People often seem to think that "A pattern is a solution to a problem in a context" is a definition of a pattern. But I think it is just an attribute of a pattern. When I say "my wife is tall" you don't take that as a definition! And so with patterns. Patterns have many characteristics. Some others are that they repeat, you can copy them, and when you copy one you usually will not copy it exactly.  

I interpret Alexander as saying that there is a reason why a pattern repeats. Patterns mean something. If patterns are caused by humans, there are human reasons for them. People do the same thing over and over because they face recurring problems, and discover that a particular solutions works for these problems. If we want to understand a pattern, we have to understand the problem it solves, not just the elements that repeat.  

A pattern is both in the world, in our heads, and on paper. A pattern on paper is a description of the pattern in our head, which generates (or maybe it is our understanding of) the pattern in the world. "Solution to a problem in a context" can also be interpreted as telling us how to write patterns. A written pattern ought to make both the problem and the solution clear. But it is still not a definition.

I admit that I've used "solution to a problem in a context" as a definition, but I think I'll go change my PowerPoint slides.

-- RalphJohnson
-----
...change them to what? -- AlistairCockburn

What about BradAppleton's one-sentence-definition at the end of PatternDefinitionThread? Or maybe you're really trying to say a pattern is something that's UsefulUsableUsed.
-----

Of course your wife is tall. You never would have noticed her otherwise, from up there. ;-> -- RonJeffries