19.
USING INDIRECT
DERIVATIONS
Direct
derivations are possible when
the original premises can be broken down (or transformed) into
individual expressions that then allow us to use our basic set of
introduction and eliminaton rules. But what if an argument form
is valid and we cannot use a direct derivation? We have the
alternative (which actually will work for any valid argument) of using
an indirect derivation,
meaning we assume that the intended conclusion
is false and show that this could happen only if we also found
something false in our premises. What it means to find something
false is that somehow, with a false conclusion, we can show that
statements in the premises or derived from them would contradict each
other.
Remember our basic definition of validity:
something about the pattern rules out having all true premises and a
false conclusion, so for an argument to be valid and still have a false
conclusion at least one of the premises must be false.
Indirect derivations will work for
any valid argument but, if a direct derivation is possible, you will
ordinarily find that it allows for a shorter proof. However,
there are times when you may not see easily enough how to work with
various rules to do this, so those are the times when it is a good idea
to do an indirect proof instead (and it may even be shorter).
The arrangement of an indirect proof is somewhat different in that we
indent for what we call a subordinate
proof. We begin with the hypothesis (HIP, standing for
"hypothesis in an indirect proof") ) that the conclusion really is
false
(just as we have with our consistency trees), then show that this would
lead to a contradiction. We then "discharge the hypothesis"
(and no longer indent) in the line following.
Example:
If Alice studies and has a good memory, then she will get most of the
answers correct and she will pass the test. Alice did pass the
test and she study, but she did not get most of the answers
correct. It follows that she does not have a good memory.
(S & M)
-> (A & P), P & S, ~A |- ~M
1. (S
& M) -> (A & P)
2. P
& S
3.
~A
/ show~M
4.
M
HIP here we start by hypothesizing that the
conclusion really is false in order to develop an indirect proof
note how we have begun indenting
5. S 2, Simp
6. S & M 4,5, Conj
7. A
& P 1,6, MP
8.
A
7, Simp
9. A
& ~A 3,8,
Conj this is the key: we
have arrived at an expression that is necessarily false
10.
~M
4-9, IP ~M must be true
since if M were true we would end up contradicting ourselves in our
premises, which we assume is not the case
If
you think back to some things you may have seen earlier in working with
truth tables, arguments in which the premises do already
contradict each other are said to be valid--and here all we mean is
that the conclusion does follow, not that it actually has to be true in
itself.
Example:
If Alice studies then she will pass. Alice did not pass, but we
know she studied. Therefore, Alice is wearing a blue
dress.
S
-> P, ~P & S |- W
If
we test for validity with a truth table we find that there is no bad
line.
P S W || S
-> P | ~ P & S | W
1 1 1
1
0 1
1 1 0
1
0
0
1 0 1
1
0
1
1 0 0
1
0
0
0 1 1
0
1
1
0 1 0
0
1
0
0 0 1
1
0
1
0 0 0
1
0
0
note that we would have the same result if
we turned around our original conclusion and said she was not wearing a
blue dress
1.
S -> P
2.
~P & S / W
3. ~W HIP
4.
S
2, Simp
5.
P
1,4, MP
6. ~P 2, Simp
7. P & ~P 5,6, Conj
8.
W
3-7, IP
A special use for indirect proofs is establishing that an expression is
a tautology (an expression that because of its form is always
true). We refer to these as zero-premise arguments.
If Alice studies and passes, then
she cannot both be studying and not passing.
|- (S & P) -> ~(S & ~P)
1. ~[(S
& P) -> ~(S & ~P)] HIP
2. ~[~(S
& P) v ~(S & ~P)] 1, Impl
3. (S
& P) & (S &
~P)
2, DM
4. S
&
P
3, Simp
5.
P
4, Simp
6. S
& ~P
3, Simp
7. ~P
6, Simp
8. P
& ~P
5,7, Conj
9. (S
& P) -> ~(S & ~P)
1-8, IP
EXERCISES
(ON YOUR OWN)
Symbolize each of the
following arguments (both of which are intended
to be valid) and use indirect derivations.
1. If Alice studies then
she will get most of the answers
correct, and if she gets most of the answers correct then she will pass
the test. Therefore, if Alice does not pass the test then she did
not study. [use S,A,P]
2. Knowing that Alice
studies only if she is not working and she
is not working only if she does not have a job proves that unless she
is not studying she does not have a job. [use S,W,J]
Go to the answer key
REVIEW EXERCISES:
Symbolize
using the suggested notation, then provide indirect derivations.
Copy and paste the problems into an email to me as you have before,
then fill in your answers. Yes,
these are the same problems you did for me in your last review
exercise, but now I am asking you to develop your skill in working with
an indirect proof. Remember to indent for the subordinate proof.
1. If
the tests are long or they are difficult, then students are not
happy. Students who work hard will be happy, and the students are
working hard. Therefore, the
tests must not be long. [L,D,H,W]
2.
Unless Alice studies she will not pass. She does need to pass in
order to graduate. Therefore, she will graduate only if she
studies. [S,P,G]
3. If every student is
ambitious then all the teachers are happy. Not all teachers are
happy. Therefore, there are students who are not ambitious.
[Sx,Ax,Tx,Hx]
4. Not all students are
ambitious. Any student who is not ambitious does not work
hard. Only students who work hard have good grades.
Therefore, some students do not have good grades.
[Sx,Ax,Wx,Gx]
5. If Jill is a good student
then she is ambitious. Anyone who is ambitious works hard.
All honor students are good students. Therefore, if Jill is an
honor student she works hard. [Gx,Sx,Ax,Wx,Hx,j]
6. No good students are
lazy. Only good students are honor students. Some
individuals are honor students. Therefore, there are students who
are not lazy.
[Gx,Sx,Lx,Hx; hint: do not symbolize
"individuals"]