20.  USING CONDITIONAL DERIVATIONS


We have already seen how we can eliminate the arrow with the rules MP and MT, but we have not seen a corresponding rule to introduce it.  We can often enough find a way of establishing a disjunction and then using the implication rule to convert it to a conditional.

P & Q |- P -> Q

1.  P & Q    / show  Q
2.  Q            1,  Simp
3.  ~P v Q     2,  Add
4.  P -> Q     3,  Impl

note that we can do something like this whenever we know that the consequent of the intended consitional is true or that the antecedent is false


But we also have a new and often quite efficient derivation technique.  Conditional or hypothetical derivations give us a way to establish the truth of a conditional (in other words, allowing us to introduce the arrow).  The key to understanding this is that if we can show that a particular consequent would have to follow once we know a given antecedent then we can say the conditional relationship exists, regardless of whether the antecedent really is true.  Again, as we did with indirect derivations, we indent for a subordinate proof, and we indicate the intention of the  hypothesis.  Here we say "HCP" to indicate this is a hypothesis in a conditional proof (CP).

Here are some very simple examples:

P -> Q, Q -> R |- P -> R

1.  P -> Q
2.  Q -> R    /show P -> R
    3.  P       HCP           we hypothesize that the antecedent P is true
    4.  Q       1,3,  MP
    5.  R       2,4,  MP
6.  P -> R    3-5,  CP      we show that we have discharged the hypothesis


|-  (~P v Q) -> (P  -> Q)

    1.  ~P v Q   HCP
       2.  P        HCP      since the consequent is itself a conditional, we can bring in another subordinate derivation
       3.  Q       1,2,  DS
    4.  P -> Q    2-3,  CP
5.  (~P v Q) -> (P -> Q)       1-4,  CP

In working with quantified expressions we often find an exception to the usual rule about not introducing a universal quantifier with a name that did not first come from a statement for which the main connective was a universal quantifier. 

(Ex)Fx -> (y)Gy  |- (x)Fx -> Gx)

1. (Ex)Fx -> (y)Gy  / show (x)(Fx -> Gx)  
    2.  Fa             HCP
    3.  (Ex)Fx       2,  EG
    4.  (y)Gy         1,3,  MP
    5.  Ga             4,  UI
6.  Fa -> Ga        2-5,  CP  
7.  (x)(Fx -> Gx)   6,  UG
Now if this last step seems as though we are saying something about everyone when we only started with a particular name (Al's , let's say), the reason it works is that the hypothesis itself could have assumed any name at all and produced the same result.  We need not even have to assume that there is anyone meets the condition of being F so as to let us know that the same individual would also have be G.


EXERCISES  (ON YOUR OWN)

Symbolize each of the following arguments (both of which are intended to be valid) and use conditional 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 FOR BONUS CREDIT (2 POINTS)
Copy and paste the arguments and your work (symbolization in predicate logic and conditional proofs) into an email to me at dmcf34@yahoo.com no later than December 12

1.  If Alice is ambitious then she is a good student.  All good students are happy.  Therefore, if Alice is a good student she is happy.
2.  No clowns are sad.  Therefore, if Bozo is a clown he is not sad.
3.  Unless they are lazy all students will work hard.  Anyone who works hard will pass.  Therefore, if Alice is a student and she is not lazy we can be sure she will pass.
4.   If there is someone winning then everyone will be happy.  Therefore, anyone who wins will be happy.