One of the most important, if not the most important, problem in modern cosmology is the problem of assigning probabilities to cosmological observations such as the value of dark energy or the spectrum of Cosmic Microwave Background (CMB) radiation. In the last few decades a number of promising attempts were made to resolve the problem in the context of inflationary cosmology and/or using the anthropic principle. Despite of remarkable phenomenological success of inflation, the more fundamental issues remain largely unexplored. In this project we would like to study the logic involved in deriving cosmological predictions or what we shall call cosmic logic. The first step would be to define a physical observer and to figure out what kind of object it is (classical or quantum, local or non-local, Markovian or non-Markovian, etc.) The second step is to formulate a set of axioms on which the reasoning in cosmology should be based. This would enable us to analyze a variety of problems posed in the literature such as Youngness paradox, Boltzmann brains problem, Q-catastrophe, Guth-Vanchurin paradox, etc.