I'm talking specifically about probability theory. I was reading some stuff about probabilistic graphical models, and they kept switching the notation in this book, but I couldn't discern the difference by context.
One possible hypothesis is that they are subtly different, e.g., $P$ is a probability measure, while $\Pr$ is an unnormalized probability measure (or something; I really have no idea).
Best Answer
In general, $Pr(.)$ is used for the probability of an event whereas $p(.)$ is used as a probability measure.
That said, notation in statistics is pretty diverse (also known as a giant mess if you're a student), so always check the book you're reading.