You can do this by unpivoting the data and using group by
:
select col
from (select a as col from table_a union all
select b from table_a union all
select c from table_a union all
select d from table_a union all
select a from table_b union all
select b from table_b union all
select c from table_b union all
select d from table_b
) t
group by col
having count(*) > 1;
If the table is large, there are more efficient ways to unpivot than using union all
.