Thus the associative law and the two distributive laws*^ carry over to the
integers:
/(i)(^(i) h(i)) = fmig o hm) = (fo(go h))(i) = ((/o g) o h)(i)
= i(fogXl))h(l) = if(l)g(l))h(l),
(/(I) + g(l))h(l) = ((/+ ^)A))^A) = ((/+ g) o h)(l)
= (foh + goh)(l)
= (/o h)(l) + (^ o h)(l) = /(I) h(l) + ^A) h(l).
Thus the integers form a ring with respect to addition and multiplication.
Since by C9) the endomorphism / corresponds to the number 1, this
number 1 is the unit element of the ring of integers:
l/(l) = /(l)/(l) = (/o/)(l)=/(l).
/AI =/(l)/(l) = (/o/)(l)=/(l).
By an isomorphism we mean a one-to-one homomorphism with respect
to the operations in question (for a ring, addition and multiplication).
Since C9) was shown to be one-to-one, we have the theorem: the mapping
C9) is an isomorphism of the ring of endomorphisms of the module of
integers onto the ring of integers.
In order to prove the commutativity of multiplication, we must note
that by the second distributive law x ^^ xa is an endomorphism: for in
fact, the image (x + y)a = xa -\- ya of x ^ y is the sum of the images
of X and>'. Application of D0) to this endomorphism gives xa = (la)x = ax,
so that the ring of integers is a commutative ring.
In an arbitrary ring (for which we denote multiplication in the same
way as for the numbers) complete induction on n enables us to generalize
the distributive laws to
nn
(E «^) ^ = z «e^' «z ^i = i^bi.
1=1 i=l 2=1 i=l
from which we have
mnm,n
Z «i Z ^fc = Z («i Z ^fc) ^ Z Z ^i^k'
■ =1 fc=l i=l fc=l i=l fc=l
** Because of the commutative law (to be proved later) only the second of the two
distributive laws needs to be proved for the integers. In any case, the first distributive
law is a simple consequence of C0) and D0).
118
and thus by B1)
D1)
PART 6
PART 6 ARITHMETIC AND ALGEBRA
By complete induction we readily obtain*^ from D1)
TO Ui
D1') n z «e-^ = z n «e;^.;
2=1 fc=i l^k^fS,n^ j=l
(i=l....,TO)
where the index set for the summation on the right-hand side is the set of
m-tuples (^1, ..., km), with I ^ ki ^ ni(i = I, ..., m).
From the distributive law we further have
a(c — d)-\- ad = a((c — d) -\- d) = ac,
(a — b)c -\- be = ((a — ^) + b)c = ac
and thus,
D2) a(c — d) = ac — ad,
D2') (a — b)c = ac — be.
Replacing a by a — ^ in D2), we see from D2') that
(a - b)(c -d) = {a- b)c -{a- b)d = ac - be + {-{ad - bd)),
and therefore by C6) and C5)
D3) (a - b){c -d) = ac -bc^bd-ad = {ac + bd) - {ad + be).
Setting c = d and a = ^ in D2) and D2') respectively, we obtain
aO = 0 = Oc, so that D2), D2'), D3) give
D3') a{-d) = -ad, {-b)c = -be, {-b){-d) = bd.
In particular, we have —a = {—l)a = a{—l) if the ring has a unit
element, which for simplicity we have here denoted by 1.
Since the product of two natural numbers is always a natural number,
the equations D3'), when applied to the ring of integers, show that the
product of a positive with a negative number is a negative number and
that the product of two negative numbers is a positive number. But an