Context

Describe the background behind the problem.

During development of ARROW-197, adding Polars support, we noticed that `polars.DataFrame.to_arrow()` returned a field defined as`large_list<item: large_string>)`.
Found this out when attempting to call `_validate_schema`. The call failed because the "checkers" did not include `pyarrow.types.is_large_list()`.

Definition of done

What must be done to consider the task complete?

It turns out that only a few changes are needed. Most importantly, we did not need to create new Builders. We could use ListBuilder and StringBuilder. Their bodies pass most of the work down to pyarrow. We then use the original table's schema to ensure roundtrip integrity.

To be done, we'll add tests of the new types.

Pitfalls

What should the implementer watch out for? What are the risks?

It would be beneficial for us to discuss good corner cases, the 'large' cases in particular.