You are on page 1of 8

Matthews 1

MA 385
Assignment 11
Caden Matthews
Section 15
Part E
19.
We first consider the base case when 𝑛 = 0. This means that φ does not have any binary
connectives. Since there are no binary connectives, φ can only contain one sentential
variable (we will not discuss the existence of ¬ in the formula, as they are not of any
concern as we care only about sentential variables and binary connectives). From this, we
note that 𝑚 = 1. It is the case that 1 = 0 + 1, so the property holds for the base case.

We now move to the inductive step. We first assume that the property 𝑚 = 𝑛 + 1 holds for
+
all formulae with “n-value” less than some 𝑘 ∈ ℤ . Consider some formula φ such that φ
has 𝑘 binary connectives. We note that since 𝑘 is in the set of positive integers, φ must have
at least one binary connective. Therefore, φ must be logically equivalent to one of the
following formulae (where ψ1 and ψ2 denote arbitrary formulae):
ψ1 ∧ ψ2
ψ1 ∨ ψ2
ψ1 → ψ2
ψ1 ↔ ψ2
We observe that the “n-value” of ψ1 and ψ2 is less than or equal to 𝑘 − 1. For future ease,
we will define two functions.
𝑁: 𝕊 → ℕ, where 𝕊 denotes the set of all sentential formulae, and ℕ is the set of natural
numbers (including zero, as is conventional). This function maps each formula to the
number of binary connectives within it.
+
𝑀: 𝕊 → ℤ , this function maps each formula to the number of sentential variables within it.
We now opt to write the following equation: 𝑁(ψ1) + 𝑁(ψ2) = 𝑘 − 1. Since 𝑁(ψ1) and
𝑁(ψ2) are each less than or equal to 𝑘 − 1, it follows that 𝑀(ψ1) = 𝑁(ψ1) + 1 and
𝑀(ψ2) = 𝑁(ψ2) + 1. We observe that 𝑀(φ) = 𝑀(ψ1) + 𝑀(ψ2) = 𝑁(ψ1) + 𝑁(ψ2) + 2.
Furthermore, we can simplify this to 𝑘 − 1 + 2 = 𝑀(φ). Therefore, 𝑀(φ) = 𝑘 + 1.
Finally since 𝑁(φ) = 𝑘, 𝑀(φ) = 𝑁(φ) + 1. Therefore, if the property 𝑚 = 𝑛 + 1 holds
+
for all formulae with “n-value” less than some 𝑘 ∈ ℤ , then it holds for all formulae with
“n-value” equal to 𝑘.
Given the result of the base case and inductive step, we can claim that the property
𝑚 = 𝑛 + 1 holds for all sentential formulae. ◼
Matthews 2

20.
We first wish to prove that a positive formula is true when all of its sentential variables are
true. Additionally, we wish to prove that a positive formula is false when all of its sentential
variables are false. We will do these both at once using Complete Induction.

Consider a positive formula with zero connectives. This positive formula is only a sentential
variable. It then, rather trivially, follows that this formula is true when all of its sentential
variables are true and false when all of its sentential variables are false.

We now move on to the inductive step. Let φ be some positive formula with 𝑛 connectives
where 𝑛 > 0. We will also assume that all positive formulae with fewer than 𝑛 connectives
are true when all of their sentential variables are true and false when all of their sentential
variables are false. Since 𝑛 is greater than zero, we know that φ must have at least one
connective. Therefore, φ is logically equivalent to either ψ1 ∧ ψ2 or ψ1 ∨ ψ2. Since ψ1 and
ψ2 have fewer than 𝑛 connectives, each is true when all of their sentential variables are true
and false when all of their sentential variables are false. This then means that when all
sentential variables of φ are true, φ is logically equivalent to ⊤ ∧ ⊤ or ⊤ ∨ ⊤, both of
which are true. Similarly, when all sentential variables of φ are false, φ is logically
equivalent to 𝐹 ∧ 𝐹 or 𝐹 ∨ 𝐹, both of which are false. Therefore, if all positive formulae
with fewer than 𝑛 connectives are true when all of their sentential variables are true and
false when all of their sentential variables are false, then a positive formula with 𝑛
connectives is true when all of its sentential variables are true and false when all of its
sentential variables are false.

Combining the base case and inductive step allows us to conclude that all positive formulae
are true when all of their sentential variables are true and false when all of their sentential
variables are false.

Assume φ is some positive sentential formula that is also a tautology. Since φ is a tautology,
it must always be true. However, when all sentential variables of φ, a positive formula, are
false, φ is false. This is absurd; therefore, φ can not be both positive and a tautology.

Assume φ is some positive sentential formula that is also a contradiction. Since φ is a


contradiction, it must always be false. However, when all sentential variables of φ, a
positive formula, are true, φ is true. This is absurd; therefore, φ can not be both positive
and a contradiction.

Therefore, a positive formula may not be a tautology or a contradiction. ◼


Matthews 3

23.
Consider a formula with zero connectives or quantifiers. We note that a formula of the
predicate calculus without any quantifiers is already in PNF. Since each formula is logically
equivalent to itself, each formula with zero connectives or quantifiers is logically equivalent
to a formula in PNF.

Now consider a formula with 𝑛 connectives or quantifiers (note this means that the number
of connectives plus the number of quantifiers is equal to 𝑛). We will also let it be the case
that 𝑛 is a positive integer. Since this proposed formula has at least one connective or
quantifier, it can be represented as one of the following:
(1) φ ∧ ψ
(2) φ ∨ ψ
(3) φ → ψ
(4) φ ↔ ψ which is equivalent to (φ → ψ) ∧ (ψ → φ)
(5) ¬φ
(6) (∀𝑥)φ
(7) (∃𝑥)φ
First note that φ must have fewer than 𝑛 connectives or quantifiers, as is the same with ψ.
Therefore, there exists some formula equivalent to φ that is in PNF, and there exists some
formula equivalent to ψ that is in PNF. Also note that by rule LE (7) and rule LE (8) we can
have it be the case that the formula equivalent to φ and the formula equivalent to ψ do not
contain any quantifiers that quantify variables that are free in the other formula. Therefore,
we can now quickly convert the seven cases into PNF in the following manner:
For all first use rule LE to replace φ and ψ with their PNF equivalents. Then,
(1): Use rules LE (12) and LE (13) to move all quantifiers to the front.
(2): Use rules LE (14) and LE (15) to move all quantifiers to the front.
(3): Use rules LE (16, 17, 18, 19) to move all quantifiers to the front.
(4): On each of the conditionals, use the strategy from (3) then use the strategy from (1).
(5): Use rules LE (1) and LE (2) to move ¬ inside all of the quantifiers.
(6): It will already be in PNF
(7): It will already be in PNF
This shows that a formula from any of these cases is equivalent to one in PNF.
Therefore, by Complete Induction, every formula of the predicate calculus is equivalent to
one in PNF. ◼
Matthews 4

Section 16
Part A
2.
Since {¬, ∨, ∧} is complete, that means every formula of the sentential calculus is logically
equivalent to one using only the connectives in the aforementioned set. To show that {¬, ∧}
is complete, we will show that each formula using only the connectives ¬, ∨, ∧ is logically
equivalent to one using only ¬, ∧. (Then by transitivity of equivalence, the {¬, ∧} must be
complete. A proof of transitivity of equivalence is include at a later page *)
Consider some formula φ that uses only the connectives ¬, ∨, ∧. Upon considering the
subformulae of φ, we note that there are two cases.
(1) There does not exist a subformula of φ in the form ψ1 ∨ ψ2.
(2) There exists a subformula of φ in the form ψ1 ∨ ψ2.
In case (1), φ is already in the desired form.
In case (2), do the following:
1. Consider the formula φ
2. If a subformula of the form ψ1 ∨ ψ2 exists, move to step 3. If not, move to step 7.
3. Use rule LE (T) to replace one of the aforementioned subformulae with ¬(¬ψ1 ∧ ¬ψ2)
4. Call this new whole formula φ'
5. Consider φ'
6. Return to step 2
7. You are done
Therefore, each formula that uses only the connectives ¬, ∨, ∧ is logically equivalent to one
using only ¬, ∧. Since each formula of the sentential calculus is logically equivalent to one
using only ¬, ∨, ∧, each formula of the sentential calculus is logically equivalent to one using
only ¬, ∧. Therefore, {¬, ∧} is complete. ◼
Matthews 5

5.
From question 16A2, we know that {¬, ∧} is complete. To show that {↑} is complete, we
must show that every formula using only ¬, ∧ is logically equivalent to one using only ↑.
Again, this follows due to transitivity of equivalence.
Let φ be some formula using only ¬, ∧.
There are two cases
(1) φ has a subformula of the form ¬ψ or ψ1 ∧ ψ2.
(2) φ does not have a subformula of the form ¬ψ or ψ1 ∧ ψ2. (In other words, φ is just a
sentential variable)
For case (2), φ is already in the desired form.
For case (1), we first note that ¬ψ is equivalent to ψ ↑ ψ. Additionally, ψ1 ∧ ψ2 is
equivalent to (ψ1 ↑ ψ2) ↑ (ψ1 ↑ ψ2). A truth table illustrating this is included on a later
page **. We use the following process:
1. Consider φ
2. If there exist a subformula of the form ¬ψ proceed to step 3. If not, go to step 4.
3. Replace said subformula using rule LE (T) with ψ ↑ ψ. Go to step 6.
4. If there exists a subformula of the form ψ1 ∧ ψ2 proceed to step 5. If not, go to step 7.
5. Replace said subformula using rule LE (T) with (ψ1 ↑ ψ2) ↑ (ψ1 ↑ ψ2). Go to step 6.
6. Consider the new (aka post rule LE) whole formula, φ'. Go to step 2.
7. You are done.
Since each formula of the sentential calculus is logically equivalent to a formula using only
¬, ∧ and each formula using only ¬, ∧ is logically equivalent to a formula using only ↑, {↑} is
complete.
Matthews 6

Part B
3.
Consider some formula φ of two sentential variables, A and B, such that φ uses only ¬, ↔.
We wish to show that the truth table of φ must contain an even number of Ts. We will do so
via Complete Induction.
Base Case: Consider the case when φ has no connectives, then φ can only be “A” or “B.”
Therefore, the φ column of the truth table will have two Ts, and two is even.
Inductive Step: Suppose φ has 𝑛 connectives with 𝑛 ≥ 1. Also suppose that for all 𝑚 < 𝑛,
a formula using only ¬, ↔ of two variables with 𝑚 connectives has an even number of Ts in
its truth table column. We note that we can represent φ in one of the following two ways
(because 𝑛 > 0)
(1) ¬ψ
(2) ψ1 ↔ ψ2
We note that ψ, ψ1, ψ2 have fewer connectives than φ and therefore must have an even
number of Ts in their truth table column.
We will first consider case (1). Since the number of rows of a truth table with two
sentential variables (excluding the row which states formulae) is four, the number of rows
is even. Since ψ must have an even number of Ts, there must also be an even number of Fs
(even-even=even). Since ψ has an even number of Fs, ¬ψ must have an even number of Ts.
We now consider case (2). On another page *** we show all possibilities of number of Ts in
ψ1 and ψ2 and how many “overlapping” Ts or Fs they have (in other words, how many Ts
ψ1 ↔ ψ2 has). Since each of these possibilities results in an even number of Ts in ψ1 ↔ ψ2,
case two is done.
By Complete Induction, φ must have an even number of Ts in its truth table column. Now
consider the formula 𝐴 ∧ 𝐵. This formula has one T, and one is odd. Therefore, no φ can be
equivalent to 𝐴 ∧ 𝐵. Since there exists a formula of the sentential calculus that is not
logically equivalent to any φ, {¬, ↔} is not complete.
Matthews 7

* Proof of Transitivity of Equivalence


{1} 1. (φ ↔ ψ) ∧ (ψ ↔ ϕ) P (For CP)

{1} 2. φ ↔ψ 1T #4

{1} 3. ψ ↔ϕ 1T #4

{1} 4. φ →ψ 2T #18,4

{1} 5. ψ →φ 2T #18,4

{1} 6. ψ →ϕ 3T #18,4

{1} 7. ϕ →ψ 3T #18,4

{1} 8. φ →ϕ 4,6T #7

{1} 9. ϕ →φ 5,7T #7

{1} 10. φ ↔ϕ 8,9T #18

{} 11. [(φ ↔ ψ) ∧ (ψ ↔ ϕ)] → (φ ↔ ϕ) 1,10CP

** Truth table to show equivalences


ψ φ ¬ψ ψ ↑ψ ψ∧ φ ψ ↑ φ ¬(ψ ↑ φ) (ψ ↑ φ) ↑ (ψ ↑ φ)

T T F F T F T T

T F F F F T F F

F T T T F T F F

F F T T F T F F
Matthews 8

*** All possibilities


We note that ψ1 must have 0, 2, or 4 Ts. The same is true for ψ2.
In the case where each has 0 Ts, the equivalence formula would have 4 Ts.
In the case where one has 0 Ts and one has 2 Ts, the equivalence formula has 2 Ts.
In the case where one has 0 Ts and one has 4 Ts, the equivalence formula would have 0 Ts.
In the case where one has 2 Ts and one has 4 Ts, the equivalence formula would have 2 Ts.
In the case where each has 4 Ts, the equivalence formula would have 2 Ts.
In the case where each has 2 Ts and they overlap for two Ts, the equivalence formula would
have 4 Ts.
In the case where each has 2 Ts and they overlap for 1 T, the equivalence formula would
have 2 Ts.
In the case where each 2 Ts and they overlap for 0 Ts, the equivalence formula would have 0
Ts.
These are all of the possibilities.

You might also like