What's the purpose of going away on a honeymoon after you get married? The ideas of the Reverand Bayes provide an answer: not his religious ideas, but his mathematical ones.
You have just started married life, and you don't know what married life is like. You have to learn about it. Bayes suggests how you do this learning - and considerable evidence now suggests this is what happens in the brain.
In a Bayesian framework, you first need to set up a belief (a "prior") about what it's like to be married. Then, any subsequent new information about married life is compared to your prior belief, and is used to update that belief. This new belief now becomes the prior belief for a new round of learning, and the cycle continues throughout married life. Crucially, the start point (your first prior belief) is incredibly important: if it's good then you tend to ignore bad information but still readily learn from good information.
The point here is that honeymoons are great (often long and luxurious holidays!) - and this helps set up a really positive prior belief with which to start married life. Your subsequent learning about married life is through rose tinted spectacles, you'll tend to ignore new negative information and find your positive belief confirmed by positive new information!
OK, so that might be the point of a honeymoon from a Bayesian perspective, but does this have any implications in the real world? It does! If you're planning a honeymoon, make sure you go immediately after getting married to set up the best possible prior for married life. Don't start married life back at work (so setting up a more mediocre prior, unless you love work) and wait for a honeymoon!
Given what we know about the neuroscience of learning, how plausible is this? Well, Bayesian ideas have been hugely influential in explaining the brain, economics and human behaviour in general. To find out more about the Bayesian Brain, have at this great (and short) talk here. And for a picture from my recent honeymoon, have a look above.