Relational generation of random data

To save lazy people like me from writing a class for this job; in a users document once `chance.name()` is used to generate `"Albert Hodges"` then the next call to `chance.gender()` should have the option to follow a previously used gender related method. Such as we sould be able to invoke `chance.gender({follow: "name"})` and then `chance.prefix({follow: "gender"})` kind of.

Start by reviewing the existing name(), gender(), and prefix() APIs and how generated values are selected. Define how follow relationships should be represented and resolved, including the requested gender({follow: "name"}) and prefix({follow: "gender"}) examples. Done should include relational generation that produces consistent related values and tests for the supported follow behavior.

Written by the indexing model from the issue text.