Some Notes on Logic
Craig Fox
Summer, 2002; Rev. Fall, 2004
1
1.1
Introduction.
What is our approach to logic?
There are many different ways one might approach the study of logic. For
example, we might think of logic as describing how we think, or better,
how we ought to think. This way of approaching the subject, however,
lends itself to some difficulties. Perhaps the most pressing difficulty might
be exactly how we would find out how we think (or ought to). Would we
somehow study people while they think? How would we determine exactly
what it was that they were really thinking? How would we determine if
someone was thinking poorly or well?
It does not seem that there are clear answers to these questions. If
we proceed in this way, logic is in danger of becoming too psychological, we
might say. There are a number of reasons why we should be wary of such
an outcome. A different approach would probably be better.
The ultimate question with which we will be concerned, with our study
of logic, is whether or not some given argument is good or bad. Partially
in order to answer questions of this sort in a general way, our logic will be
a symbolic one. This means that one of our tasks, when trying to assess an
argument, will be to translate an argument into symbolic notation. In this
way, our assessment is not simply of the one particular argument at hand,
but rather it is of a whole class of arguments sufficiently similar to the one
at hand. (Exactly how we explain what we mean by the phrase sufficiently
similar will determine different formulations of logic.)
Given a symbolic representation of our argument, then, we will want
to apply some kind of test. The test should work in such a way that if our
orginal argument was a good one, then the test outputs good (and bad
if the argument was a bad one). The first step in devising such a test will
Some Notes on Logic
be to narrow down exactly what it is that we mean by our terms good
and bad.
1.2
Definitions
Let us call an argument valid if and only if it is not possible to make its
premises true at the same time as its conclusion false. The premises are
what were arguing from, the conclusion is what were arguing for. (Our
definition of validity reveals that we are interested in deductive arguments,
as opposed to inductive ones.) An argument that is not valid, we will call
invalid. Note that this is meant to coincide with our initial judgments of
good and bad.
Let us call an argument sound if and only if (i) the argument is valid, and
(ii) its premises are (in fact) true. The collection of all sound arguments,
then, is a subset of the collection of all valid arguments. That is, all sound
arguments are valid, while not all valid arguments are sound. Note that in
our definition of validity, we did not ask whether or nor the premises were
actually true. (If an arguments premises were true and the conclusion could
still be false, then it is invalid.)
1.3
Exercises
1. Is each of the following arguments valid or invalid?
(1.1) All students are mortal. Bertrand is a student. Therefore, Bertrand is
mortal.
(1.2) Snow is black. The sky is white. Therefore, snow is black and the sky
is white.
2. Is each of the following arguments sound or unsound?
(2.1) Some chalk is white. Some pens are black. Therefore, some chalk is
white and some pens are red.
(2.2) All men love sports and cars. Your instructor is a man. Therefore,
your instructor loves sports and cars.
(2.3) Snow is white. The sky is blue. Therefore, the snow is white and grass
is green.
3. Is it possible for a sound argument to have a false conclusion? Explain.
Some Notes on Logic
2
2.1
Connectives and truth tables.
Not
The first of our logical connectives is Not. (A connective is something
that attaches to a sentenceor sentencesmaking one bigger sentence.)
Note that this is our only one-place connective, i.e. it attaches to a single
sentence. Think of the clumsy English expression It is not the case that.
If I assert: It is raining, then my assertion is true if and only if it is
actually raining. If I assert: It is not the case that it is raining, or It
isnt raining, then these are true just in case it is not true that it is raining.
More abstractly put:
p p
T
F
F
T
This is to say the following. If p is true, then its negation, p, is false.
Similarly, if p is true, then just p is false. Check that this is correct by
trying a couple of example sentences on your own.
Note that many times, in English, not sentences are somewhat disguised
as contractions. So Dogs arent dumb says the same thing as the more
explict It is not the case that dogs are dumb.
2.2
And
The rest of our connectives are two-place connectives. This is to say
that they connect two sentences. And works in the following way. It is
true if and only if each of the parts (called conjuncts) are true. Otherwise,
the and sentence (called a conjunction) is false. In table form:
p q pq
T T
T
T F
F
F T
F
F F
F
Check that the values represented in this table make sense along with your
understanding of and in English.
2.3
Or
An or sentence (called a disjunction) is true exactly when any of the
parts is true. (We call the parts disjuncts.) Note that this definition implies
that when both parts of a disjunction are true, so is the disjunction as a
Some Notes on Logic
whole. (Our or is inclusive of this true-true case.) Represented in table
form:
p q pq
T T
T
T F
T
F T
T
F F
F
Again, compare the values represented above with your intuitive understanding of or in English. (Sometimes, in English, when we assert a disjunction, we mean to assert that only one of the disjuncts can be true. Our
wedge does not correspond to such usage. We will return to this point later.)
2.4
Only if
The only if corresponds to a conditional assertion in English. Often,
this is seen as a sentence of the form If... then... (The following forms are
equivalent, for our purposes: p only if q and if p, then q.) The key to
understanding the conditional is that it is false exactly when the first part
(called the antecedent) is true and the second part (called the consequent)
is false. In all other cases, the conditional is true. As a table:
p q pq
T T
T
T F
F
F T
T
F F
T
Think of examples, such as the following: If you do your homework, then
you may watch television. Your parent would have been lying to youi.e.
asserting something falseif you did do your homework and then you were
not allowed to watch television. (That is to say, the assertion is false if the
antecedent is true and the consequent is false.) The conditional is probably
the trickiest of the connectives. One might think it odd, for example, that
a conditional with a false antecedent and a false consequent is true. This
means that the statement: If 2+2=5, then your instructor is the ruler of the
world is true. These kinds of examples notwithstanding, there are several
good reasons why we take the conditional as represented above. (This way
of understanding the conditional is called the material conditional.)
Some Notes on Logic
2.5
If and only if
We can think of this last connective basically as a kind of equal-sign
for truth values. An if and only if sentence is called a biconditional. As
such, when each of the parts (called limbs) is true, the biconditional is true.
Also, when each of the parts is false, the biconditional is true. When the
truth values of the parts differ, then the biconditional is false. Represented
in tabular form:
p q pq
T T
T
T F
F
F T
F
F F
T
One other note about if and only if. This phrase will often be abbreviated
as iff. You will see iff used frequently in definitions. For example: x is a
square iff x is a rectangle and its sides are all of equal length. See above,
throughout, for more examples of this usage of if and only if.
2.6
Definitions
Let us say that a sentence is complex iff contains at least one connective.
A sentence that is not complex is called simple. A complex sentence (since
it is well-formed) will then also contain at least one simple sentence.
The above tables are called truth tables. Truth tables lay out all of
the possible truth assignments to the component parts of a sentence. For
n component parts, there will be 2n possible truth assignments. (The base
here is 2 because there are two truth values: true and false.) So, for a
complex sentence built up out of 2 simple sentences, there will be 4 possible
truth-assignments. (See, for example, the table for if and only if above.)
For a complex sentence built up out of 3 simple sentences, there will be 8
possible truth-assignments. Check that each of the tables above captures all
of the possible truth value combinations.
2.7
Exercise
1. Suppose that you are told to give a truth table for a new two-place
connective. You are given that the connective is true if and only if both
parts are false. Let us symbolize the connective with . Produce the table
for p q.
Some Notes on Logic
3
3.1
Regimentation of arguments.
The strategy.
Our goal is to arrive at a symbolized version of a given argument. Many
sentences in arguments are not simple, therefore, some work is sometimes
required in order to come up with a symbolized sentence that corresponds to
the English version. In particular, one needs to determine what the simple
parts of the sentence are, as well as what connectives are used. We will call
this process of coming up with a symbolized version regimentation (or,
sometimes, translation, though this term can be misleading).
Let us take a single sentence as an example: The sky is blue and the
chalk is white. What are the simple sentences in this complex one? There
are two: we have The sky is blue and The chalk is white. What is the
connective? We have just one: and. So, the sentence is a conjunction of
two simple sentences. Let us symbolize simple sentences with capital letters.
We might assign S to the sky is blue and C to the chalk is white. Our
completely symbolized sentence is then: S C .
Our strategy for symbolizing entire arguments will be just the same.
Each premise will get symbolized, as will the conclusion. (It is important
that the same simple sentences get symbolized with the same capital letter,
no matter where they appear in the argument.) After doing this, we will
be able to use our truth tables to test the symbolized argument for validity
(see next section).
3.2
Examples
Example 1: Dogs bark only if cats sleep. Dogs bark and ticks bite.
Therefore, cats sleep. Take each premise and the conclusion separately.
The first premise has two simple components: Dogs bark and Cats sleep.
The connective is the only if. So symbolizing dogs bark as D and cats
sleep as C, we get the symbolized sentence: D C .
The second premise has a new simple sentence as a component: ticks
bite. Let us symbolize this with T. The whole second premise will then
be symbolized as: D T .
The conclusion has no new components. We shall symbolize it, then,
simply as C .
The whole argument, then: D C, D T, theref ore C .
Example 2: You will pass this course or you will not be happy. You
will be happy. Therefore, you will pass this course. The first premise has
Some Notes on Logic
two simple components. They are you will pass this course (P), and the
second is you will be happy (H). Note that the not is not part of the
simple component: its presence in a sentence means that the sentence is
complex. So the first premise is symbolized: P H .
The second premise has no new simple components. It will just be
symbolized as follows: H .
The conclusion has no new parts. It will be symbolized as follows: P .
The whole argument symbolized, then: P H, H, theref ore P .
3.3
Exercises
Symbolize the following arguments.
(1) Snow is white. The sky is blue. Therefore, the snow is white and grass
is green.
(2) The Cubs will win or the White Sox will win. The Cubs will not win.
Therefore, the White Sox will win.
(3) We will visit Mars only if it is inhabited. It is inhabited. Therefore, we
will visit it.
(4) If Mars is inhabited, then we will visit it. It is inhabited. Therefore, we
will visit Mars or we will visit Jupiter.
4
4.1
Testing an argument for validity.
Computing Values of Complex Expressions.
First recall our definition of validity. An argument is valid if and only
if there is no possibility of making the conclusion false while holding the
premises to be true. What we need, then, is a way of surveying the different
possible truth-value assignments. We already have such a method: truth
tables.
What we need, however, is to be able to examine the truth-value assignments of more complex sentences. (All of our previous tables were, we might
say, the most basic tables for each of the connectives.) Again, we already
have the tools before us. Since all of our complex sentences are built up out
of our five connectives (together with simple sentence letters), we can use
the basic truth tables to compute the values of the more complex sentences.
Let us look at an example.
Example 1. (A B) B.
Some Notes on Logic
Set up a truth table for this expression. As it contains two simple sentence
letters (A and B), well need 22 = 4 rows.
A B (A B) B
T T
T F
F T
F F
The real question here is how to fill in the rest of the table. Notice that the
complex sentence is an only if (conditional) statement. It happens to have
a complex antecedent, and this antecedent is itself a conjunction. We know
how to fill in the table for A B; thats just our basic and truth table. So
lets put in that part. (Well fill it in under the A B.)
A B (A B) B
T T
T
T F
F
F T
F
F F
F
So now we have the values for the antecedent filled in. We know the values
for the consequent; theyre simply the column for B from the front of the
table. Lets fill that in, under the B.
A B (A B) B
T T
T
T
T F
F
F
F T
F
T
F F
F
F
We have just one more step to go, and thats the over-all value for the whole
complex expression. What we do is to use the column for the antecedent
and the column for the consequent, and we enter the value for the only if
using those values together with our basic table for the arrow. That will
give us the complete table:
A B (A B) B
T T
T
T T
T F
F
T F
F T
F
T T
F F
F
T F
We have all Ts going down the final column, because there is no instance
(row) where the antecedent is true and the consequent is false. (Notice that
this means that the sentence is never false, no matter what truth values we
assign to A and B.)
Some Notes on Logic
Example 2. [((A B) A) (A B)]
A B [((A B) A) (A B)]
T T
T
T F T
T
T F
F
F F F
T
F T
F
T T T
T
F F
T
T T F
F
This formula is a conjunction. (Another way of saying this: the main connective of the formula is the conjunction represented by the wedge. The
main connective is above the last column you compute.) Hence, the column
in bold (under the wedge) represents the table for the expression as a whole.
The first column was computed first, then the third. Then the second column could be computed, and this represents the values for the left conjunct.
The right conjunct is simply a basic or statement. Now we used the left
conjuncts values together with the rights, and we mark a T where both
values are true. (Otherwise, we mark an F.)
4.2
Exercises
Complete truth tables for the following complex expressions.
(1) (A B) (A B)
(2) (A B) (B A)
4.3
Testing Arguments for Validity.
Our goal was to test complete arguments for validity. We now have a
method for doing this: we list all of the premises, together with the conclusion, at the top of the table. We then compute their values. When this is
complete, we examine the table. Are there any rows in which the premises
are each true, while the conclusion is false? If so, this row represents a
counterexample, and the argument is invalid. If there arent any such
rows, then there are no counterexamples, and the argument is valid.
Example 3. (A B), B, theref ore A
Some Notes on Logic
A
T
T
F
F
B
T
F
T
F
(A B)
T
T
T
F
B
F
T
F
T
10
A
T
T
F
F
So this argument is valid: there is no row in which the premises come out
true and the conclusion comes out false. Let us look at a slightly different
example.
Example 4. (A B), B, theref ore A
A B (A B) B A
T T
T
T T
T F
T
F T
F T
T
T F
F F
F
F F
So this argument is invalid: the arrow at row three indicates the counterexample. In this row, both of the premises are true and the conclusion is false.
In other words, we have a counterexample when A is false and B is true.
4.4
Exercises.
(1)(4): Test the arguments of 3.3, above, for validity.
5
5.1
Two common valid argument forms.
Modus ponens
Modus ponens (latin for affirming mode) is the name for the following
pattern of inference:
p q, p, theref ore q
An example: If you like cake, then you like frosting. You like cake. Therefore, you like frosting.
5.2
Modus tollens
Modus tollens (latin for denying mode) is the name for the follwing pattern
of interence:
Some Notes on Logic
11
p q, q, theref ore p
An example: If you like cake, then you like frosting. You dont like frosting.
Therefore, you dont like cake.
6
6.1
Two fallacious argument forms.
Affirming the consequent
The following argument form is not a valid pattern of inference:
p q, q, theref ore p
An example: If you like cake, then you like frosting. You like frosting.
Therefore, you like cake.
6.2
Denying the antecedent
The following argument form is also not a valid pattern of inference:
p q, p, theref ore q
An example: If you like cake, then you like frosting. You dont like cake.
Therefore, you dont like frosting.
6.3
Exercises.
(1) Show that modus ponens and modus tollens are each indeed valid patterns of inference.
(2) Show that affirming the consequent and denying the antecedent are
indeed not valid patterns of inference. (Do this by giving an assignment of
truth-values to the letters p and q that make the premises of the argument
form true while making the conclusion false.)
Some Notes on Logic
7
7.1
12
Use and Mention
Motivation.
There is a difference, in our language, between using an expression and
mentioning one. It should be obvious that it is important, for whatever
purposes we have, that we know what were saying with our words. When I
mention an expression, I am talking about the expression itselfas opposed
to talking about whatever it is to which the expression refers. Here is an
example of mentioning an expression:
The first letter of dog is a consonant.
We use the quotation marks to designate the fact that we are mentioning
the expression, instead of using it. (We can think of the quotation marks
as providing us a kind of name for the word, enabling us to talk about it.)
Consider the same sentence without the quotation marks around dog:
The first letter of dog is a consonant.
This sentence is meaningless! It purports to talk about something called
the first letter of dog. This is why it is important to distinguish between
use and mention: so that our sentences at least make sense. We can evaluate
the philosophical adequacy of what were talking about laterbut if were
not even making sense with our words (if arent clear about what it is that
were talking about) then our philosophical conversation cant even really
begin.
Notice the use of quotation marks in the sentence before this last example
sentence. There, I wished to talk about the sentence not having quotation
marks around a certain word, so I needed to use quotation marks when
referring to that word. (Otherwise, I would have been talking about some
actual dog that happened to have quotation marks hovering above it...)
We shall use single quotes () and double quotes () much as we
use parentheses () and brackets [ ]. Single quotes will be used to mention
expressions inside of an expression that is already mentioning an expression.
An example of this:
The first letter of dog is a consonant is true.
Some Notes on Logic
7.2
13
Exercises.
Place quotation marks in the following sentences in order to have them make
sense grammatically and factually.
(1) Snow is white is true if and only if snow is white.
(2) Grass is green is true but grass is only five letters long.
(3) Use to mention to mention to mention.
Some concluding exercises.
(1) (Jeffrey, p.20; also Smullyan) Knaves always lie, knights always tell the
truth, and in Transylvania, where everybody is one or the other (but you
cant tell which by looking), you encounter two people, one of whom says
Hes a knight or Im a knave. What are they? (Hint: you dont necessarily need to try to regiment this argument; rather, try reasoning through the
different possibilities.)
(2) Explain the process of regimenting an argument.
(3) Symbolize the following argument and test it for validity:
If Bob is a snark, then Bob is a dwark. Bob is not a dwark or Bob is a
thwark. Bob is not a thwark. Therefore, Bob is not a snark.
(3.1) Is the argument in (3) sound or unsound? Explain your answer.
(4) Place quotation marks in the following, in order to have it make sense
grammatically:
I used to use to mention to mention to mention.
Some Notes on Logic
14
References
Hart, W.D. Some exercises on use and mention. (unpublished)
Jeffrey, Richard Formal Logic: Its Scope and Limits 3rd Edition (McGrawHill: 1991)
Klenk, Virginia Symbolic Logic 3rd Edition (Prentice Hall: 1994)
Quine, W. V. O. Philosophy of Logic 2nd Edition (Harvard: 1986)
Quine, W. V. O. Elementary Logic Revised Edition (Harvard: 1980)
