I checked, and here is the problem. Given the age of the original post, I'm pretty sure that
SQL 92 was the standard applicable at that time. Might still be right for ACE but I haven't looked. I was more interested in the syntax issue than details of which standard is in force now.
http://www.contrib.andrew.cmu.edu/~shadow/sql/sql1992.txt
From the SQL92 language specification, it appears that a "comment introducer" is referred to with that dreaded phrase "implementation dependent." I.e. It is possible that Access SQL (or really, JET SQL/ACE SQL) has the option to pick ANYTHING THEY WANT as the comment marker and as long as it doesn't conflict with the rest of the syntax, it can be anything. And that EXPRESSLY includes "nothing."
So what works in SQL Server doesn't have to work in Access / JET / ACE. The fact that
/* ... */ works in some SQL implementations but not in others? Not a problem. Not even SLIGHTLY a problem.
I think part of the issue is that Access's JET and ACE don't store sequences of queries as a stored procedure, whereas some of the active SQL engines (ORACLE, ShareBase, Sybase, etc. - then and now) allowed stored sequences of SQL statements to be triggered by activating the single procedure name.
JET and ACE do not store native sequences of queries. All queries are passed to them by the Access environment. And the Access environment uses macros or VBA to run query sequences. Therefore, comments are tagged on the querydef as an object (because all objects can have descriptions), or on the steps of a macro (because a description is allowed for each macro step), or on the VBA instruction that executes the OpenQuery or DB.Execute or whatever is launching the individual query.
Basically, the person asking the original question was "looking for love in all the wrong places."