[b said:
Quote[/b] (nepenthes_ak @ July 20 2006,11:06)]uhmmm, how do I use these in writing a proof though, I thoght I under stood but I guess I just don under stand how IM suposed to write a proof with it I guess.
Give me an example of what you're trying to prove. I'm not sure where your misunderstanding is.
Generally speaking, proofs work by changing your premises (the axioms, or rules of your system, and your assumptions, or the ideas that you're exploring in the system) into theorems (conclusions about your system) through an operation called
direct consequence. The direct consequences of a theorem are the other theorems that are a single proof-step away.
So, say that I know that if theorem P is true, theorem Q is also true. This might be written symbolically as
P -> Q, or
P implies Q. If I then assume that P is true (written
P; P is false would be something like
-P or
not P,) then a direct consequence of
P and
P -> Q is
Q, because if P were true and Q were not, then it would not be true that P implies Q. This direct consequence follows from the formal definition of implication in symbolic logic. Symbolic logic, AKA mechanical logic, has strict, unwavering definitions of truth and falsehood for all the basic logical operations; And, Or, Implies, Not, For All, and others. You can find these definitions online. The nice thing about symbolic logic is that there's no guesswork to it; it's called mechanical logic because a machine could do it, exploring all possible theorems by manipulating one string of logical symbols into another according to a table of rules. These manipulations are called
Rules of Inferrence.
Logic isn't all operations, though. Like the Ps and Qs we need variable values to represent the notion of truth and falsehood. There are
propositions (which is what P and Q are in the example above) which hold the value of true or false and often stand in place for ideas about the properties of something. For example,
P could stand for "Joe grows carnivorous plants" and
Q for "Joe grows plants." Then
P -> Q could be interpreted as "If Joe grows carnivorous plants, Joe grows plants." You're probably also using
predicates, which are like propositions that are more generic. Predicates take abstract objects, referred to as
variables, as input and return truth or falsehood. For example, equality is a
binary predicate, meaning that it takes two variables and decides on truth or falsehood depending on the values given. In the case of equality, truth is returned only when the given objects are identical. We could have a 'grows' predicate
G, and if I were object
j and carnivorous plants were object
c, then we could write
G( j, c ) or
Gjc for "Joe grows carnivorous plants." There are also predicates that have one, three, four, or even undefined amounts of variables. In any case, prepositions and predicates by themselves are referred to as
atoms; when logical operators are used to join them, the resulting theorems are called
molecules. Written out entirely in operators of formal logic along with predicates and propositions, theorems are known as
formulas, in the same way that y = x + 1 is a formula of algebra.
So a proof proceeds by starting with an assumption and axioms, all in written in terms of logical formulas, and then, by using rules of inferrence, you create a chain of direct consequences that lead to your conclusion. Oftentimes, however, you start with information that isn't in the form of formal logic. From here, you have two options. You can translate your axioms, assumptions and wanted conclusions into symbolic logic by formally defining all of the predicates and theorems involved. However, this is often an impractical way to proceed; you must rigidly define the truth values for every predicate, formalize all the relevant axioms, and after all that work the mechanical deduction may be very complicated and/or drawn out.
The other option is to conciously practice the virtue which the method of symbolic logic gives us for free, which is to leave no possibility unexplored or vaugely argued. You don't use logical operators in terms of the symbols and rules for manipulating them, but it helps to model your arguments in the same way, because the rules of inferrence all come from readily evident logical truths. If your proof is about some object, you must make sure to deal with every possible type of object it could be. So, if I'm trying to prove some property of square numbers, I must make sure that my deductions are valid for even squares as well as odd squares. This basically amounts to not making unnecessary assumptions. Never be satisfied with, "It
can't be this way," or, "It
must be that way." Every argument, every property or circumstance you discuss, must have a valid and obvious justification. It may seem stupid to only work with ideas that are obvious consequences of one another, but after three or four steps, you can get to some pretty un-obvious conclusions. At first, doing proofs in this manner is very difficult, but if you can practice going between formal proofs (with symbols and the rules of inferrence) and 'real' proofs (the informal, longhand proofs that most math uses) then the most complicated part of proofs becomes writing them out in English in such a way that people won't misinterpret your grammar.
Hope that helps some.... I know it's a lot to take in all at once.
Best luck,
~Joe