Consider implementing likelihood() as log_likelihood() and wrapping with logLik()

[jamesmbaazam]

I'm now wondering if we should rename the function log_likelihood and remove this argument as in offspring_ll, especially if we eventually want to wrap this by an S3 function for epichains or related objects with the standard name logLik. The only difference is the exp call at the end but the user could also really do that themselves externally.

It would have the benefit that the mentions of "log-likelihood/likelihood" would be shortened to "log-likelihood".

Originally posted by @sbfnk in #58 (comment)