1919.] CONTINUOUS MATHEMATICAL INDUCTION.
17
A NOTE ON " CONTINUOUS MATHEMATICAL
INDUCTION."
BY DR. YUEN REN CHAO.
(Read before the San Francisco Section of the American Mathematical
Society April 5, 1919.)
1. Special case.—Let the function f(x) be defined in some
interval of a real variable x.
Hyp. 1. Let there be a point a in the interval such that
f (a) = 0.
Hyp. 2. Let there be a constant A for the interval, such
that f(x) = 0 implies f(x + S) = 0, whenever 0 < S S A.
Then for any b in the interval, where b > a, ƒ(&) = 0.
Proof.—I. If b — a S A, then by Hyp. 2 the conclusion
follows.
II. If b — a > A, then first apply Archimedes' postulate,
that is, there will be an integer n and a fraction 0(0 S 0 = 1)
such that
b — a = (n+ 0) A, or b = (a + $A) + nA.
Next, apply ordinary mathematical induction, thus: By
Hyp. 1 and 2, since 0A < A,
/ . / ( a + 0 A ) = O.
Therefore, by 2, again,
(1) / [ ( a + 0 A ) + l.A] = O.
By 2, if ƒ [(a + SA) + m-A] = 0, then
(2) /[(a + 0A) + ( m + l ) A ] = O.
Hence, combining (1) and (2),
ƒ (a +6A + nA) = 0,
that is,
m = o.
2. General case.—Let <p(x) be any propositional function,
defined in some interval of a real variable x.
18 CONTINUOUS MATHEMATICAL INDUCTION. [Oct.,
Hyp. 1. Let there be a point a in the interval such that
<p(a) is true.
Hyp. 2. Let there be a constant A for the interval such
that <p(x) implies <p(x ± 5), whenever 0 < S ^ A.
Then for any b in the interval such that b < a, respectively,
<p(b) is true.
The proof will be the same as for the special case, except
for obvious changes of wording or sign.
Remarks.—The theorem rests essentially on Archimedes'
postulate and on ordinary mathematical induction, but it is
not a generalization of the latter, in the sense of including it
as a special case. It is not a theorem in mathematical logic,
since it is concerned with a real variable x. But it is more
general than ordinary theorems dealing with equalities, in that
<p(x) may be a statement about continuity, convergence, inte-
grability, etc., that cannot be put in the simple form of
f(x) = 0.
The theorem is a mathematical formulation of the familiar
argument from "the thin end of the wedge," or again, the
argument from " the camel's nose ":
Hyp. 1. Let it be granted that the drinking of half a glass
of beer be allowable.
Hyp. 2. If any quantity, x, of beer is allowable, there is
no reason why x + S is not allowable, so long as 5 does not
exceed an imperceptible amount A.
Therefore any quantity is allowable.
Like all mathematical theorems, the conclusion is no surer
than its hypothesis. In this case, if the argument fails, it is
usually because a constant A required in the second hypothesis
does not exist. Take the very wedge itself. If a wedge is
driven with a constant force between two sides which are
pushed together by elastic forces, it will be stopped when
balanced by the component of the increasing resistance. In
this case the A within which S may increase for <p(x + 8) to
continue to hold will not be " uniform for the interval,"
so to speak, but will become smaller and smaller as x ap-
proaches the dangerous point, beyond which the conclusion
ceases to be true.
UNIVERSITY OF CALIFORNIA.
