In these lectures, I want to approach the subject from the third point of view. We shall therefore always try to have in mind the questions that statistical mechanics tries to answer when critical phenomena are discussed. These include (a) the classification of universality classes, and, equally important, the development of criteria to decide whether a given system is in a particular universality class; (b) the calculation of critical exponents; (c) the calculation of other universal amplitudes, including finite-size effects at criticality, and ratios of thermodynamic quantities away from criticality; and (d) the calculation of correlation functions, both at and away from criticality.