Sei sulla pagina 1di 28

Logic & Proofs

Chapter 2 Content
Sentential Logic
Syntax and Symbolization
Contents:
Atomic Formulae and Logical Connectives
Compound Formulae and Symbolization
Overview
Conjunction
Disjunction
The Conditional
Negation
Combination of Connectives
Formal Syntax
Parse Trees
Conventions
Studying this chapter will enable you to:
1. Discern the logical structure of English sentences.
2. Symbolize English sentences as formulae of sentential logic.
3. Explain the grammar of the logical language of sentential logic.
4. Construct and identify formulae of sentential logic.
5. Construct and use parse trees.
Logic & Proofs
Sentential Logic: Syntax and Symbolization
Page 1 of 27
Atomic Formulae and Logical Connectives
One of the first things you probably did when you learned English grammar, way back in
grade school, was to become familiar with the basic parts of speech, that is, the different
categories of words, such as nouns, verbs, adjectives, and so on. We are going to start
in the same fashion when presenting the syntax of sentential logic. In English there are
many different categories of words, and you probably still don't know what all of them
are, or what words in each category are supposed to do. What exactly is a gerund,
anyway? The good news is, in sentential logic there are only two different categories of
words or basic expressions: ATOMIC FORMULAE and LOGICAL CONNECTIVES. Okay, so what
are atomic formulae and logical connectives?
We have already mentioned that the formulae of sentential logic correspond to sentences
of English, so atomic formulae, being formulae, correspond to sentences. The only
question is what makes these sentences atomic. The answer is that they correspond
to sentences that havefrom the point of view of sentential logicno logically relevant
internal structure. From other points of view, many of the sentences do have logically
relevant internal structure, as we shall see later.
So, then, what counts as logically relevant internal structure? You have probably guessed
that this has something to do with the other type of basic expressions, the logical
connectives. Indeed, it does. The logical structure we have in mind is reflected in
sentential logic by means of these connectives. We should thus go on to take a look at
what they are, and what they correspond to in English.
To begin with, logical connectives are a lot like what they sound like they would be:
they serve to connect formulae in order to create new and more complex formulae.
In our system of sentential logic, we start out with precisely four logical connectives,
called CONJUNCTION, DISJUNCTION, THE CONDITIONAL, and NEGATION. The connectives
correspond to certain words and phrases of English that we call LOGICAL OPERATORS,
which connect sentences, as follows: conjunction corresponds to the word and;
disjunction corresponds to or (or the phrase either...or...); the conditional corresponds
to (the phrase) if...then...; and negation corresponds to not.
Of course, when we say that a connective corresponds to a particular word or phrase
in English, we are not trying to claim that the connective should be considered to have
exactly the same meaning as the English word. All we are trying to claim is that the
connective captures a particular, logically important part of the meaning. The logical
connectives represent logical IDEALIZATIONS of the corresponding words and phrases in
Englishidealizations that allow us to focus on the logical structure of English sentences.
It is time we got to a few examples. Consider the following:
Example Sentences
1. John ran.
2. Mary laughed.
3. Harry said that Mary laughed.
Logic & Proofs
Sentential Logic: Syntax and Symbolization
Page 2 of 27
4. John thinks that Mary laughed at his running.
5. John ran and Mary laughed.
6. Either John ran, or Mary laughed.
7. If Mary laughed, then John ran.
8. John didn't run.
Which of these sentences should be represented using atomic formulae, and which ones
should not? Think about it for a moment before reading on.
Okay, now that you have thought about it, we will tell you. Since each of sentences 5-8
contains one of our logical operators, none of these sentences should be considered
atomic. Sentences 1-4, on the other hand, don't contain any of the operators, so each
one of theseeven 4, which doesn't look particularly simpleis atomic.
Now, take another look at sentences 5-8. We've changed the presentation a bit to make
their structure more obvious, enclosing occurrences of sentences 1-4 that appear within
these sentences in square brackets, with a subscript corresponding to its number from
the list:
5. [John ran]
1
and [Mary laughed]
2
.
6. Either [John ran]
1
, or [Mary laughed]
2
.
7. If [Mary laughed]
2
, then [John ran]
1
.
8. [John]
1a
didn't [run]
1b
Notice how sentences 1 and 2 occur in each of these logically COMPOUND sentences,
connected by the logical operators? This is exactly what we meant earlier when we said
that the logical operators serve to connect sentences to make new sentences. We will
see in a little bit how these sentences look when translated into sentential logic.
At this point, you might be wondering about sentence 3:
3. Harry said that [Mary laughed]
2
.
This sentence also contains sentence 2, but we said that it is an atomic sentence.
How can a sentence contain an atomic sentence, and yet still be an atomic sentence
itself? It probably is not particularly enlightening to say that it is because sentence 3
doesn't contain any of the logical operators. After all, that does not explain the difference
between making a new sentence out of atomic sentences by using logical operators and
making a new sentence by sticking Harry said that in front of an atomic sentence.
The secret here has to do with the fact that the logical operators are indeed logical,
and more precisely, that they are TRUTH-FUNCTIONAL. By contrast, the expression Harry
said that is not. In order to get clear on this, we take a brief detour into SEMANTICS, the
subject of the next chapter, which deals with the MEANING of expressions. We are, in
particular, interested in the TRUTH-VALUES of sentences. Compare sentences 3 and 5:
Logic & Proofs
Sentential Logic: Syntax and Symbolization
Page 3 of 27
3. Harry said that [Mary laughed]
2
.
5. [John ran]
1
and [Mary laughed]
2
.
In order to determine whether or not sentence 5 is true, we need to know whether or not
Mary laughed, as well as whether or not John ran. On the other hand, in the case of 3, we
don't need to know whether or not Mary actually laughed in order to determine whether
or not it was true that Harry said she had. In other words, the truth-values of sentence
2 and sentence 5 are related to one another (as are the truth-values of sentence 1 and
sentence 5), while the truth-values of sentences 2 and 3 are not relatedknowing one
won't help us determine the other, even though one sentence is contained in the other.
Since the truth-values of sentence 2 and 3 are not related, these sentences must have
the same truth-functional "status": they are both atomic sentences. Since the truth-
values of sentences 2 and 5 are relatedwe need to know the truth-value of sentence
2 (and that of sentence 1) in order to determine the truth-value of sentence 5they
can not have the same truth-functional "status," so 5 cannot be (and indeed, is not) an
atomic sentence.
So that is why some sentences can contain atomic sentences as parts, and still be atomic
sentences themselves. For the most part, though, all you need to know is that a sentence
that does not contain any logical operators is an atomic sentence. We should thus go on
to take a look at these logical operators in more detail.
Logic & Proofs
Sentential Logic: Syntax and Symbolization
Page 4 of 27
Compound Formulae and Symbolization
Overview
As we noted in the last section, a non-atomic sentence is called a logically COMPOUND
SENTENCE, since it is constructed out of a compound of atomic sentences and logical
operators. Similarly, in sentential logic, any non-atomic formula is called a COMPOUND
FORMULA. In this section, we will meet each logical connective in more detail, paying
particular attention to recognizing them as they appear in English expressions and
interpreting them in some tricky cases. We will also learn how to SYMBOLIZE English
expressions, and thereby TRANSLATE them into the language of sentential logic.
Logic & Proofs
Sentential Logic: Syntax and Symbolization
Page 5 of 27
Conjunction
We've already seen one example of a conjunction in English:
Example A Conjunction
John ran and Mary laughed.
Why don't we go ahead and symbolize this sentence? If we replace the word and with
the symbol & in the above sentence, we get the following:
John ran & Mary laughed.
We are now half-way to having symbolized our first sentence. The only thing left is to
determine what to do with the atomic sentenceswhich we call CONJUNCTS, since they
are connected by a conjunction. In order to symbolize these, let a single capital letter
represent each conjunct. That gives us the following translation:
(J&M)
We could have used any capital letters we liked to stand for the atomic sentences, but
it is traditional (and purely conventional) to pick a letter that has some relationship to
the corresponding sentence. Here we went with the first letter of the name of the person
mentioned in each sentence, but we could have used the first letter of the verb just as
well:
(R&L)
That is pretty much all there is to symbolizing a simple conjunction like this one. We
replace the word and with its symbol & and represent each of the conjuncts, the atomic
sentences, with a single capital letter each. Is that all there is to conjunctions, then? No,
not by a long shot.
As you well know, conjunctions can come in many forms. Consider the following:
Example Non-Sentential Conjunctions
1. The cat yawned and stretched.
2. The cat and the dog have been fed.
These sentences both contain the word and, so they should count as conjunctions,
rather than atomic sentences, right? Right. They do not consist, however, of two
sentences stuck together with the word and, so it is not clear how we should symbolize
them. The trick with sentences like these is to paraphrase themget them into the form
of two sentences stuck together by the word and:
1. The cat yawned and the cat stretched.
Logic & Proofs
Sentential Logic: Syntax and Symbolization
Page 6 of 27
2. The cat has been fed and the dog has been fed.
Now it is perfectly clear how we can symbolize them:
1. (Y&S)
2. (C&D)
Great, so every time we see the word and we know we have a conjunction on our
hands, we might just need to paraphrase it, right? Unfortunately, it is not quite so simple.
Consider the following:
Example Non-Truth-Functional Conjunction
Two scoops of vanilla ice cream and a generous serving of
chocolate syrup make a great sundae.
This is certainly true, but if we paraphrase it the same way as the previous sentences,
we end up with the following:
Two scoops of vanilla ice cream make a great sundae, and a generous serving
of chocolate syrup makes a great sundae.
It is obvious that the paraphrase trick does not work right in this casethe paraphrase
is evidently false, while our original sentence is true. Two scoops of ice cream by
themselves, regardless of flavor, do not make a sundae at all, let alone a good one.
Similarly, a generous serving of chocolate syrup by itself is just a generous serving of
chocolate syrup. Since the result of paraphrasing does not match what we started with,
it looks like we should symbolize this sentence as an atomic sentence:
S
To add another wrinkle, we have more words than just and to worry about when it comes
to conjunctions. Consider the following:
Example Conjunction Words Other Than 'And'
The cat is napping, but the dog is chasing his tail.
Although the cat is sharpening her claws on the dog, the dog is sleeping
soundly.
The cat is purring, though the dog is howling.
Mary has just taken the dog to the vet; however, the dog's appointment
is tomorrow.
Mary is fond of cats, whereas John likes dogs.
Logic & Proofs
Sentential Logic: Syntax and Symbolization
Page 7 of 27
Each of these sentences consists of two atomic sentences connected by some word other
than and. If we paraphrase each sentence by removing that word and sticking and
between the two atomic sentences, the paraphrases have the same truth-functional
meaning as the original sentences.
The cat is napping and the dog is chasing his tail.
The cat is sharpening her claws on the dog, and the dog is sleeping soundly.
The cat is purring and the dog is howling.
Mary has just taken the dog to the vet, and the dog's appointment is tomorrow.
Mary is fond of cats and John likes dogs.
The moral here is that conjunctions in English don't always contain the word and, though
they always consist of two sentences, called conjuncts, connected by means of some
word or phrase which, when replaced by and, results in a paraphrase that is truth-
functionally equivalent in meaning, though some of the original subtlety may be lost.
See the online version for interactive material here!
In the Introduction, we discussed one argumentative step or rule of inference involving
conjunctions (well, at the time we discussed only the word and explicitly, but as we
have just seen, and is one of many different conjunction words in English): any time we
have a conjunction, we can eliminate it to obtain one of its two conjuncts. Of course,
we can also introduce conjunctions, as follows:
(P1) P
(P2) Q
(C) (P&Q)
Using the conjunction symbol, now, we can call the rule corresponding to this step
&-introduction, since it adds a & to our argument. Again, this just mirrors our intuitive
understanding of conjunctions: if we know two things individually, then clearly we know
them together.
Just a quick note on notation before we move on: we chose to use the symbol & to
represent conjunction. The symbol is also commonly used for this purpose. The use of
the ampersand is well motivated, since the symbol is also used in place of the word and
in written English. The choice of the symbol (which is called the wedge) may not be
as perspicuous, but it should become evident why the wedge is common when you see
the symbol for disjunction in just a momenteach one is just the other upside-down.
Logic & Proofs
Sentential Logic: Syntax and Symbolization
Page 8 of 27
Disjunction
Identifying conjunctions in English might be a bit tricky. For one thing, we have to look out
for more words than just and. As if that were not enough, we can not always count on
a sentence being a conjunction just because the word and occurs in it. Disjunctions are
a bit easierthe only word you have to look for is or, plus the occasional either along
with it (that is, the phrase either...or...). We've already seen an example of a disjunction:
Example A Disjunction
John ran or Mary laughed.
In order to symbolize this sentence, we'll take the same approach as we did with
conjunction, first replacing the logical phrase with a symbol. For disjunction we use the
symbol , called vee:
John ran Mary laughed.
Note that the word or between the two sentences is removed from the sentence when
we symbolize the connective. If the word either appears at the begin of the sentence
(as it does in Either John ran or Mary laughed, for instance), it is also removed when
we symbolize the disjunction. All we have left to do now is to symbolize the atomic
sentences, which in this case we call DISJUNCTS, as they are connected by a disjunction:
(JM)
Just as with our example of a conjunction, all we do is replace each of the atomic disjuncts
with a suitably chosen capital letter. Also, just like conjunctions, disjunctions can come
in many forms. Consider the following:
Example Non-Sentential Disjunction
1. Either John or Mary laughed.
2. Mary either laughed or sneezed.
Here we can apply the same paraphrasing trick as we did with conjunctions to figure out
how to symbolize these sentences:
1. Either John laughed, or Mary laughed.
2. Either Mary laughed, or Mary sneezed.
Now we straightforwardly replace the logical words or and either with the appropriate
symbol and replace the atomic sentences with capital letters:
1. (JM)
Logic & Proofs
Sentential Logic: Syntax and Symbolization
Page 9 of 27
2. (LS)
You might be wondering at this point if we have the same sort of difficulty with or as
we did with and in our sundae example. Fortunately not. Consider the sundae example
again, this time with or in the place of and, plus its paraphrase:
Example Disjunction is Always Truth-Functional
Either two scoops of vanilla ice cream or a generous serving of
chocolate syrup makes a great sundae.
Either two scoops of vanilla ice cream make a great sundae, or a
generous serving of chocolate syrup makes a great sundae.
In this case, the paraphrase does have the same meaning as the original sentence,
even though the original sentence is obviously false. This time, we should symbolize the
sentence as a disjunction, not as an atomic sentence:
(TM)
At this point, you must be thinking that disjunction cannot be all that easyafter all, we
don't have the same wrinkles to worry about with disjunction that we do with conjunction.
Unfortunately, you are rightit isn't that easy. This time, however, the tricky part has to
do with the interpretation of or itself. The problem is that there are two possible ways
of understanding the word or as it is used in English. Consider the following:
Example Ambiguity of Disjunction
Either Mary will buy ice cream, or John will buy ice cream.
This sentence can be interpreted in two different ways. Under one interpretation, it
would mean that one or both of John and Mary will buy ice cream. That is, that the
sentence is true if John buys ice cream, if Mary buys ice cream, or if both of them do.
This interpretation is the INCLUSIVE interpretation of or. The other way to interpret the
sentence is to take it to mean that either John or Mary will buy ice cream, but not both.
That is, the sentence is true if either of Mary or John buys ice cream, but false if both of
them do. This is the EXCLUSIVE interpretation of or.
We have to decide to which of these two possible interpretations we want the symbol for
disjunction to correspond. As it turns out, there is a way to easily express the exclusive
version of disjunction using the inclusive version; just think of the English expression
either, but not both. So we will use the inclusive version as the one to which the
symbol will correspond. We will also always interpret or in English sentences as
the inclusive version, unless we make a point of specifying otherwise. The tricky bit
about disjunctions, then, is in determining whether to interpret a particular disjunction
as having the inclusive or the exclusive meaning. You already know how to symbolize the
Logic & Proofs
Sentential Logic: Syntax and Symbolization
Page 10 of 27
inclusive kind of disjunction, and we'll come back to the exclusive kind in a little while,
but first we need to look at the other two logical connectives.
See the online version for interactive material here!
Logic & Proofs
Sentential Logic: Syntax and Symbolization
Page 11 of 27
The Conditional
In English, the phrase if...then... connects two sentential clauses to form a CONDITIONAL
sentence. We have already seen one example:
Example A Conditional
If John ran, then Mary laughed.
We can symbolize this sentence as follows, using the same old letters for our atomic
sentences, and the symbol to replace the logical phrase if...then...:
(JM)
Unlike conjunction and disjunction, where we call the two sentences that are connected
by the same name (conjuncts and disjuncts, respectively), the two sentences connected
by a conditional are called by different names. The if clause, which appears to the left
of the is called the ANTECEDENT, and the then clause, which goes to the right, is
called the CONSEQUENT. Thus, in the above example, J is the antecedent, and M is the
consequent of the conditional (JM).
Conditionals, like conjunctions, can be a bit tricky in English, since they can disguise
themselves in a number of ways. All of the following sentences are equivalent to the
conditional we've just symbolized:
Example Conditionals in Various Forms
If John ran, Mary laughed.
Mary laughed, provided that John ran.
Given that John ran, Mary laughed.
Mary laughed if John ran.
John ran only if Mary laughed.
All of these sentences we symbolize in the same way as the original, despite the fact that
the atomic sentences do not always appear in the same order in the English sentences
as they do in the symbolization. In particular, notice the difference between the if and
the only if constructions. There are, of course, many other words and phrases that can
serve to form a conditional than just those in the examples above. We can't list them all
here, but the examples we have seen should be enough to give you the idea.
While the sentences we consider to be conjunctions and disjunctions are pretty
straightforward as far as their interpretations go, at least once we sort out the inclusive
versus exclusive issue for disjunction, conditionals are anything but straightforward.
Logic & Proofs
Sentential Logic: Syntax and Symbolization
Page 12 of 27
In fact, there is a long-standing debate about how to interpret conditionals that has
occupied quite a number of philosophers and logicians to this day.
We won't present that debate here, since it goes well beyond the scope of this course.
We adopt the traditional truth-functional viewpoint on the matter: a conditional is true if
either its antecedent is false, or its consequent is true.
See the online version for interactive material here!
Of course, there are argumentative steps involving the conditional, as well. Perhaps the
most natural, and certainly an important one, is -elimination:
(P1) (PQ)
(P2) P
(C) Q
This rule reflects the traditional truth-functional viewpoint we just adopted: if we know
that a conditional is true, and its antecedent is as well, we can then infer the truth of
the consequent.
Before moving on, there is one final term associated with the conditional that we should
discuss: unless. Consider the following:
Example Unless
John will pick up Henry at the airport, unless Mary does it.
The most perspicuous question to ask in order to translate this sentence is: when would it
be false? If neither John nor Mary goes to pick up poor Henry (who must then presumably
take a taxi home), clearly it will be false. Since a conditional is false only when its
antecedent is true and its consequent false, we can symbolize this sentence using a
conditional in the following manner:

(MJ)
That is, if Mary does not go to pick up Henry, then John will do it.
You might or might not consider this symbolization to be quite complete, however, since
it would be true in case both John and Mary went to pick up Henry, i.e., if John went
despite Mary going. This is very similar to the issue that arose with disjunction concerning
the inclusive and exclusive ways of interpreting a disjunctive sentence. Here, on the
inclusive interpretation, if both John and Mary go to pick up Henry, the sentence is true.
On the exclusive interpretation, it is false. While the single conditional would thus suffice
to capture the inclusive interpretation, we would need a conjunction of conditionals to
capture the exclusive one:

((MJ)&(MJ))
Logic & Proofs
Sentential Logic: Syntax and Symbolization
Page 13 of 27
Negation
Our final connective is negation. It might seem a bit odd to call negation a connective,
since it only "connects" one formula. We do indeed call it a connective, however, so odd
it will have to be. Actually, negation is a UNARY connective for this very reason, while
the other connectives are BINARY connectives, as they connect two formulae. We have
already seen one example of a negation in English:
Example A Negation
John didn't run.
Negation is different from conjunctions, disjunctions, and conditionals in that, in English,
it tends to occur in the middle of sentences. This can not only make it hard to see how
to go about symbolizing negations, it can also make it a bit difficult to sort out precisely
what is being negated. Yet again, we have a paraphrasing trick that can help on both
counts:
It is not the case that John ran.
By taking our original atomic sentence John ran and sticking the expression it is not
the case that in front of it, rather than sticking didn't in the middle (and consequently
changing the tense of the verb, because English is just odd that way), we end up with a
paraphrase that preserves the form of the original sentence, and gives us a better idea
of how to symbolize it. Let's go ahead and replace the expression it is not the case that
with the symbol for negation, , called the hook:
John ran.
Now it's clear that all we have left to do is symbolize our atomic sentence as usual:
J
We have already seen two different ways that negations can manifest themselves in
English, and you have probably realized that there are more. Consider each of the
following:
Example Non-Sentential Negations
1. The cat is unhappy.
2. Mary's lawyer is dishonest.
3. My shampoo is non-allergenic.
Even though the word not doesn't appear in any of these sentences, each of them can
be interpreted as a negation, as paraphrasing illustrates:
Logic & Proofs
Sentential Logic: Syntax and Symbolization
Page 14 of 27
1. It is not the case that the cat is happy.
2. It is not the case that Mary's lawyer is honest.
3. It is not the case that my shampoo is allergenic.
In each case, we remove the negative prefix from one of the words in the original
sentence, and insert the expression it is not the case that in front of the remainder of the
sentence instead. Not all negative prefixes can be treated in this fashion, but it is pretty
clear when you ought not to do this, since the result of removing the prefix in such cases
usually does not result in a real word, as would be the case with the following sentences:
Example Non-Truth-Functional Negatives
The dog is disturbed.
Mary is nonplussed.
Other than having to look out for negative prefixes, identifying negations is generally
easy, since the word not is going to appear somewhere in the sentence (possibly
contracted, as in didn't). The big difference between negation and the other connectives
is that with negation, you only have a single sentencerather than twoto worry about.
Now that we have met all of our connectives and learned how to use them to symbolize
English sentences, we should move on to take a look at more complicated sentences,
which meansyep, you guessed itmore than one connective at a time.
Logic & Proofs
Sentential Logic: Syntax and Symbolization
Page 15 of 27
Combination of Connectives
Consider the following sentence:
Example A More Complicated Sentence
Bob will either feed the ducks, or feed the chickens and gather
their eggs.
This sentence obviously involves more than just one logical operatorwe have both
an and and an or to deal withso it might not be entirely clear how to symbolize
this sentence. The first thing to do is determine the basic structure of the sentence.
Paraphrasing it yields the following:
Bob will feed the ducks, or Bob will feed the chickens and gather their eggs.
It is now clear that the basic structure of this sentence is that of a disjunctiontwo
sentences connected together by the phrase either...or.... Accordingly, we will say that
disjunction is the MAIN CONNECTIVE of the sentence, since the basic top-level structure
of the sentence is that of a disjunction. We might as well go ahead and replace the
`either...or' with the disjunction symbol right away:
Bob will feed the ducks Bob will feed the chickens and gather their eggs.
Since one of the disjuncts is an atomic sentence, we symbolize it at this point as well,
leaving only the compound sentence that makes up the right disjunct to be dealt with:
D Bob will feed the chickens and gather their eggs.
The remaining compound sentence is obviously a conjunction, but we should paraphrase
it in order to get clear on precisely what the conjuncts are. We consider just the
conjunction for this part:
Bob will feed the chickens and Bob will gather the chickens' eggs.
Okay, now that we have paraphrased it, it is clear how to symbolize it:
(F&G)
The only remaining question, then, is how to fit this conjunction into the translation as a
whole. If we just stuck it into the translation by itself, we would get the following:
D F & G
This is not going to work, though, since it is not at all clear from this translation that
the disjunction is the main connective and that (F & G) is one of the disjuncts.
We need some sort of punctuation that indicates the fact that the conjunction as a
whole represents one of the disjuncts of the original sentence. We use parentheses as
punctuation, like so:
Logic & Proofs
Sentential Logic: Syntax and Symbolization
Page 16 of 27
(D(F&G))
Now it is clear that the main connective is a disjunction, while the right hand disjunct
is itself a conjunction.
The parentheses here indicate that the influence or SCOPE of the conjunction is
contained within the parenthesesit connects only the two atomic formulae inside
the parenthesesits influence does not extend as far as the disjunction. The scope of
the disjunction, on the other hand, includes everything in the formula, including the
conjunctionwhich is, of course, the reason we called it the main connective in the first
place.
That wasn't so bad, was it? Let us try another example:
Example Another Complex Sentence
If Bob gathers eggs, then either he won't feed the chickens or
he won't feed the ducks.
Since this sentence includes the phrase if...then..., we can already tell that the structure
of this sentence involves a conditional. It is in fact the basic structure of the sentence,
since the only other logical phrases in the sentence (either and not) occur within
the then clause. The conditional is the main connective. Consequently, we replace the
English phrase with the symbol for the conditional, doing some paraphrasing along the
way:
Bob gathers eggs Either Bob won't feed the chickens or Bob won't feed the
ducks.
Since the antecedent of our conditional is obviously an atomic sentence, we might as
well symbolize it, and while we are at it, we insert parentheses around the consequent
(as it is not atomic):
G (Either Bob won't feed the chickens or Bob won't feed the ducks)
Now all we have to do is to translate the consequent. It is clear that the main connective
of the consequent should be a disjunction:
G (Bob won't feed the chickens Bob won't feed the ducks)
As each of the disjuncts is itself a negation, all we need to do is to translate these, and
we are done:
(G(CD))
You will note here that we don't need to put any parentheses around the negations, since
there is only one sentence to which each negation could possibly attach.
In the two examples we have seen, we did not have any trouble determining the top-
level structure (main connective) of the sentence. Sometimes, however, a sentence
Logic & Proofs
Sentential Logic: Syntax and Symbolization
Page 17 of 27
is AMBIGUOUS, and it is impossible to tell what the structure should be. Consider the
following:
Example An Ambiguous Sentence
It will rain and it will hail or it will snow.
This sentence could be paraphrased in two different ways, both of which correspond to
perfectly good readings of the original sentence:
Either it will both rain and hail, or it will snow
It will rain, and it will either hail or snow.
The original sentence is ambiguous, so we have no way of telling which of these
paraphrases is the right onethey are equally good. If you want to translate an
ambiguous sentence like the above, what should you do? Well, you could always pick
whichever disambiguation you think is appropriate, but if you are not absolutely sure,
you can play it safe and just use the disjunction of the two possible translations.
Before we proceed to look at the formal syntax of sentential logic, we have to fulfill one
promise we made a little while back. We mentioned that we had a way of expressing the
exclusive sense of disjunction in terms of the inclusive sensea way that corresponds
to the English phrase either, but not both. At the time, we could not symbolize this, but
now we can. Recall the example we were considering:
Example Exclusive Disjunction
Either Mary will buy ice cream, or John will buy ice cream.
On the exclusive interpretation of disjunction, we understand the sentence to mean that
either Mary will buy ice cream, or that John will, but that not both will buy ice cream.
Here is the symbolization of this exclusive reading of the sentence:
((MJ)&(M&J))
That promise fulfilled, we can finally head on to learn about formal syntax.
See the online version for interactive material here!
Logic & Proofs
Sentential Logic: Syntax and Symbolization
Page 18 of 27
Formal Syntax
A formal language can be defined by specifying both the simplest expressions of the
language and all of the rules for combining expressions together to generate formulae.
You met the simplest pieces of the language of sentential logic; its basic expressions
are the following:
Definition Basic Expressions of Sentential Logic
All and only the following are basic expressions of sentential logic:
1. The capital letters we use to symbolize atomic sentences, called
ATOMIC FORMULAE or occasionally SENTENTIAL LETTERS: A, B, C, and so
on (possibly with numeric subscripts).
2. The symbols for the logical connectives: &, , , and .
3. The parentheses used to disambiguate the scope of the connectives:
( and ).
Knowing what the basic expressions are, we can define the notion of an EXPRESSION of
sentential logic:
Definition Expressions of Sentential Logic
Any finite sequence or string of basic expressions is an expression of
sentential logic.
We are now in a position to state the rules that define the set of syntactically WELL-
FORMED FORMULAE, or, simply, FORMULAEthe grammatical expressions of sentential
logic. Here the crucial thing we need is some way of talking about any formula of
sentential logic. We will use lowercase Greek letters, like and as VARIABLES that range
over the set of all formulae in the language of sentential logic to accomplish this. We
make a point here of using something that can not be mistaken for sentential letters to
indicate that the variables are not themselves sentential letters, nor any other formulae
of sentential logic. Less frequently, we'll want to talk about any expression of sentential
logic, including expressions that are not formulae. For that purpose, we'll use uppercase
Greek letters, like and . That said, here are the syntactic rules for the language of
sentential logic:
Definition Formulae of Sentential Logic
1. Every atomic formula is a formula of sentential logic.
2. If is a formula of sentential logic, then so is .
Logic & Proofs
Sentential Logic: Syntax and Symbolization
Page 19 of 27
3. If and are formulae of sentential logic, then so are each of the
following:
a. (&)
b. ()
c. ()
4. An expression of sentential logic is a formula only if it can be
constructed by one or more applications of the first three rules.
As you can see, rules 1 through 3 are nothing more than a precise version of the informal
discussion we presented in the earlier parts of this chapter. The fourth rule is very
important: it is a catch-all rule to make sure that only the expressions generated in
accordance with the first three rules count as formulae. If we left out rule 4, we would
not be able to rule out expressions we do not count as formulae: the first three rules
only allow us to rule things in.
This is an INDUCTIVE definition, giving rules for generating or constructing the objects
falling under the definition. It is for the purpose of formulating these rules that the
variables , , etc., were introduced. We will see that clause 4 is crucial for deciding that
a particular expression is not a formula.
This definition says nothing about what formulae mean, despite the fact that we did make
some mention of how to interpret them in our previous discussion. That's because the
rules are strictly syntactic onesthe question of what the formulae mean, particularly
what their truth-values are, is a separate matter from syntax: it is a part of SEMANTICS.
Since we were learning how to translate from English to sentential logic, we did have
to talk about the meanings of expressions, but this is not properly part of syntax. We'll
go on to look at semantics in the next chapter, but first, there are a few more syntactic
issues we need to cover, as well as looking at an example of how to build up a formula
according to the syntactic rules. Let's do that right now:
Movie Constructing Formulae
Text/printable version available as a separate PDF!

Take a moment to see if you can use these grammatical rules fluently:
See the online version for interactive material here!
There is one last and very important thing regarding syntax that we need to discuss
before heading on to the next chapter on the semantics of sentential logic. That is the
matter of parse trees, and is a topic in its own right. So on to the next section.
Logic & Proofs
Sentential Logic: Syntax and Symbolization
Page 20 of 27
Parse Trees
A PARSE TREE graphically represents the internal structure of an expression of sentential
logic, making it possible to test whether or not the expression is well-formed, i.e.,
whether or not it is a formula. As such, parse trees are a very important syntactic tool,
and you need to know how to go about constructing them. First, however, let us take
a quick look at an example of a parse tree, so you see what sort of thing it is you will
learn to construct:
Example A Parse Tree
Okay, so here is how to construct the parse tree for an expression: Start out by writing
down the entire expression. (If you are constructing a parse tree for a known formula, be
sure to include all parentheseswhatever the expression at the root of the parse tree
may be, it is assumed to have no parentheses omitted.) It helps to write it out near the
middle of the top of the page you are working on, since parse trees grow downwards
and out to either side. As you write it down, check to make sure that it contains only
basic expressions of the languageif it contains a symbol not included in the basic
expressions, then it isn't actually an expression at all, and as such cannot possibly be a
formula. There's no point in continuing to work on a parse tree in that case. That done,
the next thing you need to do is to identify the general form of the expression. Is it of
one of the following forms?
1. A sentential letter;
2. A negation symbol, followed by any subexpression, i.e., something of the form ;
3. A left parenthesis, a subexpression, a conjunction symbol (occurring within no
matched pairs of parentheses other than the outermost ones), a subexpression, and
finally a right parenthesis, i.e., something of the form (&);
4. A left parenthesis, a subexpression, a disjunction symbol (occurring within no
matched pairs of parentheses other than the outermost ones), a subexpression, and
finally a right parenthesis, i.e., something of the form ();
5. A left parenthesis, a subexpression, a conditional (occurring within no matched pairs
of parentheses other than the outermost ones), a subexpression, and finally a right
parenthesis, i.e., something of the form ().
When you evaluate your expression, it is not only helpful, but as we shall see in more
detail momentarily, important to ignore everything inside matched pairs of parentheses
Logic & Proofs
Sentential Logic: Syntax and Symbolization
Page 21 of 27
within the outermost ones, to ensure that you find the top-level structure of the
expression. Thus, for example, when considering the expression ((PQ)&(RS)),
ignore everything inside the embedded matched pairs of parentheses, as though they
had nothing in between: (()&()). This makes it quite easy to see that the & is the
top-level connective.
What if there are multiple "top-level" binary connectives in the expression, though? In
(P&Q)&(Q&R), for example, occurrences of both & and occur within only the
outermost parentheses, as is evident when we ignore the contents of embedded pairs
of parentheses: ()&(). In any such case, the expression under consideration is not
well-formed, since the grammatical rules only permit construction of formulae having
one top-level binary connective. Without ignoring the contents of embedded matched
parentheses, we would need to attempt to construct a parse tree for each binary
connective in an entire expression to rule out each one as a possible top-level connective
ignoring the contents of embedded matched parentheses allows us to immediately
identify a formula as not well-formed if there are multiple binary connectives at the "top-
level" of any given expression.
If your expression does not fit into any of the five categories listed (i.e., it is not a
sentential letter, does not begin with a negation symbol, and is either not enclosed in
parentheses, contains no top-level binary connectives, or contains multiple top-level
binary connectives) then you know that the expression is not well-formed. After all, it
could not be obtained by any of the rules for building formulae, thus by clause 4, it is
not a formula. If the expression is a sentential letter, you are also done. If, however, you
are dealing with one of the four remaining categories, there is work left to be done.
See the online version for interactive material here!
Once you have identified the general form of the expression, you know how many
immediate subexpressions there are: Expressions of the form have one immediate
subexpression, the expression . Expressions of the form ( & ), ( ), and
( ) have two subexpressions each, the expressions and . The next step in
constructing the parse tree is to create one branch for each subexpression, including the
subexpression itself at the end of the new branch as a leaf. Here are the details:
Definition Parse Tree Construction Rules
Logic & Proofs
Sentential Logic: Syntax and Symbolization
Page 22 of 27
Repeat this process for each new leaf of the tree created that is not a sentential letter.
If you are able to keep going until each branch of the tree ends in a sentential letter,
the expression you started with is a well-formed formula. The parse tree, read from the
bottom up, reflects exactly the rules that are used to construct the expression as a
formula. If, however, you encounter an expression that is not a sentential letter and could
not be produced by any syntactic rule, the expression is not well-formed.
See the online version for interactive material here!
Why don't we take a look at a couple of examples at this point? Let us start out by
constructing the parse tree for ((P&Q)R), which is a well-formed formula:
Movie Parse Trees I
Text/printable version available as a separate PDF!

Now let's try constructing a parse tree for an expression that is not a well-formed formula.
Pay careful attention to the way we use the fourth syntactic rule (that if an expression
can't be constructed via the other rules, then it is not a formula) in this example:
Movie Parse Trees II
Text/printable version available as a separate PDF!

See the online version for interactive material here!
Parse trees are good for more than just checking well-formedness, however, as we shall
see in the next chapter when we discuss some of the semantic uses to which they can
be put. With the notion of a parse tree now at our disposal, there is just one last notion
we introduce with regard to syntaxthe notion of a SUBFORMULA:
Logic & Proofs
Sentential Logic: Syntax and Symbolization
Page 23 of 27
Definition Subformula
A formula is a subformula of a formula if and only if appears (as a
node) in the parse tree of .
There are two things to note about this definition of subformula: First, any formula that
appears in a parse tree for a formula is a subformula of , including itself. Second, only
a formula can be a subformula. This means that logical connectives and parentheses,
while they can occur in subformulae (just as they can in formulae), cannot themselves
be subformulae.
That about covers it for the syntax of sentential logic. Believe it or not, in this one chapter,
you have learned an entire new language! A little practice with what you have learned,
and you will be set to begin with semantics, so let us head on into the exercises.
Logic & Proofs
Sentential Logic: Syntax and Symbolization
Page 24 of 27
Conventions
A convention governs a practice of doing something in one particular way, if that practice
could have been done another way and still achieve its implicit goal. We have already
seen a few examples of conventions: our choice of symbols to represent the logical
connectives. We chose the ampersand symbol &, for example, to represent conjunction,
but we could have chosen other symbols, including the previously mentioned wedge, ,
and the dot, . (The latter symbols have indeed been used for that very purpose.)
The choice of symbols for connectives may be an obvious convention, but it is by no
means the only one at play in our syntax. We have discussed reasons to choose a
particular letter over another as the atomic formula to symbolize a given sentence, and
those rubrics (e.g., choose a letter that is mnemonic for the verb, or mnemonic for the
subject of the sentence, etc.) are themselves conventional. We could just as easily have
settled on a practice of always using A first, B second, etc., to symbolize sentences in
alphabetical order, or from longest to shortest, or in some other fashion. Furthermore,
the very choice of capital letters from the Roman alphabet to symbolize atomic formulae
is just as much a convention as the choice of symbol for each connective. Even the
choice of parentheses as the symbols we use to provide punctuation is conventional, and
somewhat peculiar to recent timesparentheses were much less commonly used in the
late 19th and early 20th centuries, with periods doing much of the work of punctuation.
Okay, so the symbols we use are matters of convention. What about the syntactic rules
themselves?
One aspect of the syntactic rules is conventional: the decision to put the symbols in the
order we do is one of convention. Instead of building a binary formula by putting the
symbol for the connective in the middle of the two subformulae, and wrapping the whole
thing in parentheses, we could just as easily have put the connective symbol first or last,
with the two subformulae next to one another. (Actually, all three styles have names:
putting the connective in the middle is called INFIX notation, in front is PREFIX or POLISH
NOTATION and at the end is POSTFIX or REVERSE POLISH NOTATION.)
So, what is it about the syntactic rules that isn't conventional? Well, no matter what
convention we may choose for the way we arrange two formulae and a binary connective
to make a new formula, the fact remains that two formulae plus a binary connective
can be combined to make a new formula. Of course, similar remarks apply to unary
connectives (negation) and a single formula, even though there may be fewer options
as to how one can arrange the components. After all, our goal here is to have formulae
reflect logically important structure.
Relatedly, there is an important feature of our set of syntactic rules that goes beyond
convention: for any well-formed formula, there is one and only one way it can be built
up according to the syntactic rules. Why is this important?
As we have seen with parse trees, it makes it possible to always be able to determine
whether or not an expression is well-formed, and if it isn't, we can always give an
explanation of why it can not be produced by the syntactic rules.
Looking at it another way, it means that we can always tell which formula or formulae
go with which connectives. As we have seen, this is crucially important when it comes
Logic & Proofs
Sentential Logic: Syntax and Symbolization
Page 25 of 27
to avoiding ambiguity. You may recall that this is the very reason we decided to use
parentheses as punctuation in the first place, which brings us to the topic of a slightly
different kind of convention.
You have undoubtedly noticed that a lot of parentheses appear in the formulae of
sentential logic; in fact, every formula we form by means of a binary connective is
enclosed in parentheses. While we do need parentheses to keep formulae unambiguous,
it isn't actually necessary to always include all the parentheses. On occasion, we have
omitted the outermost pair of parentheses in a formulaafter all, since they are the
outermost pair, there is no way to make a formula ambiguous by omitting this pair.
In order to help keep things readable by omitting some parentheses, we can adopt the
same sort of conventions as those we find in arithmetic for determining the order of
operations. Consider the following:
3 4 + 5
The order of operations in arithmetic specify that multiplication should be carried out
first, then addition, like so:
(34)+5=12+5=17
If we had wanted the addition to be carried out first, on the other hand, we would include
parentheses to indicate this:
3(4+5)=39=27
Why do we carry out multiplication before addition when there are no parentheses?
Actually, this is just another matter of conventionthe order of operations could just as
well be different. We just happened to settle on doing it one way, rather than another.
All we need to do is thus to settle on an "order of operations" for the sentential
connectives, i.e., we need to choose the order of precedence for the connectives. We
adopt the following order: , &, , .
Let's take a look at an example:
Example A Formula with Parentheses Omitted
A & B C D
We read this formula as follows:
(((A&B)C)D)
As a further example:
Example A Formula with Parentheses Omitted
We read the following formula:
Logic & Proofs
Sentential Logic: Syntax and Symbolization
Page 26 of 27
A B C & D
as
(A(B(C&D)))
There is just one last thing we need with respect to parentheses. In cases where we have
multiple occurrences of the same connective in a row, we group them from right to left.
Thus, for example, we read A B C as (A(BC)).
The conventions we have adopted can also be expressed as a procedure for inserting
parentheses back into a formula from which some have been omitted. Here is that
procedure:
Definition Procedure for Reinserting Omitted Parentheses
1. First, insert parentheses around every occurrence of & and its two
conjuncts, starting with the rightmost & and ending with the leftmost,
2. Next, insert parentheses in the same fashion for each and its two
disjuncts, from rightmost occurrence first, to the leftmost occurrence
last,
3. Finally, insert parentheses for each , and its antecedent and
consequent, from rightmost occurrence first, to the leftmost
occurrence last,
As you follow this procedure, keep in mind that parentheses are never
inserted around negations, and that a single parenthesis should never be
inserted within another set, i.e., do not break up any existing pairs of
parentheses with ones you insert.
Once you have finished inserting parentheses, one easy check you can make to help
ensure that you have the inserted parentheses right is to count them. There should be
the same number of left and right parentheses, and the same number of pairs as there
are binary connectives (conjunctions, disjunctions, and conditionals).
That is all there is to inserting and omitting parentheses. In practice, you won't find
nearly as many parentheses omitted by logicians as our conventions allow, since a few
parentheses here and there can make a formula much more easily readable, and not
omitting parentheses keeps things completely unambiguous. Throughout the course, we
will generally refrain from omitting any parentheses, other than perhaps the outermost
ones for a formula. The notable exception is this: on occasion, we will omit parentheses
from formulae with multiple and adjacent occurrences of the same connective, e.g.,
(P&Q&R) rather than (P&(Q&R)), or ((P&Q&R)S) rather than
((P&(Q&R))S).
Logic & Proofs
Sentential Logic: Syntax and Symbolization
Page 27 of 27
Note: In homework assignments and activities, be careful to follow
directions concerning what parentheses may or may not be omitted. In
most cases, all parentheses should be includedif in doubt, include them
all!
See the online version for interactive material here!

Potrebbero piacerti anche