This paper attempts to contrast the basic elements of statistical learning theory and statistical decision theory, but I'm still confused about how the two are related.
Have you read the Wikipedia articles? Decision theory is a subset of or problem in statistical learning, in my view; both driven by statistics -- data. It concerns the optimal making of decisions such as choosing between alternatives, once or over a period of time (possibly without termination), or deciding when to stop an experiment, etc. Statistical learning theory is a much broader concept. It is the most popular paradigm for machine learning. Maybe you know that?