Lecture 23

Modal Logic

W e hold there to be two different senses of truth: those sentences

that happen to be true and those sentences that must be true.
Our logic so far does not distinguish between garden-variety truth
and necessary truth. Can we develop a logic that does account for
the difference between these kinds of true sentences? This is what we
call modal logic. “Modality” is the term that philosophers use for the
concepts “possible” and “necessary.” Modal logic is an augmented
version of our first-order relational logical language that includes new
operators for possibility and necessity.

Modality
⊲⊲ We have been using a particular word repeatedly in these lectures
as if it is clear what it means. The word is “true.” What does it
mean for a sentence to be true? Philosophers have debated this
for centuries.

⊲⊲ One standard definition, what we call the correspondence theory

of truth, holds that a sentence is true if what it says about the
universe is actually the case in the universe.

⊲⊲ Philosophers have long realized that there are two different

types of true sentence. One type of true sentence happens to
be true, but the world could be otherwise. These contingent
truths are possibilities.

⊲⊲ But then there are sentences we think must be true no matter

how the world is. We call these necessary truths. We have
thought that these necessary truths are sprinkled throughout
 human endeavors—in such areas as mathematics, science,
ethics, and metaphysics.

⊲⊲ The concept of necessary truth is stronger than mere truth or

that which happens to be true. The term “modality” is used by
philosophers to refer to the concepts of necessity and possibility.

⊲⊲ The logic we have been exploring is only geared to a single

concept of truth. If we want to expand it to be able to look at
arguments involving possible and necessary truths, then we
need to create an augmented system. We need modal logic.

⊲⊲ We have seen some necessary truths already. Think back to truth-

functional logic. We saw that there are tautologies, sentences like
“I have a brother or I don’t have a brother” that are always true.
We can extend this to our first-order predicate language and
come up with necessary truths like “All dogs are dogs.”

⊲⊲ There are sentences like these that are necessarily true. These
sentences are necessarily true because of their form. Any
sentence of the form av−a or of the form ∀x(Dx → Dx) will be
necessary truths.

⊲⊲ There are some examples from mathematics, science, ethics, and

religion that are supposedly necessary truths because of their
content, not because of their form. This means that we will need
some additional machinery to deal with the logic of modal concepts.

⊲⊲ We need to add two new operators to our first-order predicate

language. The box (o) will be our symbol for “necessary,” and
the diamond (u) will be our symbol for “possible.” They will
function syntactically sort of like quantifiers in that they will be put
in front of the sentences that they are intended to modify.

⊲⊲ We might say that the sentence “Everything is F” is true—that

is, we can assert ∀xFx. But if we want to make a stronger claim

218 An Introduction to Formal Logic

 and say that it is no accident of the universe that everything is F,
but rather that it could not be otherwise—that everything must be
F—then we can write o ∀xFx. This says that the sentence ∀xFx is
a special sentence; it is a necessary truth.

⊲⊲ In the same way, we could add a special notation to indicate that

we don’t know if ∀xFx is true or not, but it is certainly possible that
all things have the property F. In this case, what we are asserting
is u∀xFx. This says that the sentence ∀xFx is like a contingency
in truth-functional logic. It is a sentence that could be true.

⊲⊲ There is a logical relation between these two. In the same way

that we have the equivalence quantifier negation that allows us to
switch the order of a negation and a quantifier—for example, “not
everything” is the same as “something is not”—the same holds
true with our modal operators.

⊲⊲ We can employ the equivalence modal negation (MN) and switch

the order of a modal operator and a negation. If we say that it is
not the case that “Everything is G” is necessary, then we have the
sentence −o ∀xGx. But to say that something is not necessarily
true is to say that it is possible that it is false. But this is to say that
u−∀xGx.

⊲⊲ Similarly, if we say that it is not possible that everything is F,

then we are asserting the sentence −u∀xGx. But to say that it is
not possible for it to be true, we are saying that it is necessarily
false—that is, we are asserting the sentence o−∀xGx. If we move
the negation to the other side of the modal operator and switch
the operator, we end up with an entirely equivalent proposition.

⊲⊲ One way that philosophers make sense of these new operators

is to give them an interpretation in terms of possible worlds.
By the term “possible world,” we mean a way the world could
be. Our world, the real world, is just one among the infinite
possible worlds.

Lecture 23—Modal Logic 219

⊲⊲ When we write a diamond in front of a sentence, we are saying
that the sentence is possibly true. By that, we mean that there
is a possible world (or some set of possible worlds) in which
this sentence is true. When we then use that sentence in valid
arguments, we are seeing what else must also be true in this
particular possible world.

⊲⊲ When we put a box in front of a sentence, we are saying that the

sentence is necessarily true. A sentence is necessarily true if it is
true in every possible world. No matter what world you consider,
the necessary truths will be true in them. If we put a box in front of
a negated sentence—that is, we say it is necessarily false—that
means that it is true in no possible world.

The Mechanics of Modal Logic

⊲⊲ The translation of spoken-language modal sentences into our
expanded first-order modal logical language is not different from
our regular translation in our first-order language, except for the
new machinery.

⊲⊲ With all of the other additions we have made, we first learned to

translate and then jumped right into how to use them in proofs.
What are the new rules and equivalences associated with the
introduced elements? With modality, however, this is more
complicated because there is not single logical meaning for
them. Rather, there are several possible meanings.

⊲⊲ There are multiple viewpoints for modal logic. They are called
modal systems of different strengths because each stronger level
commits us to a new rule, where the weaker ones do not. We
account for different meanings of modality by having different
modal languages that use different additional axioms.

220 An Introduction to Formal Logic

 ¹¹ K (named for the philosopher-logician Saul Kripke) is the
weakest modal language, according to which a sentence is
necessary only if we can prove it using no premises. In this
language, the axiom o(P → Q) → (oP → oQ) is true and can be
inserted into a proof for any sentences P and Q.
¹¹ M or T (named independently by Kurt Gödel and Georg Henrik
von Wright), the next stronger language, adds the claim that if
a sentence is necessarily true, then it must also be true—that
is, oP → P.
¹¹ S3 (where “S” stands for “strength”) adds the additional claim
to M or T that if a sentence is necessarily true, then it must also
be possibly true—that is, oP → uP.
¹¹ S4 adds to S3 the additional claim that if a sentence is
necessarily true, then it is necessarily true that it is necessarily
true —that is, oP → ooP.
¹¹ S5, the strongest modal language, adds to S4 the claim that all
modal claims are necessary—that is, not only are necessary
truths necessarily necessary, but that sentences that are
possible are necessarily possible. In other words, uP → ouP.

⊲⊲ With modal logic, we are left with the question that mathematicians
faced when suddenly they had multiple geometric systems: Which
one is right? Which of these five modal systems is the real one?
The stance that logicians take is similar to the one David Hilbert
took with respect to mathematical systems—take whichever one
you want and then ask the philosophers questions.

Deontic Logic
⊲⊲ Philosophers have made another interesting use of modal logic,
turning it into what is called deontic logic. The term “deontology”
refers to an approach to logic that is based on duties, absolute
rules that must be followed in order to act morally.

Lecture 23—Modal Logic 221

⊲⊲ Moral theory utilizes three different statuses for possible acts.

¹¹ An action could be morally necessary, meaning that you have

to do it.
¹¹ An action could be morally prohibited; these are things that
you should never do.
¹¹ An action could be morally permissible, meaning that you
could do it if you want or don’t do it if you don’t want—either
way is fine.

⊲⊲ Deontological ethicists thought that perhaps modal logic could

be modified to become deontic logic. Morally necessary
actions would be represented with the box (o). Instead of oP
representing the necessary truth of some sentence, it would
represent the necessity of a person’s undertaking the act P.
Morally impermissible acts would be represented as o−P. This
means that it is necessary to avoid doing P. The diamond ()
would be used for morally permissible acts.

⊲⊲ The hope was that in adapting modal logic to a moral logic,

we could take the mushiness and hurt feelings that often
accompany arguments surrounding questions of ethics and
replace these feelings with the rigor and objectivity that we see
in logical argumentation.

⊲⊲ Unfortunately, the project has not taken off as much as early

advocates had hoped. A few snags have surfaced. Unlike with
the logic of truth in which every sentence is either true or false
and we never have to choose between truths, there can be
situations in which we have to choose between morally required
acts. Additionally, are ethical sentences intrinsically necessary or
only necessary in a context?

222 An Introduction to Formal Logic

Readings

Epstein, Propositional Logics, chap. 6.

Layman, The Power of Logic, chap. 12.

Questions

1.
Is S5 too strong? Is everything necessarily the way it has to be?
Could things be otherwise? Are the alternative possible worlds really
possible? Are things the way they are because that is the way they
have to be?

2.
Some philosophers have argued that to say that a sentence is true is
to say that it describes reality. For example, to say that the sentence
“Madagascar is an island” is true, it must be the case in reality that
Madagascar is, indeed, an island. So, if we hold that there are true
sentences about possibility—for example, if we think it is true that
“It is possible that Madagascar wouldn’t have been an island”—then
there is a possible world in which Madagascar is not an island. But if
to be true is to describe an aspect of reality, then if we hold sentences
about possibility to be true, then the possible worlds must be part of
reality—in other words, the possible worlds must all actually exist. Do
we want to give up on having true sentences about possibility? Do
we want to allow for reality to include not only our world but every
possible world, or is there another way to understand the truth of
sentences about possibility?

3.
Is a deontic logic possible? Could there be a logic that helps us
determine what we should do if we want to act ethically?

Lecture 23—Modal Logic 223