The use of different perspectives on a problem is a very powerful principle in statistical physics, and has been especially important in mathematical physics. I will illustrate this theme with old and recent applications. These include the interpretation of QFTs at imaginary time as statistical fields, the relation of statistical fields to lattice models such as percolation, polymers, or spanning forests, the bosonization of fermionic theories, and many others. These connections are useful in both directions, and often provide complementary insight on a problem, very difficult understand without them.

Recording of Talk: http://www.kaltura.com/tiny/0c98q