I have followed with interest the PCCList discussion on what to do about a person who writes under both her real name and a pseudonym when the institution needs for its collections only one of the names.
It seems to me that the question is answered by LCRI 22.2B, Multiple Headings--Contemporaries, 3) ... Note: Establish separate headings only for pseudonyms used (or likely to be used) on separately published works. Although the name in question here is the real name, the principle remains--establish all names used (or likely to be used) on published works. We know the real name has been used from the RLIN record. Therefore, both names should be established, each connected with 500 fields.
Providing a 400 field reference from the name not established but used on a published work(s) only asks for problems when the authority record is downloaded into a system with authority control and the reference therefore conflicts with a heading on the bibliographic record.
The suggestion of adding a 667 field with the text: The following is a valid AACR2 heading: should also not be followed. That particular text in a 667 field was used exclusively on retrospectively converted headings that followed earlier rules of subsuming earlier names under the latest name or of establishing only one name for writers using multiple names.
A possible middle ground would be to create the authority record for the name needed, citing the other name(s) used in the 670 field(s), followed by [no publs. in [name of institution or database]]. This follows what may be done with corporate entities that have earlier names that are unlikely to be needed. However, this is not quite the same thing. A person can use any name at any time, while corporate entities basically abandon earlier names.
I believe the NACO participants have an obligation in helping to create the national authority file by doing all the work necessary according to the guidelines to which they agreed, so feel establishing the other name(s) used is the proper course of action.