Thank you for your patience and guidance.

> Just define the new face to inherit from `region', so as to preserve
> the old behavior:
>
> (defface semantic-idle-symbol-highlight
>   '((t :inherit region))
>   "Face used for highlighting local symbols.")

It's weird that the region face would be used for this, because it has
nothing to do with the region. The fact that it can make me think I
have a region marked when I don't is a motivator for this bug report.
My primary concern is that I can customize it personally, so I revised
the patch to have the face inherit region.

> Not quite what I had in mind (taking something that was a variable
> and turning it into a face alias won't work right, will it?).

I probably misinterpreted your earlier tip. I looked off of
diary-face:

  (define-obsolete-face-alias 'diary-face 'diary "22.1")

I tested the last patch before sending it and it worked. Although that
patch used the defface variable and not the obsoleted one.

> Then these bits:
>
> -                     region semantic-idle-symbol-highlight-face)
> +                     region 'semantic-idle-symbol-highlight)
>
> are not necessary (or appropriate).

I don't know why using obsolete variables is most appropriate, but
I've done as requested in the revised attached patch.