Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Vacuous truths

  1. Nov 4, 2011 #1
    I dont understand the concept (or need for) of vacuous truths/implications. Why is it that if say a statement a is false then we can conclude that any implication a => b is 'true'?

    Ive been reading online on this but everything has been vague so far, the most sensible explanation Ive seen so far goes like this: All implications in logic are either true or false (and not both) so an implication of the form a => b (where a is false) is either true or false but the only case where such an implication is false is when a is true and b is false which is not the case therefore we must conclude that it is true.

    Now this almost elucidates the issue for me but then it evokes another issue for me. If we can argue in the manner that is done above then we can essentially argue that implications between completely unrelated things are true. Consider for example the statement "If the derivative of sinx is cosx then quadratic equations have a general solution/formula". Now clearly this is either true or it is false, it clearly is not false as both (the derivative of sinx is cosx) and (quadratic equations have a general solution/formula) are true so therefore this implication is true. In general we can use such an argument to conclude that any implication a => b (where it is not true that a is true and b is false) is true. But this seems completely counterintuitive to me because there appears to be no casual link between the a and b in my concrete example and this idea suggests that implications can be true for statements that are completely unrelated and have no casual relationship between them which appears to contradict what I thought was the definition of an implication (namely that a entails b, or that a CAUSES b).

    In fact what do we mean when we say then that an implication statement a => b is "true", do we mean that the truth of a actually necessarily causes the truth of b or that is possible for a to cause b even if we dont know there is a casual relationship?

    By the way I havent had any exposure to undergraduate Maths at all yet so please no heavy use of set theory or formal logic etc as Im just a beginner
  2. jcsd
  3. Nov 4, 2011 #2
    It's interesting that I made it this far in math (final year PhD student) without thinking about this. Of course, I'm not a logician.

    You can have silly implications like that where there's no causal link. The reason why you never see them is precisely because they are silly. They aren't really worth mentioning. If you have already proved that both a and b are true, the fact that a implies b is not really worthy of note. So, the causality comes into play because without that, no one would be interested.

    If you think about it, if you are trying to prove a implies b, you are allowed to assume anything that you have already proven, as well as the fact that a is true. There's no rule that says you must necessarily USE a in the proof. But if you don't use a, it would be silly to make the statement.

    Implication means IF a is true, then b is true. The key word is if. So, you don't really care what happens if a is false. You could declare the implication to be false if you wanted, if a is false. It's just a matter of convention. Someone has to make up definitions (and then the mathematical community agrees on them). They aren't written in the sky. The accepted convention is to define it to be true. In our everyday language, it doesn't really even come up because we normally only deal with implications where a is true.
  4. Nov 4, 2011 #3
    Ok that makes sense but then I dont really see why we call it implication at all since its not really implication in the normal sense (where a is a sufficient condition for b), but just a seemingly arbitrary convention...
  5. Nov 4, 2011 #4
    No, it is implication in the normal sense. It's just that the normal sense is sort of incomplete. It doesn't cover all possible truth values for a and b.
  6. Nov 4, 2011 #5
    I'm not a logician either but I have studied a fair amount of it and I believe we have these things called disjunctive normal form and conjunctive normal form, and if I am recalling this correctly this has something to do with the fact that all statements in classical logic can be reduced to or's and and's, and negation of course. The conditional connective adds no extra meaning beyond its representation in terms of or's and and's, the fact that it is the formalization of our experience of causation makes it a useful, but ultimately superfluous, psychological crutch, as classical logic admits no temporal component.
  7. Nov 4, 2011 #6
    Yes, implication can be phrased as (a AND b) OR NOT a. So, yes, it's a psychological crutch. Not strictly needed.

    So, actually, not only is it implication in the normal sense, but the whole point is that it is implication in the normal sense, but since it's a technical version of it, we have to decide on what it means precisely, and that leads to weird things that we wouldn't normally say outside of a math/logic context.
  8. Nov 4, 2011 #7
    Why cant we just not assign truth values for the implication for the truth values of a and b that arent the actual case (since they are pretty much meaningless as the state of affairs is not the actual case), i.e., leave it incomplete?

    I suppose that if we actually want to assign truth values to the implication with all possible truth values for a and b and that we assume that the only case where a => b can be false is if a is true and b false, then it obviously follows that, with the exception of a being true and b false, all truth values for the implication are "vacuously" true since they cant be false.

    I think I now understand what homeomorphic means by when he says its just silly because it seems to be pointless anyways.
  9. Nov 4, 2011 #8
    If pigs can fly then I'm the King of Siam.

    That's a true statement. When you understand that you will be enlightened.

    Hint: Tell me how that statement could possibly be false.
  10. Nov 4, 2011 #9
    Well before I may have objected that how does pigs flying cause the fact that you are the King of Siam. Well instead I say your statement is either true or false just for the sake of assigning it a truth value then it cant be false because it is false that pigs are flying and you are not the King of Siam so its true. Is this correct?
  11. Nov 4, 2011 #10

    Stephen Tashi

    User Avatar
    Science Advisor

    I think the simplest way to understand vacuuous implications is to consider what constitutes a valid counterexample. If someone claims "A implies B" is true then what must we do to have a counterexample? We must show a situation where A is true and B is false.

    So the negation of "A implies B" must be "A and (not B))". The negation of that negation must be equivalent to the original statement "A implies B". By some uncontroversial logical manipulations we can show "not ( A and (not B))" equivalent to "(not A) or B".

    What is sufficient to make "(not A) or B" true? One thing that works is to have A be false. This is why vacuuous implication works. If you don't like vacuuous implication, you could try to define a new form of implication, but what are you going to do for its truth table? You have to say whether "A implies B" is true or false in the cases when A is false.

    Suppose you say "A implies B" is false when A is false. Then when you are being considered for the Fields Medal because of your masterful proof of the theorem that "X implies Y", someone will point out that your theorem is false. His countexample will be to show a case where X is false.
  12. Nov 4, 2011 #11

    D H

    User Avatar
    Staff Emeritus
    Science Advisor

    It's a lot more than a matter of convention. It is a very necessary concept. Suppose we know that (a) P implies Q and (b) P is true. We therefore also know Q is true; it cannot be false. This is the rule of modus ponens:

    If P then Q
    Therefore Q.

    This gives us half of the truth table for implication:

    Code (Text):
        |  Q
     P→Q| 0  1
       0| ?  ?
     P  |
       1| 0  1
    Knowing that P→Q tells us nothing about Q in the case that P is false. Logically it has to be this way; just because P causes Q doesn't mean other things can't cause Q as well, even if P is false. Thus those two (so far) unknown values in the truth table must be the same value; otherwise P→Q would say something in the case that P is false. In other words, we have

    Code (Text):
        |  Q
     P→Q| 0  1
       0| X  X
     P  |
       1| 0  1
    So how to fill in that upper half? The answer lies in looking at what P→Q tells us about P when we know Q. Knowing that Q is false is the key. If P being true always causes Q to be true and we know that Q is not true then we can logically conclude that P is false. This is the rule of modus tollens:

    If P then Q
    Not Q
    Therefore not P.

    Thus the P=0, Q=0 element of the P→Q truth table is 1. Since the P=0, Q=1 element of the table must have the same value as the P=0, Q=0 element, the truth table becomes

    Code (Text):
        |  Q
     P→Q| 0  1
       0| 1  1
     P  |
       1| 0  1
  13. Nov 4, 2011 #12
    I just said that because, in principle, you are free to assign any truth values you want to it, since it is a definition. It just doesn't make much sense, unless you use the usual definition.
  14. Nov 5, 2011 #13


    User Avatar
    Science Advisor

    i like to view P→Q as: Q contains P.

    so what if you're not in P? (P is false). that doesn't change the relationship of P to Q. the only way we can falsify P→Q is to be inside of P, but outside of Q. when we're outside of P, we might be outside of Q, we might not be, it's irrelevant to whether or not Q contains P.

    for example: if you are the President of the United States, you are an American citizen.

    the set of Presidents is a subset of American citizens. considering people who aren't President, doesn't change the truth of that relationship.
  15. Nov 5, 2011 #14
    Are you familiar with the truth table for implication? That's a good place to start.

    Another key element here is that you used the word "cause." AHA, this is the source of your psychological block. Logical implication is not causality. Nothing to do with causality. Put causality far, far out of your mind; and study the truth table for implication.

    What would be evidence that "If pigs can fly then I am the King of Siam?" is false? What would have to happen in order for that statement to be true? You almost sort of have it in your paragraph above.
  16. Nov 5, 2011 #15

    Honestly, just read this thoroughly, it's actually a really good page on it (which is strange for wiki).

    I think the most satisfying reason for this convention (and it is a convention, unlike what some people are suggesting) is the following (taken from the above page):

    and also this:

  17. Nov 5, 2011 #16
    Steve, just replace "true" with "false" in your sentence. Vacuous truth is a convention- it isn't clear that we must assign such things truth values as you seem to be suggesting. It just turns out that it makes more sense to use that convention rather than a different one (for many reasons, some listed above) but you could still have a perfectly legitimate logical system if you considered such statements as false.

    This makes it very hard to introduce people to it when they first see it- they ask "why?" and the answer is "it's convenient".
  18. Nov 5, 2011 #17
    Yes, this is all in some way an equivalent question to: "is the empty set a subset of all other sets?". Asking that is a vacuously true sort of statement: the empty set is contained in any other set if for all elements x in the empty set, they are also in the other set. There are no such elements, so this is vacuously true. Conversely, if you consider your logic as a sort of set theory, we will consider vacuously true statements as true precisely when we consider the empty set as being a subset of all others (note, if you disagree with vacuous truth, then you can't consider the empty set as being a subset of all others, which I imagine would be rather inconvenient).
  19. Nov 5, 2011 #18


    User Avatar
    Science Advisor

    Suppose your teacher tells you, at the beginning of the course, "If you get an "A" on every test, you will get an "A" for the course".

    You do NOT get an "A" on every test but do pretty well. Now consider these outcomes:
    1) You receive an "A" for the course. Was the teacher lying at the beginning of the course?

    2) You receive a "B" for the course. Was the teacher lying at the beginning of the course?

    Note that the teacher's statement, at the beginning of the course says nothing at all about what would happen if you did not get an "A" on every test.
  20. Nov 5, 2011 #19


    User Avatar
    Science Advisor

    It's important to note though in Hall's example that "If you get an "A" on every test, you will get an "A" for the course" as a logical proposition is not a statement about the teachers intentions. In that case he could be lying no matter of your performance if he had no intentions of giving you an A for the course even though you get A on every test.
  21. Nov 5, 2011 #20
    Thanks for the replies guys, I do see that it makes sense now and follows from the basic axioms of logic. I guess the problem was just that I was trying to reconcile it with my intuitive/concrete understanding of implication (an understanding based on casuality hence problematic).
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook