1: %%%%%%%%%%%%%% Paper 041
2: \documentstyle[twoside,fleqn,gracos]{article}
3: \jheads{V.V. Kassandrov}
4: {Biquaternion Electrodynamics and Weyl-Cartan Geometry of Space-Time}
5:
6: %\font\B=msbm9
7: %\font\Bb=msbm10 scaled 1200
8: %\font\Bbb=msbm10
9: %
10: \def\B{\cal}
11: \def\Bb{\cal}
12: \def\Bbb{\cal}
13:
14: \begin{document}
15: %\English
16: \twocolumn[
17: \Arthead{3}{216}{222}
18:
19: \Title{BIQUATERNION ELECTRODYNAMICS\\[5pt]
20: AND WEYL-CARTAN GEOMETRY OF SPACE-TIME}
21:
22: \Author{V.V.Kassandrov}
23: {Russian People's Friendship University, Department of General
24: Physics,\\ 3 Ordjonikidze Str., Moscow 117302, Russia}
25:
26: \Rec{5 July 1995}
27:
28: \Abstract
29: {The generalized Cauchy-Riemann equations (GCRE) in biquaternion
30: algebra appear to be Lorentz-invariant. The Laplace equation is in this case
31: replaced by a nonlinear {\B C}\,-eikonal equation. GCRE contain a 2-spinor
32: and a {\B C}\,-gauge structures, and their integrability conditions take the
33: form of Maxwell and Yang-Mills equations. For the value of electric
34: charge from GCRE only the quantization rule follows, as well as the
35: treatment of Coulomb law as a stereographic map. The equivalent
36: geometrodynamics in a Weyl-Cartan affine space and the conjecture of
37: a complex-quaternion structure of space-time are discussed.}
38:
39: ] %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
40:
41: \section{Introduction} %Sec1
42: %
43: In the frames of the {\it geometrodynamic\/} approach all
44: fundamental physical quantities and above all the equations of
45: physical dynamics should be of a purely geometric nature. The twistor
46: program, the Ka\-lu\-za-Kle\-in theories and string dynamics give
47: representative examples of this concept, perhaps the most general ones up
48: to now. In essence, any physical interaction may be regarded as a
49: manifestation of geometry (by using multi-dimensional spaces, fiber bundles,
50: etc.).
51:
52: However, the diversity of admissible geometries and their
53: invariants makes the ``kinematic'' part of this procedure (selection
54: of space and geometric identification of physical quantities), as
55: well as the dynamic one (choice of a Lagrangian) quite ambiguous.
56: Even for the electromagnetic (EM) field one has a lot of different
57: geometric interpretations (Weyl's conformal factor, bundle
58: connection, the Kaluza metric field, torsion [1] or nonholonomic [2]
59: structures of space-time (ST), etc.).
60:
61: Alternatively, within the {\it algebrodynamic\/} pa\-ra\-di\-gm [3] the ST is
62: regarded as a manifold supplied with a basic algebraic structure, the
63: structure of linear algebra in the simplest case. But it is well known
64: that the exceptional algebras --- algebras with division and positive
65: norm --- exist in the dimensions $d=4$ (Hamilton quaternions) and $d=8$
66: (Cayley octonions). So it would be natural to suppose that the
67: ST algebra (STA) [4] should be exceptional in its internal
68: mathematical properties. If it is the case, the group of automorphisms
69: (Aut) of STA would generate the ST geometry, for example, by
70: operating as an isometry group.
71:
72: Moreover, the STA structure can completely determine
73: physical dynamics as well. Indeed, if we consider the physical fields
74: as algebra-valued functions of an algebraic variable, the {\it generalized
75: Cauchy-Ri\-emann equations\/} (GCRE), i.e., the differentiability conditions
76: in the STA, become fundamental equations of field dynamics.
77: Wonderfully, the generally accepted physical equations (in
78: particular, the Maxwell or Yang-Mills equations) become a direct
79: consequences of GCRE, namely their integrability conditions (see below).
80:
81: From an epistemological point of view, the algebrodynamic (AD)
82: concept returns us to the ideas of Pythagoras, Hamilton and
83: Eddington on {\bf a crucial role of Numbers in the structure of the
84: Universe}. At the modern stage we deal with the primary structure of
85: multidimensional {\it ST arithmetics}, completely different from the
86: classical arithmetics of the Macroworld, or the world of reversible
87: processes and weakly interacting objects.
88:
89: {\bf A genuine ST arithmetics ought to be non-commutative and even non-%
90: associative!} Indeed, these properties are just algebraic equivalents
91: of causal and interactive structures of the physical World (ensuring the
92: dependence of ``out-state'' on the order and composition of reactions).
93: For such reasons the most suitable STA candidate is the {\it octonion
94: algebra}, the unique exceptional non-associative algebra. However, the
95: difficulties of ``intercourse'' with octonions are well-known (see,
96: nevertheless, [6\,7]).
97:
98: Meanwhile, the non-commutativity of algebraic structures
99: is closely connected with the non-linearity of the corresponding dynamic
100: equations (this is the case, in particular, for the Yang-Mills fields).
101: We will see later that the GCRE in non-commutative algebras also
102: possess a nonlinear structure and are therefore capable to describe both
103: quantum phenomena and physical field interactions.
104:
105: In this paper we choose for a STA the {\it algebra of biquaternions\/}
106: {\Bbb B}\,,
107: the extension of real Hamilton quaternions {\Bbb H} to the field of complex
108: numbers {\Bbb C}. The {\Bbb H}\, algebra is known to have
109: ${\rm Aut}\,(\mbox{\Bbb H})={\rm SO}\,(3)$ and is in
110: perfect correspondence with the structure of the 3-dimensional space.
111: {\bf We are unaware of a similar algebra for the case of Minkowski
112: 4-space!} For obvious reasons one often considers the Clifford-Dirac
113: algebra C(1,\,3) to be the STA [4,\,5]. However, a reduction from
114: the 16-dimensional total vector space of C(1,\,3) to a 4-dimensional
115: physical ST is a completely ``voluntaristic'' procedure; even the metric
116: signature of the basic generator space may be chosen in different ways [8].
117:
118: The {\Bbb B}-algebra, isomorphic to the Clifford algebra~C(3,0) of
119: smaller dimension $d=8$, is preferable from this point of view. On the
120: other hand, the {\Bbb B}\,-dynamics, based on GCRE, appears to be Lorentz
121: invariant, so {\bf the {\Bbb B}\,-algebra may be treated as a minimal STA}.
122: This choice leads to the conjectures on a fundamental role of {\it null
123: divisors\/} as a subspace of STA and on complex-valued structure of ST;
124: these questions will be discussed below.
125:
126: Now we are ready to present the contents of the paper.
127: In Sec. 2 we begin with the basic definitions of the {\Bbb B}\,-algebra
128: and {\Bbb B}\,-differentiability. The general problems of (bi-)quaternionic
129: analysis are also briefly discussed.
130: Then, in Sec. 3, after preliminary
131: physical identifications, we demonstrate the 2-spinor structure of the
132: basic GCRE and obtain a complexified eikonal equation for each
133: component of the {\Bbb B}\,-field. Global symmetries of the model are
134: studied as well.
135:
136: Sections 4 and 5 are devoted to {\Bbb B}\,-electrodynamics as the basic case
137: of {\Bbb B}\,-differentiability. Firstly (Sec. 4) the self-duality
138: conditions are obtained from GCRE, whence follow the Maxwell equations.
139: Gauge invariance of a model of special type is demonstrated in Sec. 5. From
140: the eikonal equation, a geometrical origin of the Coulomb law as a
141: stereographic projection becomes evident, and we get for the admissible
142: values of an electric charge $q =\pm 1$, i.e., a quantization rule!
143:
144: In Sec. 6 we demonstrate the equivalence of the theory to
145: geometrodynamics in a complexified Weyl-Cartan space. A reduction to
146: Minkowski space identifies the magnetic monopole field as that of
147: torsion and the Coloumb electric one as the ST Weyl nonmetricity.
148: We conclude in Sec. 7 by the establishment of complex-valued Yang-%
149: Mills equations as the integrability conditions of GRCE and a
150: discussion of general consequences of a complex-quaternionic structure
151: of physical space. Finally, we discuss the relation of the
152: AD approach to binary geometrophysics.
153:
154: \section{{\Bb B}\,-algebra and {\Bb B}\,-differentiability} %Sec.2
155:
156: Let {\bf z}$\in${\Bbb M}\,(4,\,{\Bbb C}), ${\bf z}=\{z^\mu,\ \mu=0,1,2,3\}$
157: be an element of the complex vector
158: space {\Bbb M}\,(4,\,{\Bbb C}) of dimension $d=4$. The function
159: \beq %1
160: {\bf F}({\bf z})=\left\{F^\mu(z)\right\}=\left\{F^\mu(z^0,z^1,z^2,z^3)\right\}
161: \eeq
162: {\bf F}$\in${\Bbb M}, maps an open domain {\Bbb O}$\subset${\Bbb M} to the
163: domain {\Bbb O}$\!{}^\prime\!\subset${\Bbb M}\,; let its components
164: $F^\mu(z)$ be complex and analytic.
165:
166: Then a structure {\Bbb B} of associative algebra of complex quaternions
167: (biquaternions) {\Bbb M}$\times${\Bbb M}$\to${\Bbb M} may be introduced on
168: {\Bbb M}\,. According to the
169: isomorphism {\Bbb B}\ =\ {\Bbb L}(2,\,{\Bbb C}), {\Bbb L} being the full
170: 2$\times$2 complex matrix algebra, we
171: shall use the matrix representation of {\Bbb B}
172: \beq %2
173: \forall{\bf z}\in\mbox{{\Bbb M}} : {\bf z}=z^\mu\sigma_\mu=
174: \left\|\begin{array}{cc}
175: u&w\\p&v\end{array}\right\|,
176: \eeq
177: $\sigma_\mu=\{e,\sigma_a\}$, $e$ being the unit 2$\times$2 matrix and
178: $\{\sigma_a,\ a=1,2,3\}$ the Pauli matrices;
179: $u,v=z^0\pm z^3$; $p,w=z^1\pm iz^2$ are the DeWitt coordinates
180: on {\Bbb M}\,. Now the
181: multiplication $(\ast)$ in {\Bbb B} is equivalent to the usual matrix one;
182: the function (1) becomes a matrix-valued, or {\Bbb B}\,-valued function of a
183: {\Bbb B}\,-variable. Let for some {\bf z}$\in${\Bbb O}
184: \beq %3
185: d{\bf F}={\bf F}({\bf z}+d{\bf z})-{\bf F}({\bf z})
186: \eeq
187: be an infinitesimal increment (differential) of {\bf F}({\bf z}),
188: corresponding to a differential of a {\Bbb B}\,-variable $d{\bf z}$
189: and according to the usual
190: Euclidean metric $\rho^2=\sum_\mu|z^\mu|^2$. Then we come to the
191: following definition.
192:
193: The function (1) {\bf F}({\bf z}) is said to be {\Bbb B}\,-{\it
194: differentiable\/} in some
195: domain {\Bbb O}$\subset${\Bbb M} if for $\forall${\bf z}$\in${\Bbb O} there
196: are some {\bf G}({\bf z}), {\bf H}({\bf z}) such that
197: {\bf the differential (3) may be presented in the invariant form}
198: \beq %4
199: d{\bf F}={\bf G}({\bf z})\ast d{\bf z}\ast{\bf H}({\bf z}),
200: \eeq
201: i.e.\ only through the operation of multiplication in {\Bbb B}\,.
202:
203: For the commutative algebra of complex numbers, from (4) the Cauchy-Riemann
204: (CR) equations follow in the coordinate representation, ${\bf F}^\prime=
205: {\bf G}\ast{\bf H}$ being a derivative of {\bf F}({\bf z}). So the
206: relation (4)
207: naturally generalizes the CR equations to the case of a non-commutative
208: associative {\Bbb B}\,-algebra. Eqs.\,(4) will be further designated as GCRE
209: (in the invariant form).
210:
211: A detailed study of {\Bbb B}\,-differentiability and analyticity,
212: based on GCRE (4), may be found in [3], and a review of other
213: approaches in [9]. The most profound is perhaps Fueter's
214: work [10]; G\"ursey et al.\ [11] applied it within the $d=4$ gauge
215: and chiral theories (see also [12]).
216:
217: \section{Spinor splitting and the eikonal eq\-uation} %\Sec.3
218:
219: Let us turn now to the construction of field theory, based on
220: the concept of {\Bbb B}\,-differentiability. Consider a subspace $\mbox{{\Bbb
221: M}}_+\subset\mbox{\Bbb M}$ of the
222: points with real coordinates
223: ${\bf x}=\{x^\mu\}=\{z^\mu:{\rm Im}\,(z^\mu)=0\}$, or else the
224: subspace of Hermitian matrices with elements ${\bf z}^+={\bf z}$.
225: The {\Bbb B}\,-norm
226: $N^2({\bf z})={\rm Det}\,({\bf z})$ then generates on $\mbox{{\Bbb M}}_+$ the
227: real Minkowski metric
228: \bearr %5
229: N^2({\bf x})={\rm Det}\,({\bf x})=uv-pw \nnn
230: \cm = \left(x^0\right)^2-\left(x^1\right)^2
231: -\left(x^2\right)^2-\left(x^3\right)^2
232: \ear
233: so $\mbox{\Bbb M}_+$ may be identified with the physical ST. Solutions
234: to (4) on {\Bbb M} may be obtained by analytic continuation from
235: $\mbox{\Bbb M}_+$. We will return to a detailed study of the relation
236: between {\Bbb M} and $\mbox{\Bbb M}_+$ in Sec. 7.
237:
238: It is now evident that the {\Bbb B}\,-differentiable functions {\bf
239: F}($x$), realizing the mappings ${\bf F}\colon\mbox{\Bbb M}_+\to\mbox{\Bbb
240: M}$, should be considered as a fundamental physical field; its spinor
241: nature will be seen below. We will assume {\bf the dynamics of a basic
242: F-field to be completely determined by the GCRE (4) with} ${\bf z}={\bf
243: x}\in\mbox{\Bbb M}_+$, i.e.\
244: \beq %6
245: d{\bf F}={\bf G}(x)*d{\bf x}*{\bf H}(x)
246: \eeq
247: Except direct physical identifications of the abstract variables, in what
248: follows no other assumptions will be necessary.
249:
250: Let us rewrite now the matrices {\bf F}, {\bf H} in (6) in the form
251: \beq %7
252: {\bf F}=\|\psi(x),\eta(x)\|,\quad{\bf H}=\|\alpha(x),\gamma(x)\|
253: \eeq
254: each of $\psi,\eta$ and $\alpha,\gamma$ being a matrix-column with two
255: components; the columns transform independently through left multiplication.
256: Then (6) splits into a pair of equations:
257: \beq %8
258: d\psi={\bf G}*d{\bf x}*\alpha,\quad d\eta={\bf G}*d{\bf x}*\gamma.
259: \eeq
260:
261: From (7) and (8) it follows that each solution to (6) may be
262: presented in the form ${\bf F}(x)=\|\psi^\prime(x),\psi^{\prime\prime}(x)\|$
263: where $\psi^\prime,\psi^{\prime\prime}$ are two arbitrary
264: solutions of the unique irreducible equation
265: \beq %9
266: d\psi={\bf G}(x)*d{\bf x}*\alpha(x).
267: \eeq
268:
269: The functions $\psi(x),\alpha(x)$ belong to the {\it left-side ideal\/} of
270: the Clifford algebra $\mbox{\Bbb B}=\mbox{C(3,\,0)}$ and are therefore
271: obviously {\it 2-spinors}. A conjugated spinor reduction of (6) is also
272: possible, if the row splitting of {\bf G}($x$) is used; a double reduction
273: may be realized as well.
274:
275: These properties stand side by side with the widest symmetry
276: group of Eqs.\,(4) or (6), including the transformations
277: \beq %10
278: \left.\begin{array}{@{}r@{\,}c@{\,}lr@{\,}c@{\,}l}
279: {\bf z}&\to&{\bf m}*{\bf z}*{\bf n}^{-1},&{\bf F}&\to&{\bf k}*{\bf
280: F}*{\bf l} \\[2pt]
281: {\bf G}&\to&{\bf k}*{\bf G}*{\bf m}^{-1},&{\bf H}&
282: \to&{\bf n}*{\bf H}*{\bf l}
283: \end{array}\right\},
284: \eeq
285: {\bf m},\,{\bf n},\,{\bf k},\,{\bf l} being arbitrary
286: constant biquaternions of
287: unit norm (neglecting the dilatations ${\bf z}\to\lambda{\bf z}$,
288: $\lambda\in\mbox{\Bbb C}$), ${\bf m}^{-1},{\bf n}^{-1}$ are the inverse ones.
289:
290: {\bf Z}-transformations in (10) define a 6{\Bbb C}\,-parameter group of
291: rotations SO\,(4,\,{\Bbb C});
292: the restriction of this group to $\mbox{\Bbb M}_+$
293: (with ${\bf n}^{-1}={\bf m}^+$)
294: leads to the Lorentz transformations for {\bf x}. Now, if we put in (10)
295: ${\bf k}={\bf n}$, ${\bf l}={\bf m}^{-1}$, the functions {\bf F}, {\bf G},
296: {\bf H} manifest their nature as
297: 4-vectors (${\bf F}\to{\bf n}*{\bf F}*{\bf m}^{-1}$ etc.). However, when
298: ${\bf k}={\bf n}$, ${\bf l}={\rm Ident.}$, {\bf G}
299: transforms as a 4-vector, for {\bf F} and {\bf H} we have ${\bf F}\to
300: {\bf n}*{\bf F}$, ${\bf H}\to{\bf n}*{\bf H}$, preserving the structure of
301: the spinor splitting (8). Moreover, a double
302: row-column splitting of (6) corresponds to the case ${\bf k}={\bf l}={\rm
303: Ident.}$, when
304: {\bf F}($x$) has to be considered as a scalar, while
305: {\bf G} and {\bf H} transform
306: as conjugated spinors ${\bf G}\to{\bf G}*{\bf m}^{-1}$, ${\bf H}\to{\bf n}*
307: {\bf H}$.
308:
309: Let us return now to the dynamical consequences of GCRE (6) and (9).
310: Using the {\it Fiertz identity\/} and the double row-column splitting of (8),
311: for every matrix component $\psi={\bf F}_{AB}$; $A,B\!=\!1,2$ of {\bf F}-field
312: we get~[13]
313: \beq %11
314: (\partial_0\psi)^2-(\partial_1\psi)^2-(\partial_2\psi)^2-(\partial_3\psi)^2=0.
315: \eeq
316:
317: Hence, {\bf every matrix component of
318: a {\Bbb B}\,-differ\-en\-tiable function
319: satisfies the nonlinear, Loren\-tz invariant, complexified eikonal
320: equation} (11). For the {\Bbb B}\,-algebra it plays a role similar to that
321: of the Laplace equation in complex analysis; as for physics, its
322: fundamental properties (for $\psi\in\mbox{\Bbb R}$) were emphasized by
323: V.A. Fock~[14].
324:
325: \section{{\Bb B}\,-electrodynamics.~Self-duality co\-nditions and Maxwell
326: equations} %Sec. 4
327:
328: We shall further restrict ourselves to the case of the spinor
329: equality $\alpha(x)=\psi(x)$ in (9), i.e.\ to the fundamental equation
330: \beq %12
331: d\psi={\bf G}(x)*d{\bf x}*\psi(x).
332: \eeq
333: For (12) the global continuous symmetries (10) are reduced to
334: the transformations of the Lorentz group
335: \beq %13
336: {\bf x}\to{\bf m}*{\bf x}*{\bf m}^+,\ \psi\to{\bf s}\psi,\
337: \overline{\bf G}\to{\bf m}*\overline{\bf G}*{\bf m}^+,
338: \eeq
339: where ${\bf s}=({\bf m}^+)^{-1}$, $\overline{\bf G}$ is a {\Bbb
340: B}\,-conjugated field: $\overline{\bf G}*{\bf G}=({\rm Det}\,{\bf G})^2$.
341:
342: So relativistic invariance is ensured, and the conjugated field
343: $\overline{\bf G}(x)$ forms a 4-vector. Later on $\overline{\bf G}(x)$ will
344: be regarded as a {\Bbb C}\,-valued
345: matrix of electromagnetic (EM-) 4-potential {\bf A}($x$). Precisely, we set
346: \beq %14
347: A_\mu(x)=2\overline{G}_\mu(x)\equiv 2G^\mu(x).
348: \eeq
349: Such an identification will be justified further by its dynamic and
350: geometric consequences, as well as by the establishment of gauge
351: invariance of (12). Therefore, the latter {\bf can be considered as the
352: basic equations of {\Bbb B}\,-electrodynamics}, i.e., some type of classical
353: {\it spinor electrodynamics}, generated by solely the GCRE-structure.
354:
355: Written in components, Eqs. (12) form the set of differential equations
356: \beq %15
357: \left.\begin{array}{@{}l@{\ }l@{\ }l@{\ }l@{}}
358: \partial_uf\!=\!G^u\!f,&\partial_pf\!=\!G^w\!f,&\partial_wf\!=\!G^u\!h,&
359: \partial_vf\!=\!G^w\!h\\[2pt]
360: \partial_uh\!=\!G^p\!f,&\partial_ph\!=\!G^v\!f,&\partial_wh\!=\!G^p\!h,&
361: \partial_vh\!=\!G^v\!h
362: \end{array}\right\}.
363: \eeq
364: Here $f(x)$ and $h(x)$ are the components of a 2-spinor field $\psi(x)$, and
365: $\partial$
366: denotes a partial derivative with respect to the corresponding DeWitt
367: coordinate.
368:
369: The equations for the EM field follow from the over-det\-er\-mined system
370: (15) as its {\it integrability (compatibility) conditions}
371: \[
372: \partial_\mu(\partial_\nu\psi)-\partial_\nu(\partial_\mu\psi)=0,\quad
373: \psi=\{f(x),h(x)\}.
374: \]
375: Assuming then both $f(x),\ h(x)\not\equiv 0$ (otherwise
376: we would have obtained the same final results), we obtain after derivation
377: \beq %16
378: \left.\begin{array}{@{}r@{\,}c@{\,}l@{\ }r@{\,}c@{\,}l@{}}
379: \partial_uA_w-\partial_pA_u&=&0,&\partial_wA_w-\partial_vA_u&=&
380: \frac{1}{2}\,{\rm Det}\,{\bf A}\\[2pt]
381: \partial_vA_p-\partial_wA_v&=&0,&\partial_pA_p-\partial_uA_v&=&
382: \frac{1}{2}\,{\rm Det}\,{\bf A}
383: \end{array}\right\},
384: \eeq
385: $A_\mu(x)$ being the EM potentials (14)
386: and ${\rm Det}\,{\bf A}=A_uA_v-A_pA_w$. Going back to the
387: Cartesian coordinates, we observe that Eqs. (16)
388: are equivalent to the {\it self-duality conditions\/} (SDC)
389: \beq %17
390: \vec{\cal P}\equiv\vec{\cal E}+i\vec{\cal B}=0
391: \eeq
392: for the {\Bbb C}\,-valued electric $\vec{\cal E}=\{{\cal E}_a\}$ and magnetic
393: $\vec{\cal B}=\{{\cal B}_a\}$ components of the EM field strength tensor
394: \beq %18
395: {\cal F}_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu;
396: \eeq
397: here
398: \beq %19
399: {\cal E}_a={\cal F}_{{\scriptscriptstyle 0}a},\ {\cal B}_a=
400: {\textstyle\frac{1}{2}}\,\varepsilon_{abc}
401: {\cal F}_{bc};\ a,b,c,\dots=1,2,3.
402: \eeq
403: In addition to (17), from (16) we have
404: \beq %20
405: {\cal D}\equiv\partial_\mu A^\mu + 2A_\mu A^\mu=0,
406: \eeq
407: i.e.\ an {\it inhomogeneous Lorentz condition}.
408:
409: Combined with the definitions (18) and (19), the {\bf SDC (17) lead then
410: to the Maxwell equations in free space}
411: \beq
412: \partial_\nu{\cal F}^{\mu\nu}=\partial_\nu\left({\textstyle\frac{1}{2i}}
413: \varepsilon^{\mu\nu\rho\lambda}{\cal F}_{\rho\lambda}\right)=0.
414: \eeq
415: So the Maxwell equations represent nothing but the {\it consistency
416: conditions\/} of a basic GCRE-system and {\bf are satisfied identically
417: for each solution to the latter}. The inverse statement generally does
418: not take place!
419:
420: Now, it is easy to see that, according to the SDC (17), the {\it energy-%
421: momentum density\/} of a complex-valued EM-field turns to zero. Therefore,
422: we ought to define the {\bf physical} fields $\vec E$, $\vec B$ through the
423: real (Re) or imaginary (Im) parts of (19). For geometric reasons (see part 6),
424: we prefer
425: \beq %21
426: \vec E=2{\rm Re}\,(\vec{\cal E}\,),\quad\vec B=2{\rm Re}\,(\vec{\cal B}\,).
427: \eeq
428: The {\Bbb R}-valued vectors $\vec E$, $\vec B$ satisfy the linear Maxwell
429: equations as well.
430: However, they are mutually independent (contrary to (19)) and
431: create a non-zero energy-momentum density ($W$, $\vec P$) of the usual form
432: \beq %22
433: W\sim\left(|\vec E|^2+|\vec B|^2\right),\quad\vec P\sim[\vec E\times\vec
434: B].
435: \eeq
436: Moreover, an infinite series of conservation laws can be obtained for
437: (15) using routine procedures (see [15] for an example).
438:
439: \section{Coulomb field as a stereographic map. Electric charge quantization}
440: %Sec.5
441:
442: Let us search now for solutions to the {\Bbb B}\,-electrodyna\-mic
443: equations (15). Each of the two components $f(x)$, $h(x)$ of the spinor field
444: in (15) satisfies the {\Bbb C}\,-eikonal equation (11). Starting from one
445: of its solutions, all the other quantities, including the EM potentials (14),
446: should be derived. In particular, the wave-like solutions
447: of (11) lead to EM fields, identical to the usual EM waves [3,\,13].
448:
449: Notice now that {\bf the eikonal equation (11) possesses a wonderful
450: invariance property} under the transformations
451: \beq %23
452: \psi(x)\to\Phi\left(\psi(x)\right)
453: \eeq
454: with an arbitrary ({\Bbb C}\,-differentiable) function $\Phi(f)$.
455: Accordingly,
456: one can easily verify the {\it gauge invariance\/} of the basic system (12)
457: (and, therefore, (15)) of a special type:
458: \beq
459: \psi(x)\to\psi(x)\alpha(\psi),\ A_\mu(x)\to A_\mu(x)\!+\!\partial_\mu
460: \ln\alpha(\psi),
461: \eeq
462: $\alpha(\psi)$ being an arbitrary scalar function of the $\psi$ components
463: $f(x)$, $h(x)$.
464:
465: The {\Bbb C}\,-structure of the eikonal equation (11) essentially enlarges
466: the spectrum of its solutions. The most important are certainly two
467: static solutions found in [3]:
468: \beq %24
469: \left.\begin{array}{@{}l}
470: f^+={\displaystyle\frac{x^1+ix^2}{r+x^3}}=\tan\left(\frac{\theta}{2}\right)
471: \exp\,(i\varphi)\\[7pt]
472: f^-={\displaystyle\frac{x^1-ix^2}{r-x^3}}=\cot\left(\frac{\theta}{2}\right)
473: \exp\,(-i\varphi)
474: \end{array}\right\},
475: \eeq
476: where $\{r,\theta,\varphi\}$ are usual spherical
477: coordinates on $\mbox{\Bbb R}^3$. From a geometric
478: point of view, the expression (26) {\bf corresponds to the stereographic
479: projection} $\mbox{\Bbb S}^2\to\mbox{\Bbb C}$ of a unit 2-sphere onto the
480: {\Bbb C}\,-plane (from the south
481: (+) or north (-) poles, respectively). Substituting (26) into (15),
482: (14), we get after trivial integration:
483: \beq %25
484: \left.\begin{array}{@{}r@{\,}c@{\,}l@{}r@{\,}c@{\,}l@{}}
485: f(x)&=&f^\pm(\theta,\varphi),&h(x)&=
486: &\left[f^\pm(\theta,\varphi)\right]^2\\[5pt]
487: A_u&=&\mp{\displaystyle\frac{1}{r}},&A_v&=&\pm{\displaystyle\frac{2}{r}}\\[5pt]
488: A_p&=&{\displaystyle\frac{e^{-i\varphi}\!\left(\tan\frac{\theta}{2}
489: \right)^{\mp 1}}{r}},&
490: A_w&=&-{\displaystyle\frac{2e^{i\varphi}\!\left(\tan\frac{\theta}{2}
491: \right)^{\pm 1}}{r}}
492: \end{array}\right\}
493: \eeq
494: or, for spherical components of the {\Bbb C}\,-valued 4-potential
495: \bearr %26 or 28
496: A_0=\pm\frac{1}{2r},\quad A_r=-\frac{1}{2r},\quad A_\theta=\mp
497: iA_\varphi=\nnn
498: -\frac{1}{2r}\cot\theta\pm\frac{3}{2r\sin\theta}.
499: \ear
500:
501: Now, a transition to the physical vectors of EM-field
502: strengths (22) shows that the {\it magnetic monopole\/} and gradient-like
503: terms in (28) disappear and we get
504: \beq
505: E_\theta=E_\varphi=B_r=B_\theta=B_\varphi=0,\quad E_r=\pm\frac{1}{r},
506: \eeq
507: i.e.\ {\bf the Coulomb law with a fixed value of electric charge} $q=\pm 1$.
508:
509: Whereas the stereographic projection (26) and the transformations (24)
510: realize the conformal mappings $\mbox{\Bbb S}^2\to\mbox{\Bbb C}$, $\mbox{\Bbb
511: C}\to\mbox{\Bbb C}$ respectively, EM fields
512: behave by (25) in a gauge invariant manner, and the
513: electric charge remains quantized. An exceptional role of conformal
514: mappings in algebrodynamics has been clarified in [3] (chapter 1).
515:
516: $q$-quantization is a crucial point for {\Bbb B}\,-electrody\-na\-mics;
517: a fundamental significance of this problem has been evident to Dirac,
518: Eddington, Wheeler and other grands. In orthodox
519: field theory the $q$-quantization is postulated rather than explained.
520: The most elegant approach to this problem is produced, perhaps, by
521: multidimensional ST theories [16]; {\Bbb B}\,-electrodynamics presents
522: another possibility.
523:
524: In our approach the algebraic and purely classical origin of
525: $q$-quantization becomes evident. The fact is that the initial GCRE are
526: not invariant under the scaling ${\bf A}\to\lambda{\bf A}$, contrary to
527: the linear Maxwell
528: equations. We suppose, however, that the phenomenon of {\bf ``algebraic
529: $q$-quantization''} should have deeper topological reasons; we
530: hope to discuss them in the future [20].
531:
532: \section{Spinor connection and Weyl-Cartan geometry of ST} %Sec.6
533:
534: The fundamental equation of {\Bbb B}\,-electrodynamics (12) may be presented
535: in the form
536: \beq %29
537: \partial_\nu\psi={\bf\Gamma}_\nu(x)\,\psi(x),
538: \eeq
539: with
540: \beq %30
541: {\bf\Gamma}_\nu(x)={\bf G}(x)*\sigma_\nu
542: \eeq
543: being a {\it 2-spinor connection\/} of special type. The initial GCRE,
544: corresponding to (30), have the matrix form (6) with ${\bf H}(x)={\bf F}(x)$,
545: i.e.
546: \beq %31
547: d{\bf F}={\bf G}(x)*d{\bf x}*{\bf F}(x),
548: \eeq
549: or, in a 4-vector representation [3]
550: \bearr %32,33
551: \partial_\nu F^\mu=\Gamma^\mu_{\nu\!\rho}(x)\,F^\rho, \\ \lal
552: \Gamma^\mu_{\nu\!\rho}(x)=2(A_\nu\delta^\mu_\rho+A_\rho\delta^\mu_\nu-
553: A^\mu\eta_{\nu\!\rho}-i\varepsilon^\mu_{\cdot\nu\!\rho\alpha}A^\alpha)
554: \ear
555: where $\delta^\mu_\nu$, $\eta_{\mu\nu}$, $\varepsilon_{\mu\nu\!\rho\lambda}$
556: are the Kronecker, Minkowski and Levi-Civita
557: tensors, respectively, and $A_\mu(x)$ are {\Bbb C}\,-valu\-ed potentials
558: (14).
559:
560: Thus in the basic electrodynamic case the initial GCRE system (6) is
561: equivalent to the defining equations (32) of the {\it covariantly constant
562: vector fields\/} $\{F^\mu(x)\}$ on a {\Bbb B}\,-manifold with a
563: ``dynamically created'' effective {\bf geometry of Weyl-Cartan type},
564: represented by the affine connection (34). Note that the {\Bbb C}\,-vector
565: $A_\mu(x)$ completely determines both the Weyl part of (34) and its
566: torsion structure. A generalization by introduction of a Riemann
567: metric structure is natural as well.
568:
569: To obtain the ST geometry induced by (34), let us pass
570: from ${\bf F}(x)$ to the unitary field ${\bf U}(x)={\bf F}*{\bf F}^+$.
571: Using (32), we get
572: \beq %34
573: \partial_\nu U^\mu=\Delta^\mu_{\nu\!\rho}(x)\,U^\rho(x),
574: \eeq
575: with the {\Bbb R}-valued connection
576: \beq %35
577: \Delta^\mu_{\nu\!\rho}(x)=2(a_\nu\delta^\mu_\rho+a_\rho\delta^\mu_\nu-
578: a^\mu\eta_{\nu\!\rho}-\varepsilon^\mu_{\cdot\nu\!\rho\alpha}b^\alpha),
579: \eeq
580: where $a_\mu(x)$ and $b_\mu(x)$
581: are real and imaginary parts of the potentials $A_\mu(x)$.
582:
583: A connection similar to (36) has been introduced in Ref.\,[17] from
584: physical considerations; in [18] it was shown to be {\bf the only ST
585: connection compatible with a spinor bundle structure} with the conventional
586: notion of a covariant spinor derivative. In our approach these results
587: follow from the GCRE structure alone.
588:
589: However, the {\it torsion field\/} $b_\mu(x)$ in (36) satisfies the Maxwell
590: equations, as well as the {\it non-metricity field\/} $a_\mu(x)$. By the
591: key
592: Ansatz (28), precisely the Weyl part $a_\mu(x)$ corresponds to the ordinary
593: Coloumb electric field, justifying the previous identification of the EM
594: field with the real part of the {\Bbb C}\,-field.
595:
596: As for the imaginary part $b_\mu(x)$, for (28) it has the magnetic monopole
597: form
598: \beq %36
599: b_0=b_r=b_\theta=0,\quad b_\varphi=\mp\frac{1}{2r}\cot\theta-
600: \frac{3}{2r\sin\theta};
601: \eeq
602: we thus come to an exotic {\bf geometric interpretation of magnetic
603: monopoles as a ST torsion} (with a totally antisymmetric tensor
604: structure). Accordingly, the field $b_\mu(x)$ cannot appear in the
605: equations of geodesics. If we assume the latter to present the laws
606: of test particle motion, then {\bf monopoles should} have no
607: effect on it and therefore {\bf be entirely unobservable}!
608:
609: Let us now return to the study of the primary {\Bbb C}\,-geometry of the
610: {\Bbb B}\,-space.
611: The integrability conditions for the irreducible spinor equation (30) may
612: be written in the form
613: \beq %37
614: {\bf R}_{\mu\nu}\psi(x)=0,
615: \eeq
616: with
617: \beq %38
618: {\bf R}_{\mu\nu}=\partial_{[\mu}{\bf\Gamma}_{\nu]}-[{\bf\Gamma}_\mu,
619: {\bf\Gamma}_\nu]
620: \eeq
621: being the {\it curvature tensor\/} in the matrix representation. For
622: its self-dual components
623: \beq %39
624: (\vec{\bf R})_a={\bf R}_{{\scriptscriptstyle 0}a}+{\textstyle\frac{i}{2}}
625: \varepsilon_{abc}{\bf R}_{bc}
626: \eeq
627: with the connection of the form (31), we get
628: \beq %40
629: \vec{\bf R}=\vec{\cal P}+{\cal D}\vec\sigma-i[\vec{\cal P}\times\vec\sigma\,].
630: \eeq
631: Here the quantities
632: \[
633: \vec{\cal P}=\vec{\cal E}+i\vec{\cal B},\quad{\cal D}=\partial_\mu A^\mu
634: +2A_\mu A^\mu
635: \]
636: coincide with (17) and (20), respectively and therefore vanish
637: along with the entire self-dual tensor (40).
638:
639: It is easy to see that $\cal D$ is proportional to the curvature
640: invariant ${\cal D}=6\eta^{\mu\nu}R^\alpha_{\mu\alpha\nu}\ (=0)$.
641: So the {\bf{\Bbb B}\,-space appears to be an self-dual space with
642: zero scalar curvature}. Now, if there are two {\bf linearly independent}
643: spinor solutions of (30), then, as follows from (38),
644: \beq %41
645: {\bf R}_{\mu\nu}(x)=0,
646: \eeq
647: i.e.\ a trivial case of flat geometry and zero field strengths.
648:
649: To avoid that, the primary {\Bbb B}\,-field ${\bf F}(x)$ in (32) should
650: split into
651: two spinors $\psi^\prime(x)$, $\psi^{\prime\prime}(x)$ (see (9)),
652: proportional to each other; therefore, we have
653: \beq %42
654: {\rm Det}\,{\bf F}(x)=0,
655: \eeq
656: and the field ${\bf F}(x)$ takes the values on the subspace of null divisors
657: of the {\Bbb B}\,-algebra, or, physically, {\bf on the complex ``light
658: cone''}.
659:
660: The null {\Bbb B}\,-fields are the most fundamental objects throughout
661: the AD approach as a whole. However, they can exist only on manifolds
662: with an indefinite metric signature. So the {\bf pseudo-Euclidean structure}
663: of the World should not be postulated within the AD app\-roach, but {\bf is
664: just a necessary condition of nontrivial dynamics (and effective
665: geometry)}.
666:
667: \section{Yang-Mills fields and the {\Bb C}\,-structure of space-time}
668: %Sec.7
669:
670: Now we will demonstrate that the Yang-Mills (YM) gauge
671: fields also appear in theory in rather a natural way. To see that, let
672: us separate the {\it trace-free part\/} in the basic spinor connection (31)
673: \beq %43
674: {\bf\Gamma}_\nu(x)={\bf G}(x)*\sigma_\nu={\textstyle\frac{1}{2}}
675: \bigl(A_\nu(x)+ {\bf N}_\nu(x)\bigr).
676: \eeq
677: Then the zero component $A_\mu(x)$
678: coincides with the {\Bbb C}\,-potentials (14) of the EM field, and the
679: trace-free part ${\bf N}_\mu(x)$ can be expressed in its terms in a linear
680: way:
681: \bearr %45
682: {\bf N}_\mu(x)=N_\mu^a(x)\,\sigma_a;
683: \quad N^a_{\scriptscriptstyle 0}=A_a(x),\nnn
684: \quad N^a_b=\delta_{ab}
685: A_{\scriptscriptstyle 0}(x)-i\varepsilon_{abc}A_c(x).
686: \ear
687: The quantities ${\bf N}_\mu(x)$ can be regarded as the matrix potentials of
688: some {\Bbb C}\,-valued gauge field; its strength corresponds to the traceless
689: part of the curvature tensor (39) and may be written as usual:
690: \beq %46
691: {\bf L}_{\mu\nu}={\cal L}^a_{\mu\nu}(x)\,\sigma_a=\partial_{[\mu}
692: {\bf N}_{\nu]}-[{\bf N}_\mu,{\bf N}_\nu].
693: \eeq
694:
695: We see now that the self-dual part of (46) coincides with the traceless
696: part of the tensor (41) and, in view of (17) and (20), we have again
697: \beq %47
698: {\bf L}_{\mu\nu}+{\textstyle\frac{i}{2}}\varepsilon_{\mu\nu\!\rho\lambda}
699: {\bf L}^{\rho\lambda}=0.
700: \eeq
701: From (47) and the {\it Bianchi identity\/}, the YM equations
702: follow immediately in a usual way:
703: \beq %48
704: \partial_\nu{\bf L}^{\mu\nu}=[{\bf N}_\nu,{\bf L}^{\mu\nu}].
705: \eeq
706: So we can indeed consider the field ${\bf N}_\nu(x)$ as a {\Bbb C}\,-valued
707: YM field
708: of a special structure (45). From (38) for any non-trivial $\psi(x)$ we
709: obtain, in addition,
710: \beq %49
711: {\rm Det}\,{\bf R}_{\mu\nu}=0;
712: \eeq
713: written in components, this condition leads to an expression of the EM field
714: strength in terms of the YM ones (for each $[\mu\nu]$ separately):
715: \beq %50
716: {\cal L}^a_{\mu\nu}{\cal L}^a_{\mu\nu}=({\cal F}_{\mu\nu})^2.
717: \eeq
718: So {\bf the EM field may be regarded as a modulus of the YM triplet field
719: in the isotopic complexified 3-space.}
720:
721: Contrary to EM fields, the YM ones cannot be split into real and
722: imaginary parts (due to the nonlinearity of the YM equations) and
723: therefore are essentially {\Bbb C}\,-valued. This seems quite natural in
724: connection with the
725: pseudo-Euclidean structure of the ST (the duality operator is known to
726: have imaginary eigenvalues in the Lorentz signature). Since the
727: self-duality conditions play a crucial role both in orthodox field
728: theory and in AD, we come again to the {\bf conjecture on a {\Bbb
729: C}\,-analytic structure of real physical ST}. This possibility, discussed
730: repeatedly
731: within the frames of GRT, the twistor and string programs, as well as within
732: the {\it binary geometrophysics\/} approach [19], seems to be inevitable in
733: AD in view of the non-existence of a {\Bbb R}-valued STA with an Aut group
734: isomorphic to the Lorentz one. Thus, we suppose that close connections
735: between the field equations nonlinearity and the {\Bbb C}\, structure of
736: ST do exist, as well as its noncommutative quaternion structure (see
737: Sections 1 and 3 for the latter).
738:
739: Moreover, {\bf we may think of the {\Bbb C} structure as some natural
740: way of ST dimension enlarging} (namely, doubling), just in the sense of
741: Kaluza-Klein theories. As for physics, such an effect should be
742: essential at high energies; asymptotically, in the linear approximation,
743: the ST {\Bbb C}\, structure should split into the Minkowski space observed
744: plus a conjugated one. The same is done by the field {\Bbb C}\,-structure:
745: it exhibits a reduction to a linear {\Bbb R}-valued EM-field (doubled
746: through the SDC (17), too).
747:
748: Generally, we assume the existence of a biquaternion (i.e.\
749: complex-quaternion) algebraic structure of ST and field manifolds
750: consistent with each other. The {\bf non-commutativity} of such a
751: {\Bbb B}\,-algebra
752: {\bf results in the nonlinearity of fundamental dynamics}: the GCRE (6),
753: {\bf
754: as well as its P- and even T-noninvariance} (the connections similar
755: to (35), (36) are efficiently employed by V.G. Krechet for a 5-geometrical
756: description of electroweak interaction). Nevertheless, here the
757: usual reversible dynamics of gauge fields has been obtained in Sections
758: 4 and 7; this latter should be regarded as nothing more but some ``trace''
759: of a primary {\Bbb B}\,-structure, responsible for
760: interactions, the ``time arrow'' and the left-right preference
761: on the Minkowski ST. We expect an
762: extensive presentation of our views of these problems as well
763: as numerous generalizations of the AD approach.
764:
765: In conlusion, peculiar correlations between AD and binary
766: geometrophysics (BG) [19] should be noted. Both of the approaches
767: start from some abstract exceptional algebraic structures and deal
768: with either basic relations (in BG), or special mappings (in
769: AD). In both theories {\bf zero determinant structures} (see (43), (49))
770: {\bf are
771: of particular importance}. Finally, the ideas of {\it multipoint geometries\/}
772: [19,\,20] originate from purely algebraic considerations and should find
773: their place in AD as well. It seems plausible that other
774: deep interrelations will be found out in future.
775:
776: We see that the simplest AD model, based on the conditions of {\Bbb
777: B}\,-differentiality alone, naturally contains the geometric, spinor-gauge
778: and discrete structures, capable of solving the charge quantization and
779: monopole problems. Within this model, the Coulomb law gains an exotic
780: geometrical meaning, and the {\Bbb C}\,-eikonal equation becomes a
781: fundamental equation of field dynamics. Related problems (in particular,
782: the problem of motion law and many-sources distributions) are yet to be
783: solved.
784:
785:
786: \Acknow{I am grateful to D.V.~Alexeevsky, B.V.~Medvedev and especially to
787: Yu.S.~Vladimirov for helpful advice and (Yu.S. Vladimirov) for
788: organizational support.}
789:
790: \small
791: \begin{thebibliography}{99}
792:
793: \bibitem{1}
794: V.I.~Rodichev, {\it Izvestiya Vuzov, Fizika\/}, 1963, No.2,
795: 122 (in Russian).
796: \bibitem{2}
797: S.~Mandelstam, {\it Ann. Phys.\/} {\bf 19} (1962), 25.
798: \bibitem{3}
799: V.V.~Kassandrov, ``Algebraic Structure of Space-Time and
800: Algebrodynamics'', Peoples' Friend. Univ. Press, Moscow, 1992 (in
801: Russian).
802: \bibitem{4}
803: D. Hestenes, ``Space-Time Algebra'', N.Y., Gordon \& Breach, 1966.
804: \bibitem{5}
805: G.~Casanova, ``Vector algebra'', Presses Univers.\ France, 1976.
806: \bibitem{6}
807: F.~G\"ursey and H.G.~Tze, {\it Phys. Lett.\/} {\bf B 127} (1983), 191.
808: \bibitem{7}
809: ``Quazigroups and Nonassociative Algebras in Physics'',
810: (J.~L\"ohmus and P.~Kuusk, eds.), Proc.\ Inst.\ Phys.\ Estonia Ac.\ Sci.,
811: vol.66, Tartu, 1990.
812: \bibitem{8}
813: N.~Salingaros, {\it J.\ Math.\ Phys.\/} {\bf 23} (1982), 1.
814: \bibitem{9}
815: V.V.~Vishnewsky, A.P.~Shirokov and V.V.~Shurygin, ``Spaces over
816: Algebras'', Kasan Univ. Press, 1985 (in Russian).
817: \bibitem{10}
818: R.~Fueter, {\it Commun.\ Math.\ Helv.\/} {\bf 4} (1931--32), 9.
819: \bibitem{11}
820: F.~G\"ursey and H.G.~Tze, {\it Ann.\ Phys.\/} {\bf 128} (1980), 29.
821: \bibitem{12}
822: M.~Evans, F.~G\"ursey and V.~Ogievetsky, {\it Phys.\ Rev.\/} {\bf D47}
823: (1993), 3496.
824: \bibitem{13}
825: V.V.~Kassandrov, {\it Vestnik Peopl.\ Fried.\ Univ., Fizika}, 1993,
826: No.1, 60 (in Russian).
827: \bibitem{14}
828: V.A.~Fock., ``Theory of Space, Time and Gravity'',
829: IIL, Moscow, 1955 (in Russian).
830: \bibitem{15}
831: M.K.~Prasad, {\it Phys.\ Lett.\/} {\bf B 87} (1979), 237.
832: \bibitem{16}
833: Yu.S.~Vladimirov, ``Physical Space-Time Dimension and
834: Unification of Interactions'', Moscow Univ.\ Press, 1987 (in
835: Russian).
836: \bibitem{17}
837: Yu.N. Obukhov, V.G. Krechet and V.N. Ponomariev, {\it in\/}:
838: ``Gravitation and Relativity Theory'', Kazan, 1978, No.14--15, 121.
839: \bibitem{18}
840: V.E.~Stepanov, {\it Izvestiya Vuzov, Mathematica\/}, 1987, No.1, 72.
841: \bibitem{19}
842: Yu.I.~Kulakov, Yu.S.~Vladimirov and A.V.~Karnaukhov,
843: Introduction to Physical Structures Theory and Binary
844: Geometrophysics, Moscow, Arkhimed Press, 1992 (in Russian).
845: \bibitem{20}
846: V.Ya.~Skorobogat'ko, G.N.~Feshin and V.A.~Pielykh, {\it in\/}:
847: ``Math.\ Methods and Physico-Mechanical Fields'', Kiev, Naukova
848: Dumka, (1975), No.1, 5.
849:
850: \end{thebibliography}
851:
852: \end{document}