This excision seems wrong, as it would disallow conversions that are now permitted. At the least, it seems non-editorial.

@W-E-Brown How so? Given a type const int, it can be represented as T, where T is const int, or cv T, where cv is const and T is int.

The only context where this would present any kind of questionable difference is:

If T2 is an inaccessible or ambiguous base class of T1, a program that necessitates such a cast is ill-formed.

However, we ignore cv-qualifiers in the determination of if a type B is a base class of D, so there would be no difference in effect.

That argument is beyond me. Fewer words does not always produce a better specification.

We routinely go out of our way to indicate when a rule applies to a possibly cv-qualified type. The proposed change seems to go in the other direction and thus obscures, rather than illuminates.

We use the cv notation when we need to establish constraints based what they are, (e.g. the cv-qualification must be the same, or cv1 must be less cv-qualified than cv2). If anything, using the cv notation and then not using the resulting "variables" we introduce is more confusing, since we generally only use the notation when we care about the cv-qualification. Personally, when I see cv I automatically assume that it will be used to enforce a restriction.