) * \gamma_q)^x(1) =
158: F^x_{\gamma_q}( \phi * (\gamma_p * \delta * \bar{\gamma_q}))$, that is, in terms of the labeling,
159: $G_\delta$ is given by right multiplication by $\gamma_p * \delta * \bar{\gamma_q}$.
160:
161: It is also clear from the above calculations that any common left multiplying element,
162: and any right multiplying element, can be realized by appropriately choosing $D$ and $\delta$ respectively.
163: We summarize this in the following:
164:
165: \begin{prop}\label{dk}
166: \begin{enumerate}
167: \item The effect of a deck transformation is a common left multiplication, and any such common left
168: multiplication can be realized by a deck transformation.
169: \item The effect of $G_\delta$ is right multiplication, and any such right multiplication
170: can be realized by some $\delta$ from $p$ to $q$.
171: \end{enumerate}
172: \end{prop}
173:
174: We have noted that our labeling is not locally constant. By Proposition \pr{dk}(2) we see that any such
175: discontinuity is always given by right multiplication.
176:
177:
178: Denote by $\gr$ the group ring of $\p$ with coefficients in $\Z$. We will in fact not use the full ring structure,
179: but only the left and right actions of $\p$ on $\gr$. For each $p \in M$,
180: $H_2(\hM - r^{-1}(p))$ is a free Abelian group with basis in one-to-one correspondence with
181: $r^{-1}(p)$. This is true by a Mayer-Vietoris sequence, since $\hM$ is contractible.
182: So via our labeling, for each $p \in M$ we may identify $H_2(\hM - r^{-1}(p))$ with $\gr$.
183: Also note $H_2(\hM - r^{-1}(p)) = \pi_2(\hM - r^{-1}(p))$.
184:
185: Any deck transformation induces an automorphism of $H_2(\hM - r^{-1}(p))$ for each $p$. It follows from
186: Proposition \pr{dk}(1) that in terms of the identification of $H_2(\hM - r^{-1}(p))$ with $\gr$,
187: this automorphism (of $\gr$ as an Abelian group)
188: is given by left multiplication by some common $\psi \in \p$.
189: Similarly, given $p,q \in M$, a path $\delta$ from $p$ to $q$ naturally induces an isomorphism from
190: $H_2(\hM - r^{-1}(p))$ to $H_2(\hM - r^{-1}(q))$, and it follows from Proposition \pr{dk}(2) that
191: in terms of the identifications with $\gr$, this isomorphism is given by right multiplication by some element
192: of $\p$.
193:
194:
195: \section{Co-oriented AB equivalences}\label{imm}
196:
197: Let $M$ be an oriented aspherical 3-manifold,
198: let $F$ be a closed oriented surface, and let $\A \su \im$ be a regular homotopy class
199: of immersions $i:F \to M$ which are null-homotopic.
200: A CE point of an immersion $i:F \to M$ is a point of self intersection
201: of $i$ for which the local stratum in $\im$ corresponding to the
202: self intersection, has codimension one.
203: We distinguish four types of CEs which we name
204: $E, H, T, Q$.
205: In the notation of \ct{hk} they are respectively
206: $A_0^2|A_1^+$, $A_0^2|A_1^-$, $A_0^3|A_1$, $A_0^4$.
207: The four types may be demonstrated by the following local models, where letting $\lm$ vary,
208: we obtain a 1-parameter family of immersions which is transverse to the
209: given codim 1 stratum, intersecting it at $\lm=0$.
210:
211:
212:
213:
214: $E$: \ \ $z=0$, \ \ $z=x^2+y^2+\lm$.
215:
216: $H$: \ \ $z=0$, \ \ $z=x^2-y^2+\lm$.
217:
218: $T$: \ \ $z=0$, \ \ $y=0$, \ \ $z=y+x^2+\lm$.
219:
220: $Q$: \ \ $z=0$, \ \ $y=0$, \ \ $x=0$, \ \ $z=x+y+\lm$.
221:
222: See Figure \pr{hoi}, which corresponds to some small $\lm > 0$. If $R$ is one of the above four CE types,
223: then we denote by $|R|$ the number of sheets involved in the given configuration, that is,
224: $|E|=|H|=2$, $|T|=3$, $|Q|=4$.
225: Let $I_n \su \A$ be
226: the space of all immersions with precisely $n$ CEs, in particular, $I_0$ is the space of all stable immersions.
227:
228: \begin{figure}[t]
229: \scalebox{0.7}{\includegraphics{HOI.eps}}
230: \caption{The four CE types}\label{hoi}
231: \end{figure}
232:
233: A choice of one of the two sides of the local codim 1 stratum at a given point of the
234: stratum, is represented by the choice of $\lm<0$ or $\lm>0$ in the formulae above. We will
235: refer to such a choice as a
236: \emph{co-orientation} for the configuration of the self intersection.
237: For types $E$ and $T$, the configuration of the self intersection at the two sides
238: of the stratum is distinct, namely, for $\lm<0$ there is an additional
239: 2-sphere in the image of the immersion, and we permanently
240: choose this side ($\lm<0$) as our positive side for the co-orientation.
241: For types $H$ and $Q$, the configuration of the self intersection on the two
242: sides of the strata is the same.
243:
244: We will now further discuss the symmetries of our four configurations.
245: A symmetry of a CE configuration is an orientation preserving diffeomorphism
246: from a neighborhood of the CE in $M$ to itself, which maps sheets onto sheets.
247: A symmetry $h$ will be said to \emph{preserve} the co-orientation of the CE if the image under $h$ of motion
248: into some side of the stratum, is again motion
249: into the same side, (after identifying each sheet with its image).
250: The comment of the previous paragraph may now be stated as follows:
251: All symmetries of $E,T$ configurations preserve co-orientation, whereas configurations $H,Q$
252: admit co-orientation reversing symmetries.
253: We further notice, that configurations $E,H,T$ admit co-orientation preserving symmetries
254: that realize any given permutation of the sheets, whereas for configuration $Q$, any co-orientation
255: preserving symmetry induces an even permutation of the sheets, and any co-orientation reversing
256: symmetry induces an odd permutation of the sheets. These facts may be seen by slightly resolving the
257: CE into one of the sides, and attempting to construct the diffeomorphism $h$ for the resolved configurations.
258:
259:
260: Given an immersion $i\in I_n$, an \emph{$n$-co-orientation} for $i$ is a choice
261: of co-orientation at each of the $n$ CE points $p_1, \dots , p_n$ of $i$.
262: (This is called a temporary co-orientation in \ct{o},\ct{h},\ct{n}.)
263: A \emph{proper} $n$-co-orientation for $i \in I_n$ is an $n$-co-orientation where at each CE of type $E$ and $T$,
264: the co-orientation chosen is the permanent one mentioned above. (This definition is slightly weaker than that
265: in \ct{h}.) For immersion $i \in I_n$, and proper $n$-co-orientation $\tc$ for $i$,
266: let $p \in M$ be one of its $n$ CE points and we define $C_p(i,\tc)$ as the symbol
267: $R^{g_1\ep_1,\dots,g_{|R|}\ep_{|R|}}_d$ which is composed of:
268:
269:
270: (1) A symbol $R \in \{ E,H,T,Q \}$ which is the configuration of the given CE.
271:
272: (2) A sequence of elements $g_1,\dots,g_{|R|} \in \p$ which is determined as follows:
273: Let $\hat{i}:F \to \hM$ be a lift of $i$. There are $|R|$ sheets of $\hat{i}(F)$ passing through points of
274: $r^{-1}(p)$
275: %(not necessarily distinct),
276: which in turn are labeled by elements $g_1,\dots,g_{|R|} \in \p$ (Section \pr{lb}).
277: If the CE is of type $E$,$H$ or $T$ then the order of the $g_j$s is arbitrarily chosen.
278: As to $Q$ configuration, we choose once and for all a rule by which an orientation on a
279: simplex determines an ordering on its four faces, up to an even permutation. Now, we resolve
280: the CE of type $Q$ positively with respect to $\tc$,
281: i.e. slightly deform $i$ (not $\hat{i}$) near $p$,
282: to obtain a stable immersion in a neighborhood of $p$,
283: and write the elements $g_1,g_2,g_3,g_4$
284: in order determined by the orientation of $M$ restricted to the small simplex created by this resolution.
285: (This ordering convention is related
286: to the symmetry properties discussed above).
287:
288: (3) To each $g_j, 1 \leq j \leq |R|$,
289: there corresponds a sign $\ep_j \in \{ +, - \}$ which is determined as follows:
290: We look at $i: F \to M$ (not the lift), and resolve the CE at $p$ according to $\tc$.
291: For types $E,T,Q$ this creates a little sphere in the image of $i$, which bounds
292: a little 3-cell $V$ in $M$ (for $Q$ this is a simplex and for $E,T$
293: recall that $\tc$ is proper).
294: Each $g_j$ we have, corresponds to a sheet involved in the CE.
295: This sheet may have $V$ on its non-preferred side,
296: (determined by the orientations of $F$ and $M$),
297: in which case we set $\ep_j=+$, or otherwise
298: $V$ is on its preferred side, in which case $\ep_j=-$.
299: For type $H$ the region $V$
300: is not bounded by the local configuration,
301: but may still be defined, e.g. for $\lm>0$ in the formula for $H$ above,
302: $V$ will be a region consisting of points close to the origin and satisfying
303: $0\leq z\leq x^2-y^2+\lm$. Now the signs $\ep_j$ are determined by the \emph{opposite} convention
304: than above, namely,
305: a sheet having $V$ on its \emph{preferred} side will have $\ep_j=+$, and if
306: $V$ is on its \emph{non-preferred} side we set $\ep_j=-$.
307: The reason for the opposite convention is to obtain slightly nicer formulae in the end.
308: (It is consistent with the conventions in \ct{o}, chosen there in a similar way for the same reason).
309:
310: (4) The subscript $d \in \gr$ which will be called the \emph{degree} of the CE,
311: is determined as follows:
312: Let $\hat{i}:F \to \hM$ be the same lift as used in (2).
313: Wherever $\hat{i}(F)$ passes through a point of
314: $r^{-1}(p)$, we push it slightly into the preferred side of $\hat{i}(F)$ in $\hM$,
315: obtaining a map $F \to \hM - r^{-1}(p)$.
316: This map represents an element of $H_2(\hM - r^{-1}(p))$ which we identify with an element $d \in\gr$
317: (Section \pr{lb}).
318:
319: If $p_1,\dots,p_n$ are the $n$ CEs of $i \in I_n$ then we define $C'(i,\tc)$
320: to be the $n$-tuple $(C_{p_1}(i,\tc),\dots,C_{p_n}(i,\tc))$, so
321: $C'(i,\tc) \in \C'_n$ where $\C'_n$ is the set of all $n$-tuples of symbols of the form
322: $R^{g_1\ep_1,\dots,g_{|R|}\ep_{|R|}}_d$.
323: In addition to the permanent choices of Section \pr{lb}, $C'$ also
324: depends on the choice of lift $\hat{i}$, and various ordering choices.
325: We thus define an equivalence relation on $\C'_n$ to be the equivalence relation generated by the
326: following operations:
327:
328: \begin{enumerate}
329: \item Any permutation of the $n$ symbols.
330: \item For $R=E,H,T$, any permutation of the $|R|$ elements
331: $g_1\ep_1,\dots,g_{|R|}\ep_{|R|}$ (each pair $g_j\ep_j$ goes together).
332: \item For $R=Q$, any \emph{even} permutation of the four elements $g_1\ep_1,\dots,g_4\ep_4$.
333: \item Given $h \in \p$ replace each one of the $n$ symbols $R^{g_1\ep_1,\dots,g_{|R|}\ep_{|R|}}_d$ of the
334: $n$-tuple, by $R^{(hg_1)\ep_1,\dots,(hg_{|R|})\ep_{|R|}}_{hd}$ (the same $h$ for all $n$ symbols).
335: \item Given $h \in \p$, replace \emph{one} of the $n$ symbols $R^{g_1\ep_1,\dots,g_{|R|}\ep_{|R|}}_d$
336: by $R^{(g_1h)\ep_1,\dots,(g_{|R|}h)\ep_{|R|}}_{dh}$.
337: (it is of course the same $h$ multiplying $g_j$ and $d$ on the right
338: within one symbol, $h$ being allowed to vary only between the $n$ symbols).
339: \end{enumerate}
340:
341: Denote by $\C_n$ the set of equivalence classes of $\C'_n$ under this equivalence relation.
342: Note that because of the \emph{common} left action in item (4), $\C_n$ is \emph{not} simply
343: the set of unordered $n$-tuples of elements of $\C_1$.
344: We now define $C(i,\tc) \in \C_n$ to be the equivalence class of $C'(i,\tc)$.
345: So $C$ is well defined, independent of all choices, including the choice of lift $\hat{i}$
346: (Proposition \pr{dk}(1)). Denote by $I^P_n$ the set of all pairs $i,\tc$ where $i \in I_n$ and
347: $\tc$ is a proper $n$-co-orientation for $i$. Then $C : I^P_n \to \C_n$ and we claim:
348:
349: \begin{lemma}\label{srj}
350: The map $C : I^P_n \to \C_n$ is surjective.
351: \end{lemma}
352:
353: \begin{pf}
354: We are given an $n$-tuple $x$ of symbols of the form $R^{g_1\ep_1,\dots,g_{|R|}\ep_{|R|}}_d$.
355: Begin with any immersion $i \in \A$ and lift it to $\hat{i}:F \to \hM$.
356: Choose points $p_1,\dots,p_n \in M$ as locations for the $n$ CEs we will construct.
357: Deform $\hat{i}$ by regular homotopy
358: so that pieces of $F$ will pass the right points of $\hM$ above each $p_k$, to produce the correct labelings
359: appearing as superscripts in the symbols in $x$. Continue the deformation near each such point so that
360: for the projection back to $M$ the right configurations will be created with the right signs, and inducing the
361: right ordering in case of CE of type $Q$.
362: Finally, move some other pieces of $F$ across the right points of
363: $\bigcup_{1 \leq k \leq n} r^{-1}(p_k)$ to obtain the correct degree $d \in \gr$.
364: The projection back to $M$ gives the desired immersion.
365: \end{pf}
366:
367:
368:
369:
370:
371: We also note that for types $H,Q$,
372: the relation between the symbols obtained for a given co-orientation
373: of a CE, and that obtained for the opposite co-orientation, is as follows:
374: For $H$, all remains the same except for the signs $\ep_j$ which are all reversed.
375: For $Q$, all signs are reversed, and in addition an odd permutation is performed on
376: $g_1\ep_1,\dots,g_4\ep_4$.
377: (The degree $d$ remains unchanged since its definition does not involve the co-orientation.)
378: We call the symbol obtained from a symbol of type $H$ or $Q$
379: in this way, the \emph{reversed} symbol.
380: (Recall that $C_p(i,\tc)$ is only defined for \emph{proper} $\tc$ and so
381: for $E,T$ the co-orientation may not be reversed.)
382:
383:
384:
385: We recall the definition of an AB equivalence, appearing in \ct{o}:
386: A regular homotopy between two immersions $i,j \in I_n$ is called an AB equivalence if
387: it is alternatingly of type A and B, where
388: \begin{enumerate}
389: \item $J_t:F\to M$ ($0\leq t \leq 1$)
390: is of type A if it is of the form
391: $J_t = U_t \circ i \circ V_t$ where
392: $i:F\to M$ is an immersion and
393: $U_t: M \to M$,
394: $V_t:F\to F$ are isotopies.
395: \item $J_t:F\to M$ ($0\leq t \leq 1$)
396: is of type B if $J_0 \in I_n$ and there are
397: little balls $B_1,\dots,B_n\su M$ centered at the $n$ CE points of $J_0$
398: such that $J_t$ fixes $U=(J_0)^{-1}(\bigcup_k B_k)$
399: and moves $F-U$ within $M - \bigcup_k B_k$.
400: \end{enumerate}
401:
402:
403: Given two immersions $i,j \in I_n$ and $n$-co-orientations $\tc,\tc'$ for $i,j$ respectively,
404: we now define \emph{co-oriented} AB equivalence, or CAB equivalence,
405: between $i,\tc$ and $j,\tc'$
406: to be an AB equivalence between
407: $i$ and $j$ which respects $\tc,\tc'$, i.e. if we carry $\tc$ at each CE of $i$
408: continuously along the AB equivalence, then we arrive at $j$ with $n$-co-orientation $\tc'$.
409:
410: We now prove:
411:
412: \begin{prop}\label{cab}
413: Let $i,j\in I_n$ and $\tc,\tc'$ $n$-co-orientations for $i,j$ respectively.
414: Then $i,\tc$ and $j,\tc'$ are CAB equivalent iff $C(i,\tc)=C(j,\tc')$.
415: \end{prop}
416:
417: \begin{pf}
418: If $i,\tc$ and $j,\tc'$ are CAB equivalent, let $J_t$ ($0 \leq t \leq 1$)
419: be the given CAB equivalence, and $\tc_t$ ($0 \leq t \leq 1$), the $n$-co-orientation carried
420: continuously along (so $\tc_1=\tc'$). If a lift $\hat{J}$ of $J$
421: is used for defining $C'(J_t,\tc_t)$ a each $t$, and all choices of orderings are also
422: carried along continuously, then by Proposition
423: \pr{dk}(2), $C'(J_t,\tc_t)$ may only change along the way by
424: right multiplication. We may still have that $J_1$ is different
425: than the lift of $j$ used to define $C'(j,\tc')$,
426: in which case we will gain a common left multiplication.
427: The choices of ordering may also be different, but all together we get
428: $C(i,\tc)=C(j,\tc')$.
429:
430: For the converse,
431: assume $C(i,\tc)=C(j,\tc')$. Since our allowed ordering of superscripts, and equivalence of ordering, all correspond
432: to the possible symmetries of the configurations, there is an ambient isotopy $U_t :M \to M$ which brings the
433: CEs of $i$ onto the CEs of $j$, such that the $n$-co-orientation $\tc$ carried continuously along,
434: coincides with $\tc'$, such that the orientations of the corresponding sheets match, (since the signs $\ep_j$
435: coincide,)
436: and such that all labelings in $\p$ and degrees in $\gr$ coincide
437: up to right multiplication by an element in $\p$ and common left multiplication by an element in $\p$.
438: Since the orientations of corresponding sheets match,
439: we may continue with a regular homotopy of the form $i \circ V_t$ for isotopy $V_t : F \to F$,
440: until we have the same discs in $F$ participating in the CEs of $i$ and $j$, and the restriction of $i$
441: and $j$ to those discs coincides precisely, and we still have matching labeling and degrees up
442: to right and common left multiplication.
443: By Proposition \pr{dk} any right multiplication can be realized by dragging a CE around some loop in $M$
444: and back to its place to match $j$, and any common left multiplication can be realized by
445: changing the choice of lift. So we perform such ambient isotopies and change of lift until we have that
446: all labelings and degrees coincide precisely. This means that now, not only $i,j$ but also their chosen lifts
447: $\hat{i},\hat{j}$ coincide on the discs participating in the CEs, and that
448: if we slightly deform $i$ and $j$ by pushing each sheet of the CEs slightly into its preferred side,
449: then $\hat{i},\hat{j} : F \to \hM - \bigcup_{1 \leq k \leq n} r^{-1}(p_k)$ represent the same element in
450: $H_2(\hM - \bigcup_{1 \leq k \leq n} r^{-1}(p_k))$.
451: (Note that
452: $H_2(\hM - \bigcup_{1 \leq k \leq n} r^{-1}(p_k)) = \bigoplus_{1 \leq k \leq n} H_2(\hM - r^{-1}(p_k))$,
453: where the projections are induced by inclusion.)
454: Form this point on we may proceed exactly as in the proof of \ct{o} Proposition 3.4, where instead of
455: working in
456: $\E$ with set of designated points $\{p_1,\dots,p_n\}$, we work in $\hM$ with set of designated points
457: $\bigcup_{1 \leq k \leq n} r^{-1}(p_k)$. We then compose the obtained regular homotopy with $r$ obtaining the
458: desired CAB equivalence in $M$.
459: \end{pf}
460:
461:
462: As already seen in \ct{o}, a CE of type $H$ or $Q$
463: may be CAB equivalent to itself with the opposite co-orientation,
464: which means that the stratum corresponding to this CE in $\im$ is one sided.
465: From Proposition \pr{cab} we see that this happens iff the corresponding symbol is equivalent
466: to its reversed symbol. For this to happen clearly the number of $+$ and $-$ signs should be equal,
467: as this number is preserved under equivalence. So a one sided stratum may occur only for symbols of
468: the form $H^{g_1+,g_2-}_d$ and $Q^{g_1+,g_2+,g_3-,g_4-}_d$. If all $g_j$ are equal then we clearly
469: do get a symbol equivalent to its reversed, and so a one sided stratum.
470: In \ct{o} where $\pi_1$ is trivial ($M=\E$),
471: it is thus clear which are the one sided strata. But for general aspherical manifold this is intricately
472: related to the structure of $\pi_1(M)$. In the following two examples we show a case where the stratum
473: for $H^{g_1+,g_2-}_d$ is one sided though $g_1 \neq g_2$, and then a case of an $M$
474: where the stratum is one sided only when $g_1=g_2$, though $\pi_1(M)$ is nontrivial.
475:
476:
477: Let $S$ be the connect sum of four projective planes, with $a,b,c,d \in S$ the standard four loops
478: for which $\pi_1(S)=\langle a,b,c,d \ | \ a^2b^2c^2d^2 = 1 \rangle$. Let $M$ be the 3-manifold
479: obtained from $S \times [0,1]$ by attaching a 2-handle to the loop $b d$ in $S \times \{ 1 \}$.
480: Then $\pi_1(M)=\langle a,b,c \ | \ a^2b^2c^2b^{-2}=1 \rangle$ and one can show that $M$ is
481: indeed aspherical. We look at the symbol $H^{b^2 + , ab^2c -}_0 \in \C'_1$ (the subscript is the zero element of
482: $\Z[\pi_1]$). We note that $b^2 \neq a b^2 c$, e.g. by looking at the quotient obtained
483: by adding the relations $b=c=1$. The reversed symbol
484: is $H^{b^2 - , ab^2c +}_0 $. Left multiplication by $a$ and right multiplication by $c$ gives
485: $H^{ab^2c - , a^2b^2c^2 +}_0 = H^{ab^2c - , b^2 +}_0$
486: (since $a^2b^2c^2=b^2$), which is equivalent back to $H^{b^2 + , ab^2c -}_0$ by permuting the
487: superscripts. And so the $H^{b^2 + , ab^2c -}_0$ stratum is one sided.
488:
489: On the other hand, for $M = S^1 \times \R^2$, if $x$ denotes a generator of
490: $\pi_1(M)$ and say $k < r$ then $H^{x^k+,x^r-}_0 \in \C'_1$ is not equivalent to the reversed
491: symbol $H^{x^k-,x^r+}_0$ since the property of having the $+$ sign attached to the smaller
492: power of $x$ will be preserved under any left or right multiplication by element $x^s \in \pi_1(M)$.
493:
494:
495:
496:
497: \section{Finite order invariants}\label{fo}
498:
499:
500: Given an $n$-co-orientation $\tc$ for $i \in I_n$ and a subset $A\su \{p_1,\dots,p_n\}$,
501: we define $i_{\tc,A} \in I_0$ to be the immersion obtained from $i$ by resolving all CEs
502: of $i$ at points of $A$ into the
503: positive side with respect to $\tc$,
504: and all CEs not in $A$ into the negative side.
505: Now let $\G$ be any Abelian group and let $f:I_0\to\G$ be an invariant, i.e. a function which is
506: constant on each connected component of $I_0$.
507: Given $i\in I_n$ and an $n$-co-orientation $\tc$ for $i$,
508: $f^\tc(i)$ is defined as follows:
509: $$f^\tc(i)=\sum_{ A \su \{p_1,\dots,p_n\} } (-1)^{n-|A|} f(i_{\tc,A})$$
510: where $|A|$ is the number of elements in $A$.
511: If $\tc,\tc'$ are two $n$-co-orientations for the same immersion $i$ then
512: $f^\tc(i)=\pm f^{\tc'}(i)$ and so having $f^\tc(i)=0$ is independent of the $n$-co-orientation $\tc$.
513: An invariant $f:I_0\to\G$ is called \emph{of finite order} if
514: there is an $n$ such that $f^\tc(i)=0$ for all $i\in I_{n+1}$.
515: The minimal such $n$ is called the \emph{order} of $f$.
516: The group of all invariants on $I_0$ of order at most $n$ is denoted $V_n = V_n(\G)$.
517:
518: Let $f \in V_n$. If $i,j \in I_n$ and $\tc,\tc'$ are $n$-co-orientations for $i,j$ respectively, such that
519: $i,\tc$ and $j,\tc'$ are CAB equivalent, then $f^{\tc}(i) = f^{\tc'}(j)$, the argument being the same
520: as in \ct{o} Proposition 3.8. And so we have by Lemma \pr{srj} and Proposition \pr{cab}, that
521: each $f \in V_n$ induces a well defined function $\mu_n(f): \C_n \to \G$.
522: The map $f\mapsto \mu_n(f)$ induces an injection $\mu_n:V_n / V_{n-1} \to \C_n^*$
523: where $\C_n^*$ is the group of all functions from $\C_n$ to $\G$.
524:
525: If $i \in I_n$ and $\tc,\tc'$ are two $n$-co-orientations for $i$ which differ at precisely
526: one CE, then for any $f \in V_n$, $f^\tc(i) = -f^{\tc'}(i)$. If the CE where we have reversed the co-orientation
527: is of type $H$ or $Q$, and located at $p \in M$, then we have already noticed the relation between
528: $C_p(i,\tc)$ and $C_p(i,\tc')$, and so we now get an equation that must be satisfied by any $g \in \C_n^*$
529: in order for it to lie in the image of $\mu_n$, namely, if $Z_2,\dots,Z_n$ are any $n-1$ symbols
530: then for any symbol $H^{g_1\ep_1,g_2\ep_2}_d$, $g$ must satisfy
531: $g(H^{g_1\ep_1,g_2\ep_2}_d,Z_2,\dots,Z_n) = - g(H^{g_1\hat{\ep}_1,g_2\hat{\ep}_2}_d,Z_2,\dots,Z_n)$
532: where $\hat{\ep}_j$ denotes the sign opposite to $\ep_j$.
533: We will write such an equation in the short form
534: $H^{g_1\ep_1,g_2\ep_2}_d = -H^{g_1\hat{\ep}_1,g_2\hat{\ep}_2}_d$.
535: For $Q$ configuration we similarly get
536: $Q^{g_1\ep_1,g_2\ep_2,g_3\ep_3,g_4\ep_4}_d = -Q^{g_2\hat{\ep}_2,g_1\hat{\ep}_1,g_3\hat{\ep}_3,g_4\hat{\ep}_4}_d$
537: (note the transposition $1 \leftrightarrow 2$).
538: Also note that in the above two equations, the operation which produces the $n$-tuple
539: of symbols on the right from the
540: $n$-tuple of symbols on the left, is indeed well defined on the classes in $\C_n$.
541:
542:
543: %We will find the image of $\mu_1$, by this we will be classifying all order one invariants.
544:
545: Let $i \in \A$ be an immersion with a self intersection of local
546: codim 2 at $p$ and
547: $n-1$ additional self-intersections of local codim 1 (i.e. CEs) at $p_1,\dots,p_{n-1}$.
548: We look at a 2-parameter family of immersions which moves $F$ only in a neighborhood
549: of $p$, such that the immersion $i$ corresponds to
550: parameters $(0,0)$ and such that this 2-parameter family is transverse to the local
551: codim 2 stratum at $i$.
552: In this 2-parameter family of immersions we look at a loop which
553: encircles the point of
554: intersection with the codim 2 strata,
555: i.e. a circle around the origin in the parameter plane.
556: This circle crosses the local codim 1 strata some $r$ times.
557: Between each two intersections we have an immersion in $I_{n-1}$
558: with the same $n-1$ CEs, at $p_1,\dots,p_{n-1}$.
559: At each intersection with the local codim 1 strata, an $n$th CE is added,
560: obtaining an immersion in $I_n$. Let $i_1,\dots,i_r$ be the $r$ immersions in $I_n$ so
561: obtained. For each $1 \leq k \leq r$, choose a proper $n$-co-orientation $\tc_k$ for
562: $i_k$, such that the co-orientation chosen for each $p_j$, $1 \leq j \leq n-1$ is the same in all
563: $\tc_1,\dots,\tc_r$. Let $e_k$, $k=1,\dots,r$ be $1$ or $-1$ according to whether we are passing
564: the $n$th CE of $i_k$ in the direction of its co-orientation determined by $\tc_k$, or in the opposite direction,
565: respectively.
566: For an invariant $f$, it is easy to show that the following equation holds:
567: $\sum_{k=1}^r e_k f^{\tc_k}(i_k) = 0$.
568: Looking at $\mu_n:V_n / V_{n-1} \to \C_n^*$ we thus obtain additional equations that must be satisfied by
569: a function in $\C_n^*$ in order for it to lie in the image of $\mu_n$.
570: We will now find all equations on $\C_n^*$ obtained in this way.
571: As above, the equations will be written in short form as equations on the symbols.
572: We may assume (by moving the codim 2 singular point $p$ if necessary), that there is a neighborhood
573: $U$ of $p$ such that the labeling by elements of $\p$ is locally constant in $r^{-1}(U)$,
574: and such that all motion involved in our 2-parameter family of immersions takes place inside $U$.
575: And so we will be able to follow the labelings during our loop of immersions.
576:
577: The local codim 2 strata may be divided into six types which we name after the types of CEs
578: appearing in a 2-parameter family of immersions, transverse to the given stratum:
579: $EH$, $TT$, $ET$, $HT$, $TQ$, $QQ$. In the notation of \ct{hk} they are
580: respectively: $A_0^2|A_2$, $A_0^3|A_2$, $(A_0^2|A_1^+)(A_0)$,
581: $(A_0^2|A_1^-)(A_0)$, $(A_0^3|A_1)(A_0)$, $A_0^5$.
582: Formula and sketch for local model for such strata,
583: the bifurcation diagrams, and the equations obtained, are as follows.
584: The symbol $\cdots$ represents a string of $g_j\ep_j$s of the appropriate length
585: (the same string for all appearances of $\cdots$ within the same diagram or equation).
586:
587:
588:
589:
590: \begin{figure}[h]
591: \scalebox{0.6}{\includegraphics{SEH.eps}}
592: \caption{$EH$ configuration}\label{feh}
593: \end{figure}
594:
595: $EH$: \ \ $z=0$, \ \ $z=y^2 + x^3+\lm_1 x + \lm_2$.
596: \begin{equation}\label{eeh}
597: 0 = E^{\cdots}_d - H^{\cdots}_d
598: \end{equation}
599:
600:
601:
602:
603: \begin{figure}[h]
604: \scalebox{0.6}{\includegraphics{STT.eps}}
605: \caption{$TT$ configuration}\label{ftt}
606: \end{figure}
607:
608:
609: $TT$: \ \ $z=0$, \ \ $y=0$, \ \ $z=y+x^3+\lm_1 x + \lm_2$.
610: \begin{equation}\label{ett}
611: 0 = T^{g_1\ep_1,g_2\ep_2,g_3\ep_3}_d - T^{g_1\hat{\ep}_1,g_2\hat{\ep}_2,g_3\hat{\ep}_3}_d
612: \end{equation}
613:
614:
615:
616: \begin{figure}[h]
617: \scalebox{0.6}{\includegraphics{SET.eps}}
618: \caption{$ET$ configuration}\label{fet}
619: \end{figure}
620:
621: $ET$: \ \ $z=0$, \ \ $x=0$, \ \ $z=(x-\lm_1)^2 + y^2 + \lm_2$.
622: \begin{equation}\label{eet}
623: 0 = T^{\cdots g-}_d - T^{\cdots g+}_d - E^{\cdots}_{d-g} + E^{\cdots}_d
624: \end{equation}
625:
626:
627: \begin{figure}[h]
628: \scalebox{0.6}{\includegraphics{SHT.eps}}
629: \caption{$HT$ configuration}\label{fht}
630: \end{figure}
631:
632:
633: $HT$: \ \ $z=0$, \ \ $x=0$, \ \ $z=(x-\lm_1)^2 - y^2 + \lm_2$.
634: \begin{equation}\label{eht}
635: 0 = -T^{\cdots g+}_d + T^{\cdots g-}_d - H^{\cdots}_{d-g} + H^{\cdots}_d
636: \end{equation}
637:
638:
639:
640:
641: \begin{figure}[h]
642: \scalebox{0.6}{\includegraphics{STQ.eps}}
643: \caption{$TQ$ configuration}\label{ftq}
644: \end{figure}
645:
646:
647: $TQ$: \ \ $z=0$, \ \ $y=0$, \ \ $x=0$, \ \ $z=y+(x-\lm_1)^2 + \lm_2$.
648: \begin{equation}\label{etq}
649: 0 = Q^{\cdots g-}_d - Q^{\cdots g+}_d - T^{\cdots}_{d-g} + T^{\cdots}_d
650: \end{equation}
651:
652:
653:
654:
655: $QQ$: This configuration is a quintuple point, i.e. five sheets passing through a
656: point, any three of which are in general position.
657: We model this by five planes passing through a point in $\E$.
658: A 2-parameter family may be constructed by fixing three of the planes,
659: and allowing the other two planes to move parallel to themselves.
660: A loop in the shape of a rectangle around the origin of the parameter plane
661: may be constructed as follows: The three fixed planes $P_1,P_2,P_3$ intersect in a triple point
662: $x$. Start with the plane $P_4$ very close to $x$, and $P_5$ somewhat further away. For the first
663: edge of the rectangle, move $P_4$ across
664: $x$ passing a first quadruple point. In this process a little simplex $S^-$ has vanished and a new little
665: simplex $S^+$ has been created from the same four planes $P_1,\dots,P_4$.
666: For the second edge of the rectangle,
667: move $P_5$ across the whole of $S^+$, passing four quadruple points on the way,
668: when crossing the four triple points of $S^+$.
669: For the third edge move $P_4$ back to its original place, thus crossing $x$ again, in the opposite direction,
670: recreating $S^-$ in place of $S^+$, this being the sixth quadruple point. Finally move
671: $P_5$ back to place, across $S^-$, adding four
672: more quadruple points, and so completing a total of ten quadruple points.
673: We co-orient each of these ten quadruple points according to the direction we are passing it along this loop,
674: and so each of the ten terms in the equation that we are producing, will have a $+$ sign in front of it.
675:
676: We will call the first and third edge of the rectangle described above "short edges" (since they pass only
677: one quadruple point) and the second and fourth edges "long edges".
678: We will first find the relation between the symbols of two consecutive quadruple points along a
679: long edge, and then see that the same relation holds when passing to a short edge, and back.
680: So let $p,q$ be two vertices of the little simplex, which $P_5$ passes consecutively.
681: Let $e$ be the edge of the simplex connecting $p,q$. By affine transformation of $\E$ we may assume
682: that $p = (0,0,-1), q=(0,0,1)$, that the two planes intersecting to create the edge $e$ are the planes
683: $\{ x=0 \}$ and $\{ y=0 \}$,
684: and that the plane $P_5$ which is in motion, crossing $p$ and then $q$, is the plane
685: $\{ z = t \}$ where $t$ increases in the range $-1-r < t < 1+r$ for some small $r$.
686: For concreteness we look at the time $t=0$, and so $P_5 = \{ z=0 \}$.
687: As to the remaining two planes, one must pass $p$,
688: and the other must pass $q$.
689: They cannot be vertical planes (since we already have two vertical planes),
690: and so they are given by $\{ z=-1+ax+by \}$ and $\{ z=1+cx+dy \}$. The coefficients
691: $a,b$ must both be nonzero since if say $b=0$ then we get three planes $\{ z=-1+ax \} $, $\{ z=0 \}$,
692: $\{ x=0 \}$ which are not in general position (when parallelly moved to have a common point). Similarly $c,d$
693: are nonzero.
694: By rotating the configuration around
695: the $z$ axis we may assume $a,b>0$. We claim that we must then get also
696: $c,d >0$. Indeed, in order that $P_5$ will pass $p$ and $q$ consecutively, it must be that the other
697: two vertices of the simplex lie outside the region $-1 \leq z \leq 1$ of $\E$.
698: These two points are $s_1 = \{ z=-1+ax+by \} \cap \{ z=1+cx+dy \} \cap \{ x=0 \}$
699: and $s_2 = \{ z=-1+ax+by \} \cap \{ z=1+cx+dy \} \cap \{ y=0 \}$. The $z$ component of $s_1$ is
700: $b+d \over b-d$, and this is outside the interval $[-1,1]$ iff $d>0$.
701: (If $b=d$ then there is no intersection, which implies that the three planes are not in general position.)
702: In the same way $c>0$, using $s_2$.
703: And so the simplex $S^+_p$ created from the passage of $P_5$ across $p$ (bounded by the planes
704: $\{ x=0 \}, \{ y=0 \}, \{ z=0 \}, \{ z=-1+ax+by \}$), lies in the octant $x \geq 0, y\geq 0, z \leq 0$,
705: whereas the simplex $S^-_q$ that is about to vanish when $P_5$ arrives at $q$ (bounded by the planes
706: $\{ x=0 \}, \{ y=0 \}, \{ z=0 \}, \{ z=1+cx+dy \}$), lies in the octant $x \leq 0, y\leq 0, z \geq 0$
707: which is the exact opposite octant. And so the sign $\ep_j$ with which each of the common planes
708: $\{ x=0 \}, \{ y=0 \}, \{ z=0 \}$ appears in $S^+_p$ and $S^-_q$ is the opposite sign. But since we
709: have co-oriented all quadruple points according to our direction of motion, the simplex $S^-_q$
710: is not the one with which we determine the signs for the symbol $Z_q$ of the quadruple point at $q$, but
711: rather the simplex $S^+_q$ which will be created \emph{after} we cross $q$.
712: As we have already noticed, the signs of the faces for $S^+_q$ are all opposite to the corresponding
713: signs for $S^-_q$, and so finally the common planes with $S^+_p$ will have the same sign as in $S^+_p$.
714:
715: We apply a similar two step argument to determine the relation between the orderings of the faces in the
716: two symbols $Z_p,Z_q$. We note that if some ordering of the faces of $S^+_p$ is consistent with the orientation
717: of $S^+_p$ (restricted from $\E$) then the ordering for the faces of $S^-_q$ obtained by simply replacing the
718: plane $\{ z=-1+ax+by \}$ with the plane $\{ z=1+cx+dy \}$ is \emph{in}consistent with the orientation of
719: $S^-_q$. And so this same ordering of these same four planes, now considered as the faces of $S^+_q$,
720: \emph{is} consistent with the orientation of $S^+_q$.
721: And so for the superscripts of two consecutive symbols we finally get, that the $g_j\ep_j$ corresponding to
722: the plane which appears only in the first quadruple point and not the second (the plane $\{ z=-1+ax+by \}$
723: in our case) should be dropped, and exactly in its place should be written the $g_j\ep_j$ corresponding to
724: the new plane which participates in the second quadruple point but not in the first
725: (the plane $\{ z=1+cx+dy \}$ in our case). The three $g_j\ep_j$s corresponding to the three common planes
726: (the planes $\{ x=0 \}, \{ y=0 \}, \{ z=\pm 1 \}$ in our case) must remain in their place, and with
727: the signs $\ep_j$ unchanged.
728:
729: We now show that when passing from the last quadruple point of a long edge to the quadruple point of the
730: next short edge, or from that to the first quadruple point of the next long edge, then the same relation holds.
731: Indeed, looking say at the first and
732: second quadruple point in our rectangle described above, let $y$ be the triple point
733: of $S^+$ which $P_5$ passes first along the long edge, for the occurrence of the second quadruple point.
734: Let $P$ be one of the planes $P_1,\dots,P_4$ such that if $P$ is moved in the direction such that
735: $S^+$ increases in size, then the point $y$ will move toward $P_5$ and eventually cross it. So we can
736: start from the same initial position as before, but instead of moving $P_4$ and then $P_5$, we move only
737: plane $P$, such that $S^-$ will vanish, passing the first quadruple point, $S^+$ is created and
738: increases in size until $y$ crosses $P_5$ thus passing the second quadruple point.
739: So we have obtained the same two quadruple points, where now four of the planes are fixed and
740: only the plane $P$ is in motion. But this is precisely the motion analyzed above to obtain the
741: relation between two consecutive quadruple points along a long edge.
742:
743: Finally it remains to determine the degrees $d \in \gr$ of the ten symbols.
744: We do this again by analyzing the relation between the degrees of two consecutive
745: quadruple points. Following the intersection point between $\{x=0\}, \{y=0\}, \{z=t\}$ as it continuously moves
746: up from $p$ to $q$, we see that there is no change in the degree due to
747: these three planes.
748: (Recall that the labeling is locally constant above the region of interest.)
749: There may, on the other hand, be a change due to the planes $P_i = \{ z=-1+ax+by \}$
750: and $P_j = \{ z=1+cx+dy \}$. If the preferred side of $P_i$ is facing upward, which happens iff the corresponding
751: sign $\ep_i$ appearing in the symbol $Z_p$ for the first quadruple point is $-$,
752: then when computing the degree $d_p$ for the first quadruple point, $P_i$ is slight pushed upward
753: so as to pass above the point $p$. But as $P_5$ moves upward to $q$, $P_i$ does not participate in the second
754: quadruple point, and remains below the point $q$ which is where the degree is now computed. So in this case there
755: is a change of $-g_i$ to the degree due to the plane $P_i$. If on the other hand the plane $P_i$ is facing
756: downward, which happens iff the sign $\ep_i$ appearing in $Z_p$ is $+$, then the plane $P_i$ is
757: pushed \emph{downward} when $d_p$ is computed, and again it will remain below the point $q$ when the degree
758: of the second quadruple is computed, and so there will be no change in the degree due to $P_i$. The two cases
759: may be written in one formula as follows. For sign $\ep \in \{ +,- \}$, define
760: $$ | \ep | =
761: \begin{cases}
762: 1 & \text{if} \ \ep = + \\
763: 0 & \text{if} \ \ep = -
764: \end{cases}
765: $$
766: then the change in degree between the first and second quadruple point, due to the plane $P_i$
767: is $-|\hat{\ep_i}|g_i$. A similar analysis as to the effect of the plane $P_j$ on the degree will
768: give $+|\ep_j|g_j$. Together we get that if $P_j$ is the plane missing from the first quadruple point, and $P_i$
769: appears there with sign $\ep_i$, and $P_i$ is missing from the second quadruple point, and $P_j$ appears there
770: with sign $\ep_j$, then the change in the degree between the first and second quadruple point
771: is $+|\ep_j|g_j-|\hat{\ep_i}|g_i$.
772:
773: Looking at the circle (or rectangle) of ten quadruple points, we see that the same 4-tuple of planes participates
774: in a quadruple point twice, and with opposite co-orientation, and so the ordering of the corresponding $g_i\ep_i$ at
775: the two times differ by an odd permutation.
776: They occur at the precise opposite timing along the circle, i.e. five places apart, and so the ordering of
777: quadruple points along the circle is that all five possible 4-tuples appear one after the other, and then they
778: appear again in the same order, to complete the cycle of ten quadruple points. So given a quadruple point, either
779: it or the matching quadruple point on the opposite side of the circle, have the property, that the ordering of
780: superscripts may be chosen as
781: $g_1\ep_1,\dots,g_4\ep_4$ so that in the next quadruple point the plane corresponding to $g_1$ will be missing, in
782: the following quadruple point the plane corresponding to $g_2$ will be missing,
783: and so on. Finally, since the two matching quadruple points (i.e. those with the same 4-tuple of
784: planes) occur with opposite co-orientation, the sign attached to each $g_i$ will be opposite
785: (i.e. $g_i\ep_i$ is replaced by $g_i\hat{\ep}_i$).
786: Since we have seen that if a plane is common to two consecutive quadruple points, it appears
787: in their superscripts with the same sign, and since a given plane will be missing exactly once between
788: its appearance in a given quadruple point, and the occurrence of the matching quadruple point,
789: it follows that whenever a plane is missing from a quadruple point, then it appears
790: in the following quadruple point with opposite sign than it does in the previous one.
791:
792: To write our final formula, let us choose the first quadruple point and ordering $g_1\ep_1,\dots,g_4\ep_4$
793: as mentioned above, and we may write its degree as $d-|\ep_5|g_5$ where $g_5$ corresponds to the plane missing
794: from the first quadruple point, and $\ep_5$ is the sign with which it will appear in the second quadruple point.
795: The first symbol is thus
796: $Q^{g_1\ep_1,g_2\ep_2,g_3\ep_3,g_4\ep_4}_{d-|\ep_5|g_5}$, and the combination of all analysis above produces the
797: following equation:
798:
799:
800: \begin{multline}\label{eqq}
801: 0 = Q^{g_1\ep_1,g_2\ep_2,g_3\ep_3,g_4\ep_4}_{d-|\ep_5|g_5}+
802: Q^{g_5\ep_5,g_2\ep_2,g_3\ep_3,g_4\ep_4}_{d-|\hat{\ep}_1|g_1}+
803: Q^{g_5\ep_5,g_1\hat{\ep}_1,g_3\ep_3,g_4\ep_4}_{d-|\hat{\ep}_2|g_2}+
804: Q^{g_5\ep_5,g_1\hat{\ep}_1,g_2\hat{\ep}_2,g_4\ep_4}_{d-|\hat{\ep}_3|g_3}+
805: Q^{g_5\ep_5,g_1\hat{\ep}_1,g_2\hat{\ep}_2,g_3\hat{\ep}_3}_{d-|\hat{\ep}_4|g_4} \\
806: + Q^{g_4\hat{\ep}_4,g_1\hat{\ep}_1,g_2\hat{\ep}_2,g_3\hat{\ep}_3}_{d-|\hat{\ep}_5|g_5} +
807: Q^{g_4\hat{\ep}_4,g_5\hat{\ep}_5,g_2\hat{\ep}_2,g_3\hat{\ep}_3}_{d-|\ep_1|g_1}+
808: Q^{g_4\hat{\ep}_4,g_5\hat{\ep}_5,g_1\ep_1,g_3\hat{\ep}_3}_{d-|\ep_2|g_2}+
809: Q^{g_4\hat{\ep}_4,g_5\hat{\ep}_5,g_1\ep_1,g_2\ep_2}_{d-|\ep_3|g_3}+
810: Q^{g_3\ep_3,g_5\hat{\ep}_5,g_1\ep_1,g_2\ep_2}_{d-|\ep_4|g_4}
811: \end{multline}
812:
813: Now let $\Delta_n = \Delta_n(\G) \su \C_n^*$ be the subgroup consisting of all
814: functions $g \in \C_n^*$ which satisfy the above equations (1) -- (6) together with our
815: first two equations $H^{g_1\ep_1,g_2\ep_2}_d = -H^{g_1\hat{\ep}_1,g_2\hat{\ep}_2}_d$ and
816: $Q^{g_1\ep_1,g_2\ep_2,g_3\ep_3,g_4\ep_4}_d = -Q^{g_2\hat{\ep}_2,g_1\hat{\ep}_1,g_3\hat{\ep}_3,g_4\hat{\ep}_4}_d$.
817: Then we have obtained that the image of the injection $\mu_n : V_n / V_{n-1} \to \C_n^*$
818: is contained in $\Delta_n$. Finding the precise image of $\mu_n$ for all $n$ (as has been done for
819: the case $M=\E$ in \ct{o},\ct{h}), would give a full classification of all finite order invariants. In the next section
820: we will show that the image of $\mu_1$ is all $\Delta_1$, by this classifying all order one invariants.
821:
822:
823:
824:
825: \section{Order one invariants}\label{oo}
826:
827: We define a ``universal'' Abelian group $\G_U$, by the Abelian group presentation
828: which takes as generators the elements of $\C_1$, and as relations, the set of all
829: equations used above to define $\Delta_n$ (and in particular $\Delta_1$).
830: Note that for defining $\Delta_1$, an expression such as
831: $H^{g_1\ep_1,g_2\ep_2}_d = -H^{g_1\hat{\ep}_1,g_2\hat{\ep}_2}_d$
832: was merely a short form for writing
833: $g(H^{g_1\ep_1,g_2\ep_2}_d) = -g(H^{g_1\hat{\ep}_1,g_2\hat{\ep}_2}_d)$,
834: this being a condition on an element $g \in \C_1^*$
835: for being included in the subgroup $\Delta_1$, whereas, in the definition of $\G_U$,
836: $H^{g_1\ep_1,g_2\ep_2}_d = -H^{g_1\hat{\ep}_1,g_2\hat{\ep}_2}_d$
837: is an actual relation in the presentation of $\G_U$ by generators and relations.
838: We define the universal element $g^U \in \Delta_1(\G_U)$, to be the function that assigns to each
839: element in $\C_1$, the generator corresponding to it in $\G_U$. By definition of $\G_U$,
840: indeed $g^U \in \Delta_1(\G_U)$. We will now establish the existence of an order 1 invariant
841: $f^U : I_0 \to \G_U$ satisfying $\mu_1(f^U)=g^U$.
842: This will prove our desired result, that for any Abelian group $\G$, $\mu_1 : V_1(\G) \to \Delta_1(\G)$
843: is surjective. Indeed, given $g \in \Delta_1(\G)$, there is a (unique) homomorphism
844: $\varphi : \G_U \to \G$ such that $g = \varphi \circ g^U$, and we get
845: $\mu_1 ( \varphi \circ f^U) = g$.
846:
847:
848:
849: \begin{lemma}\label{B}
850: There exists $i_0 \in \A$ whose image is contained in a ball $B \su M$.
851: \end{lemma}
852:
853: \begin{pf}
854: Let $i \in \A$ be some immersion and $H_t:F \to M$ a null-homotopy from $i$
855: to a constant map $k:F \to M$ with $k(F)=p \in B$. Using the parallelizibility
856: of $M$ we extend $H$ to a homotopy of bundle monomorphisms, from $di$ to a bundle monomorphism
857: $b$ which covers $k$. By the Smale-Hirsch Theorem
858: applied to $Imm(F,B)$, there is an immersion $i_0:F \to B$ with $di_0$ homotopic to $b$ as bundle
859: monomorphisms. Now by the Smale-Hirsch Theorem applied to $Imm(F,M)$, $i$ is regularly homotopic
860: to $i_0$, and so $i_0 \in \A$.
861: \end{pf}
862:
863: Fix an immersion $i_0 \in I_0$ with image contained in a small ball $B \su M$, as
864: provided by Lemma \pr{B}.
865: For any $i \in I_0$ take a generic regular homotopy $J_t$ from $i_0$ to $i$,
866: and to each CE which occurs along $J_t$ choose a proper co-orientation.
867: Let $f^U(i)$ be the sum of the elements in $\G^U$ corresponding to the
868: co-oriented CEs occurring along $J_t$, each taken with a sign $\pm$ according to whether
869: $J_t$ passes it in the direction of its co-orientation. An opposite choice of co-orientation
870: for a CE along $J_t$ (relevant for types $H,Q$), will produce the negative element in $\G_U$, but
871: this element will appear with opposite sign in the sum and so this sum is well defined, for
872: given $J_t$.
873: We would like to show that in fact $f^U$ is
874: independent of the regular homotopy $J_t$, and so $f^U$ is a well defined invariant.
875: This is equivalent
876: to showing that this sum is 0 along any closed $J_t$, i.e. $J_t$ such that $J_0=J_1=i_0$.
877: Once we know $f^U$ is well defined, it is clear that it is of order 1,
878: and $\mu_1(f^U)=g^U$.
879:
880: The fact that the sum of values is 0 along any null-homotopic loop follows from
881: the definition of $\Delta_1$ and $\G_U$ as demonstrated in \ct{o},
882: and so the sum along loops induces a well defined homomorphism $\phi : \pi_1(\A,i_0) \to \G_U$.
883: We must show $\phi=0$.
884:
885: Fix a disc $D \su F$, and let $S: \A \to Imm(D,M)$ be given by restriction,
886: $S(i)=i|_D$. Then $S$ is a fibration with fiber $\A_D = \{ i \in \A \ : \ i|_D = i_0|_D \}$.
887: So we get an exact sequence
888: $$\pi_1(\A_D,i_0) \stackrel{inc_*}{\longrightarrow} \pi_1(\A , i_0)
889: \stackrel{S_*}{\longrightarrow} \pi_1(Imm(D , M ) , i_0|_D ).$$
890: Let $K \su \pi_1(\A , i_0)$ be the subgroup consisting of loops obtained by composing $i_0 : F \to B$
891: with a motion of $B$ in $M$ through embeddings, beginning and ending with the inclusion of $B$ in $M$.
892: It is clear that $S_*$ maps $K$ onto $\pi_1(Imm(D , M ) , i_0|_D )$, and so the elements of $K$ and
893: those coming from $\pi_1(\A_D,i_0)$ generate $\pi_1(\A , i_0)$. Now $\phi$ is 0 on elements of $K$, since no
894: CEs occur along such loops, and so it remains to show that $\phi$ is 0 on
895: $\pi_1(\A_D,i_0)$. It is shown in the proof of \ct{q} Theorem 3.4, that $\pi_1(\A_D,i_0)$ is generated by
896: a loop which only moves a small disc $U \su F$ which is disjoint from $D$,
897: and $U$ moves only within a small ball $B'$.
898: In our case we may take $B' \su B$, and so we finally need only to check the value of $\phi$
899: on loops moving $F$ in $B$. If we take our labeling to be locally constant above $B$,
900: then we can have all lifts used to define our symbols to be contained in the same lift of $B$,
901: constantly labeled by some fixed element $h \in \p$.
902: And so all symbols will be of the form $R^{h\ep_1,h\ep_2,\dots}_{mh}$.
903: By a permutation of the superscript, which is even in case $R$ is $Q$, we may assume that all
904: $+$ signs appear first, and so the superscript is characterized only by the \emph{number} of $+$ signs,
905: which is precisely the way the superscript of a symbol is defined in \ct{o} for immersions
906: $F \to B = \E$.
907: The subscript is characterized by the coefficient $m$ of $mh$ which is also simply the degree
908: as defined in \ct{o} for immersions $F \to B$. And so $\phi$ being 0 on all loops moving
909: $F$ in $B$, follows from the fact that $f^U$ appearing in \ct{o} for the case $M=\E$, is well defined.
910: Indeed, for $f^U$ of \ct{o}, the sum of symbols along a loop of immersions
911: in $B$, is 0 in the group $G_U$ of \ct{o}.
912: We have noticed that our symbols of the form $R^{h\ep_1,h\ep_2,\dots}_{mh}$
913: correspond to the symbols of \ct{o},
914: but furthermore, any equation on the symbols appearing in \ct{o},
915: has a corresponding equation on our symbols of the form
916: $R^{h\ep_1,h\ep_2,\dots}_{mh}$,
917: by considering the same codim 2 configuration in $B$.
918: And so the sum in our present $G_U$ must also be 0.
919: This completes the proof of:
920:
921: \begin{thm}
922: Let $F$ be a closed orientable surface, $M$ an aspherical orientable 3-manifold,
923: and $\A$ a regular homotopy class of null-homotopic immersions of $F$ into $M$.
924: Then for any Abelian group $\G$, the injection $\mu_1 : V_1 / V_0 \to \Delta_1$ is surjective.
925: \end{thm}
926:
927: We conclude with the following remark. Our classification is in terms of the group $\Delta_1(\G)$
928: which as we have seen is the same as the group $Hom(\G_U , \G)$. In \ct{o} where $M=\E$ and so
929: $\pi_1(M)=0$, the structure of $\G_U$ is completely understood and accordingly the group
930: of order 1 invariants. In the general case presented here, the structure of
931: $\G_U$ depends on the structure of $\pi_1(M)$, most notably through the equivalence relation which
932: produces $\C_1$ from $\C'_1$. The complexity of this dependence may be seen in the
933: discussion following Proposition \pr{cab}, where the question is whether a given symbol is equivalent
934: to its reversed symbol. And so a closed explicit classification as in \ct{o}
935: (not to mention explicit formulae for the invariants, as in \ct{f}), will require extensive
936: analysis of $\pi_1(M)$ for any given $M$.
937:
938:
939:
940:
941:
942:
943:
944:
945:
946: \begin{thebibliography}{StoPC}
947:
948: \bibitem[HK]{hk}
949: C. A. Hobbs, N. P. Kirk: ``On the Classification and Bifurcation of Multigerms of Maps
950: from Surfaces to 3-Space.'' - \emph{Mathematica Scandinavica}
951: 89 (2001) 57--96.
952:
953: \bibitem[N1]{q}
954: T. Nowik:
955: ``Quadruple points of regular homotopies of surfaces in 3-manifolds.''
956: \emph{Topology} 39 (2000) 1069-1088.
957:
958: \bibitem[N2]{a}
959: T. Nowik:
960: ``Automorphisms and Embeddings of Surfaces and Quadruple Points of Regular Homotopies.''
961: \emph{Journal of Differential Geometry} 58 (2001) 421-455
962:
963: \bibitem[N3]{o}
964: T. Nowik:
965: ``Order one invariants of immersions of surfaces into 3-space.''
966: \emph{Mathematische Annalen} 328 (2004) 261-283.
967:
968: \bibitem[N4]{h}
969: T. Nowik:
970: ``Higher order invariants of immersions of surfaces into 3-space.''
971: \emph{Pacific Journal of Mathematics} - to appear.
972:
973:
974: \bibitem[N5]{f}
975: T. Nowik:
976: ``Formulae for Order One Invariants of Immersions of Surfaces.''
977: \emph{Advances in Mathematics} - to appear.
978:
979: \bibitem[N6]{n}
980: T. Nowik:
981: ``Immersions of Non-orientable Surfaces.'' - Preprint.
982:
983: \end{thebibliography}
984:
985: %%%%%%%%%%%%%%
986: \end{document}
987: %%%%%%%%%%%%%%
988:
989:
990:
991:
992:
993: %\begin{figure}[t]
994: %\scalebox{0.7}{\includegraphics{prl1.eps}}
995: %\caption{}\label{f1}
996: %\end{figure}
997:
998:
999:
1000:
1001: $C_{p'_k}(i)=C_{p_k}(j)$, $k=1,\dots,n$. This means in
1002: particular, that if $B'_1,\dots,B'_n$ and $B_1,\dots,B_n$ are neighborhoods of the $p'_k$s
1003: and $p_k$s respectively, then for each $k$ there is an orientation preserving diffeomorphism
1004: from $B'_k$ to $B_k$ which takes each sheet of $i(F)\cap B'_k$ orientation
1005: preservingly onto the corresponding sheet of $j(F)\cap B_k$.
1006: These diffeomorphisms may all be realized by one ambient isotopy $U_t:\E\to\E$. There is
1007: then an isotopy $V_t:F\to F$ such that the
1008: final immersion $i'$ of the regular homotopy $U_t\circ i \circ V_t$
1009: satisfies that $i'$ and $j$ have the
1010: same $n$ CE points $p_1,\dots,p_n\in\E$, ${i'}^{-1}(\bigcup_k B_k) = j^{-1}(\bigcup_k B_k)$
1011: which we name $U$,
1012: and $i'|_U = j|_U$.
1013: Also $d_{p_k}(i') = d_{p_k}(j)$ for $k=1,\dots,n$.
1014: Now $U$ is a union of some discs $D_1,\dots,D_r$. We construct the following
1015: handle decomposition of $F$. $D_1,\dots, D_r$ will be the 0-handles. If $g$ is the
1016: genus of $F$ we will have 1-handles $h_1,\dots,h_{2g+r-1}$
1017: as follows: $h_1,\dots,h_{2g}$ will
1018: each have both ends glued to $D_1$ such that
1019: $F-(D_1 \cup h_1 \cup \cdots \cup h_{2g})$ will
1020: be a disc containing $D_2,\dots,D_r$.
1021: Then for $k=1,\dots r-1$, $h_{2g+k}$ will have one end glued to $D_k$ and the other
1022: to $D_{k+1}$.
1023: The complement of all
1024: 0- and 1-handles is again one disc, which will be the unique 2-handle.
1025: We will now construct a
1026:
1027:
1028:
1029:
1030:
1031:
1032:
1033:
1034:
1035:
1036:
1037:
1038:
1039:
1040: