Isogroups and isosubgroups
Abstract
The main goal of this paper is to give a mathematical foundation, serious and consistent, to some parts of Santilli’s isotheory. We study the isotopic liftings of groups and subgroups and we also deal with the differences between an isosubgroup and a subgroup of an isogroup. Finally, some links between this isotheory and the standard groups theory, referred to representation and equivalence relations among groups are shown.
Full text
RACSAM
Re . R. Acad. Cien. Se ie A. Ma .
VOL.97 (1), 2003, pp. 1–12
´
Algeb a / Algeb a
Isog oups and isosubg oups
Ra´ul M. Falc´
on and Juan N´u˜
nez
Abs ac . The main goal o his pape is o gi e a ma hema ical ounda ion, se ious and consis en , o
some pa s o San illi’s iso heo y. We s udy he iso opic li ings o g oups and subg oups and we also deal
wi h he di e ences be ween an isosubg oup and a subg oup o an isog oup. Finally, some links be ween
his iso heo y and he s anda d g oups heo y, e e ed o ep esen a ion and equi alence ela ions among
g oups a e shown.
Isog upos e isosubg upos
Resumen. El p incipal obje i o de es e a ´
ıculo es p opo ciona un undamen o ma em´
a ico, consis en e
y igu oso, a de e minadas pa es de la iso eo ´
ıa de San illi. En ´
el se ealiza el le an amien o iso ´
opico de
los g upos y subg upos, es udi´
andose asimismo la di e encia en e un isosubg upo y un subg upo de un
isog upo. Se mues an inalmen e algunas elaciones en e es a iso eo ´
ıa y la eo ´
ıa s anda d de g upos,
e e en es a los emas de ep esen aci´
on de g upos y de elaciones de equi alencia en e g upos.
1.In oduc ion
In 1978, he I alian-Ame ican heo e ical physicis and ma hema ic Rugge o Ma ia San illi p oposes a
gene aliza ion o con en ional Lie’s heo y by using he concep o iso opy (in he G eek sense o being
”axiom-p ese ing”, also called iso opic li ing), which implies he o igin o he ac ually known like San-
illi’s iso heo y (see [2]). To do his, he ex ends he basic uni I=(+1,diag(+1, ..., +1), ...) o he ini ial
s uc u e o a gene alized uni b
I=b
I(x, •
x,••
x, ..., µ, τ, ...), called isouni , which depends on he coo dina e x
and hei de i a i es, on he densi y µ, on he empe a u e τand, in gene al, on any magni ude o he physic
en i onmen o he sys em in which we a e. By using i , San illi does a s ep-by-s ep gene aliza ion o he
mo e impo an ma hema ical s uc u es, ob aining o he new ones, cha ac e ized by he ac o ha ing he
same p ope ies as he ini ial ones, while he new uni s sa is y mo e gene al condi ions han he e i ied
by he ini ial ones. San illi gi es he name o ma hema ical isos uc u es o hese new s uc u es. In his
way, he s udies isog oups,iso ings,iso ields,iso ec o spaces and isoalgeb as (see [3], [4], [5] and [11], o
ins ance).
I allowed him o ge in a as way some de elopmen o physical applica ions, p incipally in Quan um
Mechanics and Dynamical P oblems o pa icles and an ipa icles. San illi’s iso opies allow o map any
gi en and ixed linea , local and canonical s uc u e in o i s mos gene al possible non-linea , non-local
P esen ado po Jes´
us Ilde onso D´
ıaz.
Recibido: 6 de Feb e o de 2002. Acep ado: 30 de Oc ub e de 2002.
Palab as cla e / Keywo ds: Iso heo y, Lie-San illi, isog oup, isosubg oup
Ma hema ics Subjec Classi ica ions: 17B99.
c
°2003 Real Academia de Ciencias, Espa˜
na.
1
R. M. Falc´
on and J. N´
u˜
nez
and non-canonical o ms which a e capable o econs uc ing linea i y, locali y and canonici y in ce ain
gene alized isospaces and iso ields wi hin he ixed ine ial coo dina es o he obse e .
Howe e , in he las yea s, San illi has ound some ma hema ical inconsis encies in his ea ly o mula ion
o he iso heo y. Due o i , San illi and o he ma hema icians ha e s udied he iso opic li ings o Func ional
Analysis and Di e en ial Calculus (see [1] and [6]). I has allowed o ge some impo an applica ions in
Physics (see [7], [8], [9] and [10], o ins ance).
So, as San illi’s iso heo y needs e en a consis en ma hema ical ounda ion, San illi himsel has p o-
posed se e al subjec s o esea ch o he in e na ional ma hema ical-scien i ic communi y. One o hem
consis s on p o ing he exis ence o isos uc u es co esponding o he li ing o s uc u es al eady known,
al hough hey do no ha e a p ac ical applica ion in Physic. San illi hinks ha i would be good o con ince
he scien i ic abou he ele ance o his esea ch, which would gi e bigge consis ence and eliance o his
iso heo y.
In his pape , we y o pa ially esponse o San illi’s pe i ion, by s udying a possible li ing o he
simples algeb aic s uc u e: he g oup s uc u e. Howe e , as he e is al eady some s udies abou i (see
[11]), we comple e hem, gi e also some examples and show how he subg oups can be iso opically li ed
by using he San illi’s model o cons uc isop oduc s. Ge ing his las ques ion is he main goal o his
pape .
To do his, we p e iously gi e in Sec ion 2 some basic de ini ions ela ed o iso opic li ings. Sec ion 3 is
de o ed o he s udy o isog oups. Nex , in Sec ion 4, we ob ain Theo em 2, which assu es he cons uc ion
o an isosubg oup, gi ing some examples, oo. We also dis inguish be ween subg oups and isosubg oups o
an isog oup. Las wo sec ions a e de o ed o s udy some applica ions o San illi’s iso heo y: ep esen a ion
and equi alence ela ions among isog oups, espec i ely, and some links among hese concep s and he
s anda d g oups heo y.
2.P elimina ies
Remembe ha o a gi en and ixed ma hema ical s uc u e, an iso opy o iso opic li ing is any li ing o i ,
which gi es a new ma hema ical s uc u e e i ying he same basic axioms (o p ope ies) as he i s . This
new s uc u e is called iso opic s uc u e o isos uc u e (see [2]).
In 1978 (see [2]), San illi p oposes a possible model o iso opy, called San illi’s iso opy, which allows o
cons uc he named ma hema ical isos uc u e, based on an isouni I. This isouni can be ob ained s a ing
om he ollowing de ini ion:
Le Ebe any ma hema ical s uc u e, de ined on a se o elemen s C. Le V⊇Cbe a se wi h an inne
law ∗and an uni elemen I. Such a se Vis said o be he gene al se o he iso opy. Le b
I∈Vbe such
ha i s in e se T=b
I−I, wi h espec o he law ∗, exis s. Iwill be called iso opic uni o isouni and i will
be he basic uni in he li ing o he s uc u e E.Twill be he iso opic elemen . Finally, b
Iand ∗a e he
iso opy elemen s.
Then, San illi p oposes o each an isos uc u e b
Es a ing om he s uc u e E, by conside ing he
ollowing cons uc ion le els:
a) Con en ional le el: (see [2]) I is he ini ial ma hema ical s uc u e, o med by a se o elemen s and
he laws de ined among hem. In his le el appea he usual ma hema ical s uc u es wi h espec o
usual uni s: E=E(a, +,×, ...).
b) Gene al le el: I is he gene al se V, in which a e, pa icula ly, he iso opy elemen s used in
2
Isog oups and isosubg oups
he isop oduc cons uc ion model, ha is, V=V(α, ∗, ?, ...). I is impo an o no e ha E∗=
E(a, ∗, ?, ...)( he es ic ion o V o E) mus e i y he same axioms as he ini ial s uc u e E.
c) Iso opic le el: (see [2]) I is he ma hema ical isos uc u e ob ained when li ing, ha is b
E=
b
E(ba, b
+,b
×, ...).
I is o med by an iso opic se and he isolaws on i . Elemen s o such se , which a e usually deno ed
by using a ha , a e gi en wi h espec o he isouni o b
E. So, ixed and gi en he isos uc u e (b
E, b
×),
wi h isouni b
I, whe e Iis he uni o Ewi h espec o ∗, hese elemen s a e ba=bab
×b
I, whe e San illi
de ines he law b
×as bab
×b
b=d
a∗b. See hen ha , bab
×b
I=d
a∗I=ba=b
Ib
×ba, which implies ha b
Iis
he uni elemen o b
Ewi h espec o b
×.
I is immedia e o check ha he mapping I:E→b
E:a→bais a bijec ion, because i is on o by
cons uc ion and i is also injec i e, due o ba6=b
b, o all a, b ∈Esuch ha a6=b. Indeed, in b
E,
ba=bab
×b
I6=b
bb
×b
I=b
bwi h espec o he isouni b
Io b
E; in he same way as a=a×e6=b×e=b
in E, whe e eis he uni elemen o Ewi h espec o ×.
d) Le el o p ojec ion: (see [6]) I appea s when we conside he ma hema ical isos uc u e b
E e e ed
o he iso opy elemen s used in i s cons uc ion. I s elemen s a e deno ed by a line supe posed o he
ha bo elemen s o b
E, ha is,b.
In his way, i we use he iso opy elemen s ∗(wi h uni I) and b
I o cons uc b
E, hen we ob ain a
s uc u e b
Ein he le el o p ojec ion, whose elemen s a e e e ed o he uni I:ba=a∗b
I= (a∗b
I)∗I.
The mapping π:b
E→b
E:ba→π(ba) = bais named p ojec ion. In gene al, we say ha an elemen o
b
Eis p ojec ed on i s co esponding associa ed elemen belonging o b
E. No e ha , by cons uc ion,
he mapping πis on o.
In a i s s age, b
Eis only do ed wi h laws when πis linea wi h espec o he isolaws associa ed wi h
b
E. So, ixed an isos uc u e (b
E, b
×), i πis linea wi h espec o b
×, he law b
×is de ined on b
Eby
bab
×b
b=bab
×b
b. In such a case, π: ( b
E, b
×)7→ (b
E, b
×)is an on o mo phism .
The e o e, his le el o p ojec ion is he mos impo an in p ac ice, because i allows o ob ain some
ma hema ical models which would be no possible unde usual uni s.
The e exis s s ill ano he le el which joins bo h con en ional and iso opic le els. I is he axioma ic
le el ([2]), which iden i ies e e y ma hema ical s uc u e e i ying he same axioms.
So, in a schema ic way, as he di e en cons uc ion le els appea ing in an iso opic li ing, as he
ela ions among hem can be obse ed in he ollowing diag am:
Con en ional le el Gene al le el
(V, ∗, ?, ...)
−−−−−−−−−−−−−−−−→
(E, +,×, ...) (E, ∗, ?, ...)
↓]↓
Le el o p ojec ion P ojec ion
←−−−−−−−−−−−−−−− Iso opic le el
(b
E, b
+,b
×, ...) ( b
E, b
+,b
×, ...)
Finally, we will say ha an iso opic li ing o he s uc u e Eis injec i e i X=Y, o all X, Y ∈E
such ha b
X=b
Y. I is equi alen , by cons uc ion, o say ha he p ojec ion π:b
E→b
E:ba→π(ba) = bais
3
R. M. Falc´
on and J. N´
u˜
nez
an injec i e mapping. The e o e, as a consequence, i he iso opic li ing o Eis injec i e, hen π:b
E7→ b
E
will be an isomo phism.
3.Isog oups
S a ing om he de ini ion o isog oup (see [2]), we gi e in his sec ion some examples and p ope ies o
hem.
De ini ion 1 Le (G, ◦)be a g oup, wi h an associa i e inne law ◦, and an uni elemen e. An isog oup
b
Gis an iso opy o G, now equipped wi h a new inne associa i e law, b
◦and an uni elemen b
I, such ha he
pai (b
G,b
◦) e i ies he axioms o a g oup. I , besides, bαb
◦b
β=b
βb
◦bαis e i ied o all bα, b
β∈b
G, hen we say
ha b
Gis an isoabelian isog oup o isocommu a i e isog oup.
See ha his de ini ion o isog oup is qui e gene al. So, he uni elemen wi h espec o b
◦, which we
call isouni , is no , in gene al, he men ioned isouni in a San illi’s iso opy. Howe e , when we do ha
cons uc ion, we look a e o make i in his way so ha hese wo elemen s we e he same. Mo eo e , he
ac o w i ing b
Iin he place o be, is no casual. I we ollow he no a ion used jus he e, beis he iso opic
li ing o e, bu , in gene al, beis no he uni elemen o b
◦. I implies ha no ions, p ope ies and heo ems
s udied in he ini ial s uc u e canno be applied in he new s uc u e.
Le see i in he case ha we ha e go a San illi’s iso opy, using a ixed isouni and a ∗law. To do i ,
when we ha e he ini ial g oup (G, ◦), we conside he isouni b
I(which does no belong o Gin gene al),
and we de ine he law ∗which we wan o wo k wi h. I is al eady known he model ha we use o
cons uc he iso opic se b
G, by using he isouni b
Iand he law ∗(which is applied o any s uc u e). So,
b
G={bα=α∗b
I=αb
I|α∈G}.
Now, le see how o li he associa ed law ◦. In he iso opic le el, we de ine he new law as ollows:
bαb
◦b
β=[
α∗β,∀bα, b
β∈b
G. The e o e, in he le el o p ojec ion, we de ine he new law as ollows: bαb
◦b
β=
(α∗β)∗b
I. The new law is called isop oduc .
We can show nex ha i we impose ha (G, ∗)is an associa i e g oup wi h I∈Gas he uni elemen
wi h espec o ∗, hen (b
G,b
◦)is an isog oup. To do i , we will see ha b
◦is an inne law which e i ies he
axioms o a g oup.
Indeed, ∀bα, b
β, bγ∈b
G, we ha e:
a) b
◦is an inne law o b
G, since bαb
◦b
β=[
α∗β∈b
G, because α∗β∈G, due o ∗is an inne law o Gby
hypo hesis ( emembe ha (G, ∗)mus be a g oup).
b) (bαb
◦b
β)b
◦bγ=[
α∗βb
◦bγ=
(α∗β)∗γ=
α∗(β∗γ) = bαb
◦[
β∗γ=bαb
◦(b
βb
◦bγ). (See ha ∗is associa i e is
e y impo an so ha b
◦is associa i e).
c) b
I∈b
G, since I∈G. (See ha b
I=I∗b
I). Mo eo e , b
Iis he isouni ha we sea ch, because
bαb
◦b
I=[
α∗I=bα=b
Ib
◦bα.
d) Le bα∈b
Gbe. I will be α∈G. So, as (G, ∗)is a g oup wi h uni elemen I, he e exis s α−I∈G, such
ha α∗α−I=α−I∗α=I. Then, we mus only ake he elemen d
α−I, as he isoin e se o bαwi h espec
o b
◦,because hen we ha e ha bαb
◦d
α−I=
α∗α−I=b
I=d
α−Ib
◦bα.
e) Finally, i ∗is commu a i e, (b
G,b
◦)will be also commu a i e, because bαb
◦b
β=[
α∗β=[
β∗α=b
βb
◦bα.
4
Isog oups and isosubg oups
So, we ha e p o ed he ollowing:
Theo em 1 Le (G, ◦)be an associa i e g oup and le b
Iand ∗be wo iso opy elemen s. I (G, ∗)is
an associa i e g oup wi h uni elemen I∈G, hen he iso opic li ing (b
G,b
◦)cons uc ed by he model
o he isop oduc , has an isos uc u e o isog oup. Mo eo e , i (G, ∗)is commu a i e, hen (b
G,b
◦)is a
commu a i e isog oup. ¥
Now, we will see some examples o isog oups:
Example 1 Le (R,+) be he g oup o he eal numbe s wi h he usual sum. A i ial iso opic li ing
could be cons uc ed by using he isouni b
I= 0 and he law ∗ ≡ +(no e ha (R,∗)=(R,+) is a
g oup wi h uni elemen 0∈R); so we would ha e he pai (b
R,b
+),whe e b
R={ba=a∗0 = a+ 0 =
a|a∈R}=R.Mo eo e , as ∗ ≡ +, he isop oduc would be de ined as bab
+b
b=d
a∗b=[
a+band
bab
+b
b=[
a+b=a+b. So, we would ha e b
+≡ ∗ ≡ +.
So, he iso opy o (R,+), gi en by he isouni 0and he law ∗ ≡ +, is he same as he i ial iso opy, ha
is, he iden i y. I p o es ha he cons uc ion which we a e using is igh , since i we do no change ei he
he ini ial uni o he ini ial g oup law, hen his g oup emains in a ian when cons uc ing he iso opy. ¥
Example 2 Now, le conside he iso opy o he g oup (R∗,×)(R∗is he eal numbe s se minus he
ze o), ob ained by using he isouni b
I=iand he law ∗ ≡ • ( ha is, by he usual complex law).
So, he iso opic se is c
R∗=Im(C) {0}and he isop oduc will be de ined as bab
×b
b=d
a∗b=d
a•b=
[
a×b o all a, b ∈R.Then, bab
×b
b=bab
×b
b=[
a×b= (a×b)∗i= (a•b)•i o all a, b ∈R.¥
Mo eo e , le see ha he isog oups o he las wo examples a e isocommu a i e. I can be obse ed by
using Theo em 1, because we ha e ha he ini ial g oups a e commu a i e.
No e ha in bo h examples i has been used an isouni ac ing as a cons an . Howe e , examples in which
he isouni used depends on ini ial coo dina es can be also shown:
Example 3 Le conside (R,+), as in Example 1. We conside an iso opic li ing wi h iso opy elemen s
∗ ≡ +and b
I=b
I(x) = ½1, i x= 0
1
x2, i x6= 0 ¾. Then, b
Iis posi i e de ined and non-singula and hus we ob ain
he li ing a→ba=a∗b
I=½0, i a= 0
1
a, i a6= 0 ¾.
Finally, he isop oduc is de ined by bab
+b
b=[
a+b, whe e bab
+b
b=½0, i a+b= 0
1
a+b, i a+b6= 0 ¾in he p ojec-
ion le el. In his way, he se {0}∪{1
a:a∈R∗}can be do ed o an isog oup s uc u e, by he law b
+.
¥
To inish his sec ion we will p o e ha ixed and gi en an a bi a y isog oup, i can be conside ed as an
iso opic li ing which ollows he iso opic cons uc ion model which we a e conside ing. Indeed, we ha e
he ollowing:
5
R. M. Falc´
on and J. N´
u˜
nez
P oposi ion 1 E e y iso opy I: (G, ◦)→(b
G,b
◦)can be conside ed as an iso opic li ing which ollows
he isop oduc cons uc ion model.
PROOF. I is su icien o conside in he gene al le el he se (G, ∗), whe e ∗is de ined by a∗b=I−1(bab
◦b
b)
(i has sense, because he mapping I:E→b
E:a→bais a bijec ion by cons uc ion, as we al eady
obse ed). So, he isolaw b
◦can be de ined la e as bab
◦b
b=I(a∗b) = d
a∗b, and his is he way in which an
isolaw is de ined acco ding o he isop oduc cons uc ion le el, as we al eady saw.
In his way, we also ge , by linea i y, o do e he se (G, ∗)wi h a s uc u e o g oup. Mo eo e , he
uni co esponding o ∗will be, by cons uc ion, he elemen o E, om which he isouni o he associa ed
isolaw b
◦is iso opically li ed. ¥
4.Isosubg oups
In his sec ion we in oduce he de ini ion o isosubg oup. We gi e some examples and p ope ies o hem
and we also deal wi h he di e ences be ween an isosubg oup and a subg oup o an isog oup.
We s udy nex possible li ings o he subg oups, s a ing om he San illi’s cons uc ion model. To do
his, we mus demand ha an isosubg oup is he iso opic li ing o a subg oup Ho a gi en and ixed g oup
G. The di icul y appea s when we equi e ha e e y iso opic li ing o a gi en s uc u e is also a s uc u e
o he same ype. So, e e y iso opy o Hshould ha e s uc u e o subg oup and hus, he iso opic li ings
o Hcould no be independen o he li ing o G. The e o e, he de ini ion o isosubg oup would be as
ollows:
De ini ion 2 Le (G, ◦)be an associa i e g oup and (b
G,b
◦)be an associa ed isog oup. Le Hbe a
subg oup o G. We say ha b
His an isosubg oup o b
Gi , being an iso opy o H, he pai (b
H, b
◦)is a
subg oup o b
G, ha is, i b
H⊆b
G,b
◦is an inne law in b
Hand (b
H, b
◦)has s uc u e o g oup.
Le apply now his de ini ion o he San illi’s cons uc ion model, by an isouni and a law ∗. Suppose
ha we ha e he associa i e g oup (G, ◦)and he isog oup (b
G,b
◦), ob ained by bo h, an isouni b
Iand a law
∗p e iously ixed. Le Hbe a subg oup o G. Since we demand ha in he u u e isosubg oup b
H, he
associa ed law is b
◦, i we go on abou he gi en cons uc ion o he isop oduc , hen he law and he isouni
(bo h which will be he iso opy elemen s), mus be, espec i ely, ∗and b
I, because i no , we would no
ob ain he same law b
◦in gene al. See i wi h an example:
Example 4 Le (Z,+) be he g oup o in ege s, wi h he usual sum. We ake, unde usual no a ions,
∗ ≡ +,b
I= 2.As (Z,∗) = (Z,+) is a g oup wi h uni elemen 0∈Z,we can use he iso opy o elemen s
∗and b
I. Then, b
Z={ba=a∗2 = a+ 2 |a∈Z}=Zand he isop oduc is de ined as bab
+b
b=[
a+b, being
bab
+b
b=[
a+b= (a+b)∗2 = a+b+2, o all a, b ∈Z.In his way, we ha e ob ained he isog oup (b
Z,b
+),
which comes om he addi i e g oup (Z,+).
Le now conside he subg oup (P,+) o e en in ege s and 0. I we cons uc he iso opy ela ed o he
same elemen s as abo e (which is always possible, due o (P,∗) = (P,+) is a g oup wi h uni elemen
0∈P), we ob ain i s ly he iso opic se b
P, being b
P={bm=m∗2 = m+ 2 |m∈P}=P,and we
would ge a e he same isop oduc b
+.
Le now p o e ha (b
P,b
+) is a isosubg oup o (b
Z,b
+), aking in o conside a ion ha b
Pis an iso opy o
P⊆Z. To see i , we obse e
6
Isog oups and isosubg oups
a) b
P⊆b
Z,since P⊆Z.
b) bmb
+bn=
m+n∈b
P, o all m, n ∈P. So, b
+is an inne law on b
P.
c) b
Psa is ies he g oup condi ions, because associa i i y is om (b
G,b
◦), he isouni b
I= 2 = b
0belongs o
b
Pand bm−b
I=d
−m∈b
P,∀m∈P,since bmb
+d
−m=
m+ (−m) = b
0 = b
I=d
−mb
+bm.
So, (b
P,b
+) is an isosubg oup o (b
Z,b
+).¥
No e ha in his example, we can a oid some s eps when cons uc ing b
H, once i is p o ed ha we
can make he iso opy co esponding o elemen s used o cons uc b
G. Indeed, emembe ha by using
Theo em 1, condi ions o be sa is ied a e ha he pai (H, ∗)is a g oup, ha ing he uni elemen I, he same
o (G, ∗). So, i we p oceed simila ly as we did when p o ing ha (b
G,b
◦)had a g oup s uc u e by he
isop oduc b
◦(ob ained s a ing o ∗), we ha e he condi ions needed: b
+is an inne law on b
P, he g oup
axioms a e sa is ied by cons uc ion and inally, (b
H, b
◦)is associa i e due o b
◦is associa i e on (b
G,b
◦), o ∗
being i by hypo hesis.
The e o e, i is p o ed he ollowing:
Theo em 2 Le (G, ◦)be an associa i e g oup and (b
G,b
◦)be he associa ed isog oup co esponding o
he iso opy o elemen s b
Iand ∗. Le Hbe a subg oup o G. I (H, ∗)has s uc u e o subg oup o (G, ∗),
hen he iso opic li ing (b
H, b
◦), co esponding o he iso opy o elemen s b
Iand ∗, is a isosubg oup o b
G.
¥
In ac , as he isog oup cons uc ion model al eady poin ed ou his condi ion, we can say ha i he
iso opy co esponding o b
Iand ∗can be made, hen (b
H, b
◦)is an isosubg oup o b
G. So, he las p oblem
which could appea is ha such an iso opy could no be made o no e i ying some ini ial condi ions. We
will see i in he ollowing example:
Example 5 Le conside he g oup (Z/Z2,+) wi h he usual law +. Le w i e 1= 1 + Z/Z2and
2= 2 + Z/Z2. Conside now b
I=1and he law ∗de ined by
1∗1=1=0∗0
1∗0=0∗1=0.
I is easy o see ha ∗is associa i e, due o:
(1∗1)∗1=1∗1=1∗(1∗1)
(1∗1)∗0=1∗0=1∗(1∗0)
(1∗0)∗0=0∗0=1=1∗1=1∗(0∗0)
(0∗0)∗0=1∗0=0=0∗1=0∗(0∗0)
whe eas o he possible cases a e also sa is ied by commu a i i y. The e o e, (Z/Z2,∗)has s uc u e o
g oup, wi h uni elemen I=b
I=1∈Z/Z2.
Mo eo e , i we make now he iso opy co esponding o b
Iand ∗, we ob ain ha he iso opic se is
Z/Z2,
being
Z/Z2={b
0=0,b
1=1}=Z/Z2.
7
R. M. Falc´
on and J. N´
u˜
nez
Besides, he co esponding isop oduc b
+will be gi en by
b
0b
+b
0=[
0∗0=b
1
b
0b
+b
1=[
0∗1=b
0=b
1b
+b
0
b
1b
+b
1=[
1∗1=b
1
So, we ha e:
0b
+0=b
0b
+b
0=b
1=1
0b
+1=b
0b
+b
1=b
0=0=1b
+0
1b
+1=b
1b
+b
1=b
1=1
Thus, b
+≡ ∗ and so, (
Z/Z2,b
+) is a new isog oup, being (
Z/Z2,b
+) = (Z/Z2,∗).
Le conside sepa a ely he subg oup ({0},+) o (Z/Z2,+). We obse e ha ({0},∗)does no ha e
a s uc u e o g oup, because ∗is no an inne law on {0},due o 0∗0=1/∈ {0}.So, condi ions o
Theo em 2 o ob ain an isosubg oup by making he iso opy o elemen b
I=1and ∗, a e no sa is ied. We
would ha e, in ac , ha d
{0}={b
0}, whe e b
0b
+b
0=b
1/∈d
{0}.¥
We a e going o ask ou sel es a new ques ion. Suppose ha we ha e an associa i e g oup (G, ◦), wi h
uni elemen I, and le (b
G,b
◦)be he isog oup associa ed o he iso opy o elemen s b
Iand ∗. We know ha
e e y isosubg oup o b
Ghas s uc u e o subg oup. We also know some examples in which subg oups o
Gdo no gi e ise o isosubg oup o b
G, by using he same b
Iand ∗like iso opy elemen s. Then, we can
inally ask i e e y subg oup o b
Ghas a s uc u e o isosubg oup o (b
G,b
◦), ha is, i i comes om he
iso opic li ing o a subg oup o G. O cou se, we ha e al eady no ed ha i we ixe a subg oup (b
H, b
◦)o
(b
G,b
◦), hen he law b
◦has o be he same in bo h pai s, so he co esponding elemen s o iso opic ha e o
coincide. Tha is, i b
His an isosubg oup, hen i ha e o come om an iso opy ha ing he same elemen s
as he ones used in he cons uc ion o b
G. So, he only possible subse o Gwhich would gi e ise o he
possible isosubg oup would be H={a∈G:ba∈b
H} ⊆ G. Howe e , he pai (H, ◦)is no a subg oup o
(G, ◦)in gene al, as we can check in he ollowing:
Example 6 Le conside bo h he g oup (Z/Z2,+) and he isog oup (
Z/Z2,b
+) men ioned in he las
example. We ha e he pai s ({0},+) and ({b
1},b
+) espec i ely, as he only p ope subg oups o bo h.
As we ha e jus seen, i we ake he subg oup b
H= ({b
1},b
+) o (
Z/Z2,b
+), he only possible subse o
Z/Z2 om which we could gi e a s uc u e o isog oup o b
Hwould be H={1},because 1∗1=1,whe e
b
I=1is he isouni used in ha example o cons uc he iso opy. Howe e , ({1},+) is no a subg oup o
(Z/Z2,+), because, o ins ance, +is no an inne law on {1}, due o 1+1=0.¥
So, wi h his example, he las ques ion abo e men ioned is answe ed in he nega i e. In his way, he
link be ween g oups and isog oups is inally sol ed.
8
Isog oups and isosubg oups
5.Iso ep esen a ion o ini e isog oups
In his sec ion we y o ha e some ela ions be ween San illi’s iso heo y and he s anda d heo y o ep e-
sen a ion o g oups.
Le us ecall ha a ep esen a ion o a ini e g oup Gis a pai (V, ρ), whe e Vis a ec o k-space and ρ
is a g oup homomo phism ρ:G→Gl(V).
To gene alize his concep on bo h iso opic and p ojec ion le els we i s ly gi e he de ini ion o isoho-
momo phism o isog oups and secondly, o iso ep esen a ion o ini e isog oups.
De ini ion 3 Le (G, ◦)and (G0,•)be wo g oups and le (b
G,b
◦)and (c
G0,b
•)be associa ed isog oups,
espec i ely. An isohomomo phism de ined be ween b
Gand c
G0is he iso opic li ing o any mapping ρ:
G→G0, ha is, bρ:b
G→c
G0:bg→bρ(bg) = d
ρ(g), e i ying bρ(bgb
◦b
h) = bρ(bg)b
•bρ(b
h), o all Lbg,b
h∈b
G.
No e ha by demanding he compa ibili y o he li ing used, we ob ain ha ρis a g oups homomo phism
wi h espec o he laws o Gand G0. To see i we i s ly in oduce he ollowing esul , which is easy o
p o e:
P oposi ion 2 Le (G, ◦)be a g oup and (b
G,b
◦)an associa ed isog oup. I b
Ghas been ob ained s a ing
om an iso opy compa ible wi h espec o ◦( ha is, bgb
◦b
h=[
g◦h, o all g, h ∈G), hen Gand b
Ga e
isomo phic g oups. ¤
Mo eo e , i we conside he isop oduc cons uc ion model (which is always possible), we deduce he
ollowing:
Co olla y 1 Unde he hypo hesis o P oposi ion 2, i he iso opy used ollows he isop oduc cons uc ion
model, hen we ha e, in he gene al le el o he g oup (G, ∗), ha (G, ∗)≡(G, ◦).
PROOF.
I is immedia e by cons uc ion, because ixed and gi en a, b ∈G, we ha e d
a∗b=bab
◦b
b=d
a◦b. So,
a∗b=a◦b.¥
I is now possible o p o e he ollowing:
P oposi ion 3 Unde condi ions o De ini ion 3, i he iso opy used o cons uc b
Gand c
G0is compa ible
wi h espec o ◦and •, espec i ely, hen bρis an isohomomo phism be ween b
Gand c
G0i and only i ρis an
homomo phism be ween Gand G0.
PROOF.
a) ⇒
Le us suppose ha bρ:b
G7→ c
G0is an isohomomo phism.
Then, ixed g, h ∈G, we ha e
ρ(g◦h) = bρ([
g◦h) = bρ(bgb
◦b
h) = bρ(bg)b
•bρ(b
h) =
ρ(g)•ρ(h), and hus,
ρ(g◦h) = ρ(g)•ρ(h), which implies ha ρis an homomo phism, due o gand ha e a bi a y in G.
b) ⇐
Le us suppose ha ρ:G7→ G0is an homomo phism. Fixed gand hin G, we ha e ha bρ(bgb
◦b
h) =
bρ([
g◦h) =
ρ(g◦h) =
ρ(g)•ρ(h) = bρ(bg)b
•bρ(b
h). So, bρis an isohomomo phism. ¥
9