For example, suppose you wish to say
This particular example is true, because 5 is a natural number, and when we put 5 in for x, we get 5^{2} = 25, which is true. In contrast, "For some natural number x, x^{2} = 2" is false, because 2 is not a perfect square. However, "For some real number x, x^{2} = 2" is true, because we could put √2, the principal square root of 2, in for x. We could also put in √2 for x in this last case; it's no problem that there is more than one solution.
The form of an existential quantification
Every existential quantification statement uses a dummy variable[?], such as the x in the above example. x could be replaced with any other symbol, and the meaning of the statement wouldn't change, as long as that symbol isn't being used anywhere else. (In a less formal context, we might use a pronoun instead; for example, we could say "For some natural number, its square is equal to 25"; here, "it" takes the place of the dummy variable. Even less formally, "Some natural number has a square equal to 25".)
It's also necessary to specify the universe of discourse of the dummy variable. This allows us to capture the difference between, for example, "Something is evil" ("For some x, x is evil") and "Some human is evil" ("For some human x, x is evil"). In our example, the universe of discourse was originally the set of natural numbers; later we changed it to the set of real numbers. It's not necessary for the universe of discourse to be a set in terms of formal set theory, and existential quantification can be used in logical theories that don't make any reference to sets. Indeed, ordinary predicate logic usually takes the universe of discourse to be everything under discussion, at least at the fundamental level.
We said that the statements to be covered by an existential quantification must fit a certain template. Call this template P(x). That is, P(x) is a specific statement about x; formally, it should be a predicate in one variable. In the original example above, P(x) is the statement "x^{2} = 25". P(x) is known as the (I don't know what it's called; can anybody help out here?). By the way, it's possible that P(x) doesn't mention x at all. You could take P(x) to be "The sky is blue" and say "For some natural number x, the sky is blue". This is usually a silly thing to say, since there's no reason to mention x in that case; you could just say "The sky is blue" and be done with it. Nevertheless, such a case may come up in a degenerate situation.
Several phrasings are used for existential quantification, such as:
Informally, the "for some x" or "∃x" might well appear after P(x), or even in the middle of it if it's a long phrase. Formally, however, the phrase that introduces the dummy variable should always be placed in front. (In this example, the universe of discourse is everything under discussion, the most fundamental universe of discourse.)
How you restrict the universe of discourse depends on the nature of the underlying logical theory.
First, you might be using a theory with more than one type. For example, suppose that you're studying a formal set theory with urelements[?] (objects that are not sets), rather than ordinary set theory (where everything is a set). Then you could have two types of objects, sets and urelements, which might be indicated differently in your notation. For example, suppose that urelements are given by lowercase letters, while sets are given by uppercase letters. Then the statement
If you're using set theory (as almost all of mathematics does), then you can use sets themselves to define the universe of discourse. For example, if N is the set of natural numbers, then
Additionally, you can use logical conjunction to further restrict a universe of discourse to only those x such that some statement Q(x) is true. The statement
This can be used to cover both of the cases above. For example, to deal with urelements in untyped predicate logic (the ordinary kind), you just need a logical predicate[?] that says whether something is a set or an urelement; say, Sx iff x is a set. Then to say that A has at least one member that is an urelement, write
We need a list of algebraic properties of existential quantification, such as distributivity over disjunction, and so on.
Search Encyclopedia

Featured Article
