6

Quantifiers III:
Specialization

The previous two chapters illustrate how to proceed when a quantifier appears
in the statement B. A method is introduced in this chapter for working
forward from a statement that contains the universal quantifier “for all.”

6.1 HOW TO USE SPECIALIZATION

When the statement A contains the quantifier “for all” in the standard form:
A: For all “objects” with a “certain property,”
“something happens,”
one typical method emerges for working forward from A—specialization.
In general terms, specialization works as follows. As a result of assuming A
is true, you know that, for all objects with the certain property, something
happens. If, at some point, you were to come across one of these objects that
does have the certain property, then you can use the information in A by being
able to conclude that, for this particular object, the something does indeed
happen. That fact should help you to conclude that B is true. In other words,
you will have specialized the statement A to one particular object having the
certain property.
To illustrate the idea of specialization in a more tangible way, suppose you
know that

A: All cars with 4 cylinders get good gas mileage.

69
70 CHAPTER 6: QUANTIFIERS III: SPECIALIZATION

In this statement, you can identify the following three items:

Objects: cars.
Certain property: with 4 cylinders.
Something happens: get good gas mileage.
Suppose that you are interested in buying a car that gets good gas mileage,
so your objective is

B: To buy a car that gets good gas mileage.

You can work forward from the information in A by specialization to establish
B as follows. Suppose you are in a dealer’s lot one day and you see a particular
car that you like. Talking with the salesperson, suppose you verify that the
car has 4 cylinders. Recalling that the foregoing statement A is assumed to
be true, you can use this information to conclude that

A1: This particular car gets good gas mileage.

In other words, you have specialized the for-all statement in A to one particular
object with the certain property.
If you analyze this example in detail, you can identify the following steps
associated with applying specialization to a forward statement of the form:
A: For all “objects” with a “certain property,”
“something happens.”

Steps for Using Specialization

1. Identify, in the for-all statement, the object and its type, the certain
property, and the something that happens.
2. Look for one particular object with the certain property that you can
apply specialization to.
3. Conclude, by writing a new statement in the forward process, that the
something happens for this particular object.
The following example demonstrates the proper use of specialization. To that
end, suppose you know that
√
A: For all real numbers x, y ≥ 0, x + y ≥ 2 xy.

In the foregoing statement, you can identify the objects (real numbers x
and y), the certain property (being ≥ 0), and the something that happens
√
(x + y ≥ 2 xy). You can therefore specialize this statement to any two
real numbers that are ≥ 0 (that is, objects with the certain property). For
example, the result of specializing A to x = 3 and y = 27 is
p
A1: 3 + 27 ≥ 2 3(27), or equivalently, 30 ≥ 18.
 6.1 HOW TO USE SPECIALIZATION 71

Alternatively, for given real numbers a and b, the result of specializing A

to x = a2 ≥ 0 and y = b2 ≥ 0 is
√
A2: a2 + b2 ≥ 2 ab.

Specialization is now used in a complete proof.

Definition 14 A real number u is an upper bound for a set of real numbers

T if and only if for all elements t ∈ T , t ≤ u.

Proposition 8 If R is a subset of a set S of real numbers and u is an upper

bound for S, then u is an upper bound for R.

Analysis of Proof. The forward-backward method gives rise to the key

question, “How can I show that a real number (namely, u) is an upper bound
for a set of real numbers (namely, R)?” Definition 14 is used to answer the
question. Thus it must be shown that

B1: For all elements r ∈ R, r ≤ u.

The appearance of the quantifier “for all” in the backward process suggests
proceeding with the choose method, whereby one chooses

A1: An element, say r, in R,

for which it must be shown that

B2: r ≤ u.
(Here, the symbol r is used both for the chosen object in A1 and the general
object in the for-all statement in B1, though they have different meanings.)
Turning now to the forward process, you will see how specialization is used
to reach the conclusion that r ≤ u in B2. From the hypothesis that R is a
subset of S, and by Definition 12 on page 55, you know that

A2: For each element x ∈ R, x ∈ S.

Recognizing the keywords “for each” in the forward process, you should con-
sider using specialization. According to the discussion preceding Proposition
8, the first step in doing so is to identify, in A2, the object with its type
(element x ∈ R), the certain property (there is none) and the something that
happens (x ∈ S). Next, you must look for one particular object with the
certain property with which to specialize. Recall call that, as a result of the
backward process, you chose the particular element r ∈ R (see A1). The
final step of specialization is to conclude, by writing a new statement in the
forward process, that the something in A2 happens for the particular object
r ∈ R. In this case, specialization of A2 allows you to conclude that

A3: r ∈ S.
72 CHAPTER 6: QUANTIFIERS III: SPECIALIZATION

The proof is not yet complete because the last statement in the backward
process (B2) has not yet been reached in the forward process. To do so,
continue to work forward. For example, from the hypothesis you know that
u is an upper bound for S. By Definition 14 this means that

A4: For every element s ∈ S, s ≤ u.

Once again, the appearance of the quantifier “for every” in the forward process
suggests using specialization. Accordingly, identify, in A4, the object with its
type (element s ∈ S), the certain property (there is none), and the something
that happens (s ≤ u). Now look for one particular object with the certain
property with which to apply specialization. The same element r chosen in
A1 serves the purpose noting, from A3, that r ∈ S. Specialization then allows
you to conclude that, for this particular object r ∈ S, the something in A4
happens, so

A5: r ≤ u.
The proof is now complete because A5 is the last statement obtained in the
backward process (see B2).
In the condensed proof that follows, note the lack of reference to the
forward-backward, choose, and specialization methods.

Proof of Proposition 8. To show that u is an upper bound for R, let

r ∈ R. (The word “let” here indicates that the choose method is used.) By
hypothesis, R ⊆ S and so r ∈ S. (Here is where specialization is used.)
Furthermore, by hypothesis, u is an upper bound for S, thus, every element
in S is ≤ u. In particular, r ∈ S, so r ≤ u. (Again specialization is used.)

When using specialization, be careful to keep your notation and symbols

in order. Doing so involves a correct “matching up of notation,” similar to
what you learned in Chapter 3 when using definitions. To illustrate, suppose
you are going to apply specialization to a statement of the form:
A: For all objects X with a certain property,
something happens.
When looking for a particular object, say Y , with which to specialize, it is
necessary to verify that Y satisfies the certain property in A. To do so, replace
X with Y everywhere in the certain property in A and see if the resulting
condition is true. Similarly, when concluding that the particular object Y
satisfies the something that happens in A, again replace X everywhere with Y
in the something that happens to obtain the correct statement in the forward
process. (This is done when writing statements A3 and A5 in the foregoing
analysis of the proof of Proposition 8.) Be careful of overlapping notation,
for example, when the particular object you have identified has precisely the
same symbol as the one in the for-all statement you are specializing.
 6.2 READING A PROOF 73

6.2 READING A PROOF

The process of reading and understanding a proof that uses specialization is

demonstrated with the following proposition.

Definition 15 A real number u is a least upper bound for a set S of real

numbers if and only if (1) u is an upper bound for S and (2) for every upper
bound v for S, u ≤ v.

Proposition 9 If v∗ and w ∗ are least upper bounds for a set T , then v∗ = w ∗ .

Proof of Proposition 9. (For reference purposes, each sentence of the

proof is written on a separate line.)
S1: From the hypothesis, both v∗ and w ∗ are upper bounds for T .
S2: Because v∗ is a least upper bound for T , v∗ ≤ u, for any upper
bound u for T .
S3: In particular, w ∗ is an upper bound for T , so v∗ ≤ w ∗ .
S4: Similarly, w ∗ is a least upper bound for T and, because v∗ is
an upper bound for T , w ∗ ≤ v∗ .
S5: It now follows that v∗ = w ∗ .
The proof is now complete.

Analysis of Proof. An interpretation of statements S1 through S5 follows.

Interpretation of S1: From the hypothesis, both v∗ and w ∗ are upper bounds
for T .
The author is working forward from the hypothesis using part (1) of the
definition of a least upper bound to claim that

A1: v∗ and w ∗ are upper bounds for T .

When reading a proof, it is advisable to determine where the author is

heading. To do so, work backward from B yourself. In this case, you are led
to the key question, “How can I show that two real numbers (namely, v∗ and
w ∗ ) are equal?” Read forward in the proof to see how the author answers this
question. From S3 and S4, the answer in this case is to show that

B1: v∗ ≤ w ∗ and w ∗ ≤ v∗ .

Interpretation of S2: Because v∗ is a least upper bound for T , v∗ ≤ u, for

any upper bound u for T .
The author is continuing to work forward by stating part (2) of the defini-
tion of a least upper bound applied to v∗ ; that is,

A2: For every upper bound u for T , v∗ ≤ u.

Interpretation of S3: In particular, w ∗ is an upper bound for T , so v∗ ≤ w ∗ .

74 CHAPTER 6: QUANTIFIERS III: SPECIALIZATION

It is here that specialization is applied to the for-all statement in A2, indi-

cated by the words “in particular.” Specifically, A2 is specialized to the value
u = w ∗ , which is an upper bound for T (see A1). The result of specialization,
as the author claims is S3, is

A3: v∗ ≤ w ∗ .

Interpretation of S4: Similarly, w ∗ is a least upper bound for T and, because

v∗ is an upper bound for T , w ∗ ≤ v∗ .
The author is using the same analysis as in S3 but, this time, applied to
the least upper bound w ∗ and the upper bound v∗ for T . The result of this
specialization is that

A4: w ∗ ≤ v∗ .

Interpretation of S5: It now follows that v∗ = w ∗ , and so the proof is

complete.
The author is working forward by combining v∗ ≤ w ∗ from A3 and w ∗ ≤ v∗
from A4 to claim correctly that v∗ = w ∗ . Finally, the author states that the
proof is complete, which is true because the conclusion B has been established.

Summary

You now have various techniques for dealing with quantifiers that can appear
in either A or B. As always, let the form of the statement guide you. When B
contains the quantifier “there is,” the construction method is used to produce
the desired object. The choose method is associated with the quantifier “for
all” in the backward process. Finally, if the quantifier “for all” appears in the
forward process, use specialization. To do so, follow these steps:
1. Identify, in the for-all statement, the object with its type, the certain
property, and the something that happens.
2. Look for one particular object with the certain property that you can
apply specialization to. (This object often arises as a result of the back-
ward process, especially when the choose method is used.)
3. Conclude, by writing a new statement in the forward process, that the
something happens for this one particular object.
It is common to confuse the choose method with the specialization method.
Use the choose method when you encounter the keywords “for all” in the
backward process; use specialization when the keywords “for all” arise in the
forward process. Another way to say this is to use the choose method when
you want to show that “for all objects with a certain property, something
happens”; use specialization when you know that “for all objects with a certain
property, something happens.”