Foundations: Sample Spaces, Events & the Axioms

the algebra of events (union, intersection, complement)

Because events are sets of outcomes, we can combine them the same way we combine any sets, and each set operation translates into a plain-language statement about what happens. There are three core moves. The union of A and B, written A ∪ B, is the event 'A or B (or both) occurs' — it gathers every outcome that is in A, in B, or in both. The intersection, written A ∩ B, is the event 'A and B both occur' — only the outcomes lying in both sets. The complement of A, written A^c (or 'not A'), is the event 'A does not occur' — everything in the sample space outside A.

Reading the translation table the other way is what makes this practical: the word 'or' in a probability question becomes a union, the word 'and' becomes an intersection, and the word 'not' becomes a complement. A small picture helps — imagine two overlapping circles inside a rectangle (the sample space). The whole shaded pair is A ∪ B; the lens where they overlap is A ∩ B; the part of the rectangle outside circle A is A^c. From these three you can also build differences, like 'A but not B' = A ∩ B^c, the part of A that misses B.

These operations obey the familiar laws of set algebra — they commute (A ∪ B = B ∪ A), associate, and distribute (A ∩ (B ∪ C) = (A ∩ B) ∪ (A ∩ C)), and the complement of the complement gives you back A. This matters because it lets you rewrite a messy event into an equivalent, easier form before computing its probability. The algebra of events is the grammar; the axioms of probability, coming next, are the rules for attaching numbers to whatever sentences this grammar lets you write.

Roll one die. Let A = 'even' = {2,4,6} and B = 'at least 4' = {4,5,6}. Then A ∪ B = {2,4,5,6} (even or at least 4), A ∩ B = {4,6} (even and at least 4), and A^c = {1,3,5} (not even).

Or = union, and = intersection, not = complement — the three translations you reach for constantly.

Probabilities do not simply add across a union when events overlap: P(A or B) = P(A) + P(B) - P(A and B), because the overlap would otherwise be counted twice.

Also called
event operationsset operations on events事件運算集合運算