Yes: check whether the original or new column are null, then assign one to the other as appropriate:

begin
  if :new.col1_new is null then
    :new.col1_new := :new.col1;
  else
    :new.col1 := :new.col1_new;
  end if;
end;

Doing this does have drawbacks though:

- You double your storage requirements for this column
- Doing this in a trigger adds overhead, slowing down DML
- This only works for insert; for UPDATEs you need more advanced checking as you may be setting the column to null
- The values could get out of sync (e.g. because someone disables the trigger); add a check constraint to ensure they're equal to avoid this

You can avoid these by using views and renaming the column in the view definition.

This could either be a view that renames the table:

create or replace view t_new as 
  select c2, ..., c1 c1_new from t;

When you update the code, you point it to the view and the renamed column.

Or you could:

- Rename the table (e.g. with the same name but in lowercase)
- Create a view with the original table name that selects both the original

rename t to "t";
create or replace view t as 
  select c1, c2, ..., c1 c1_new from "t";

With either view approach, you're accessing the same column. This avoids problems that can arise from duplication (extra storage, desync, etc.)

Whatever you do it adds complexity.

So it's worth asking: are the current column names so bad that renaming them is worth the cost of updating all the code?