Universal Statements and Counterexamples
Universal Statements and Counterexamples
Notice in this context that the …rst statement has the potential to be confusing.
Did the author mean that every real number has a nonnegative square, or did
the author mean that there is some real number with a nonnegative square?
Because of this potential for confusion, you should be careful to include univer-
sal quanti…er words when writing universal statements. Also because of this
potential for confusion, it is important that you carefully read mathematical
statements to see if there are implicit universal quanti…ers.
Some authors use "any" as a universal quanti…er, writing "The square of any
real number is nonnegative." Unfortunately, this con‡icts with a di¤erent use
of "any" as in "Are there any solutions to cos(x) = 1?," the answer to which is,
"Yes, x = 0 is such a solution." As a consequence of the ambiguities associated
with "any", you should not use "any" as a universal quanti…er.
Consider a few more examples of true universal statements that convey the
same information:
Why is the last statement universal? Because "always" contains the uni-
versal quanti…er "all".
Compare the two statements, "The square of a real number is nonnegative,"
and "The square of a real number equals the number itself." The …rst is a
universal statement with an implied "every". What about the second state-
ment? It actually is a restricted statement, and should be quanti…ed as follows:
1
"The square of some real number equals the number itself." In this second
formulation, it is clearer that the author does not mean that every real number
equals its square. Notice that the restricted statement is clari…ed by using the
restricted quanti…er "some".
How do we assess the truth of a universal statement? Either we show
that it is always true (a proof ), or we show that it is false in at least
one instance. Such an instance is a counterexample. A counterexample
should be as explicit and particular as possible. Consider the statement: "If
a function has a zero derivative at x = a, then it must have a local minimum
or maximum at x = a:" This is actually a universal statement since it can be
rewritten as: "Every function that has a zero derivative at x = a has a local
minimum or maximum at x = a." We might try to prove that this is true
for all functions that have zero derivatives at x = a, or we might try to show
that the statement is false. In fact, the statement is false. To establish
its falsity, we only have to …nd one example of a function that has a zero
derivative at x = a, but which does not have a maximum or minimum at x = a:
Further, we want to be as speci…c about such a function as possible so that it is
immediately obvious that we have a function that disobeys the claimed fact. A
good counterexample here would be f (x) = (x a)3 since f 0 (a) = 0 but x = a
is neither a local minimum nor a local maximum. A less good counterexample
would be f (x) = (x a)n where n is an odd integer with n 3. A poor
counterexample would be f (x) = a3 b2 x5 2a3 bx6 a3 x7 + 3a2 b2 x6 + 6a2
bx7 + 3a2 x8 3ab2 x7 6abx8 3ax9 + b2 x8 + 2bx9 + x10 where b is some …xed
real number.
Consider the universal statement, "For all x > 7; x2 100." It does not
matter that we can …nd lots of choices of x that satisfy both x > 7 and x2 100:
What matters is that we can …nd at least one x for which x > 7 and x2 > 100:
For the given statement, x = 11 is a great counterexample, and x = 1; 000; 000
is a good counterexample. While x = e is a valid counterexample, it is a poor
counterexample because it is not obvious that e > 7 and ( e )2 > 100:
In summary,