In 639-5, collections are not assumed to be "exclusive" (i.e., encompassing only entities within the semantic range denoted by the name of the collection but that are not coded as individual languages). An important application of ISO 639-2 is use in MARC, and in that context collections are treated as exclusive, however. So, depending on the application context, [ber] may or may not be assumed to include every Berber language.
Btw, you refer to "smaller Berber dialects". We need to be careful with our terminology: dialects are not coded in ISO 639-1/-2/-3/-5; dialects are in scope for coding only in Part 6. Presumably in this case you meant "smaller, less-developed languages" or something along that line.
As for the question of a macrolanguage, in principle a "Berber" macrolanguage could be coded, but [ber] should not be re-defined to change its scope from a collection to a macrolanguage: that would constitute some narrowing of scope, which would be problematic. If a "Berber" macrolanguage were to be coded, then it would need a distinct alpha-3 ID.
But before coding a new macrolanguage, I'd question the need. In particular, I question that there is _one_ Berber language (the correspondent describes themself as a speaker of "the Berber language"). It's entirely possible that the 25 Berber languages currently coded in 639-3 over-differentiate, but I strongly doubt that there are no real language divisions. (The fact that there are significantly different orthographies in use across the family points to a lack of unifying factors.) That makes me inclined to think that this person is a speaker of one of the languages already coded.
That said, one thing that can help maintain vitality of threatened languages is unification of closely-related languages, either through a shared orthography or through more aggressive socio-linguistic engineering. Over time, that can result, effectively, in a new language; but in the short term a coded macrolanguage entity may be useful in that it captures the ambiguity of there being multiple distinct languages while at the same time attempts are being made to communicate in just one form. But, before we go coding new entities, we ought to see evidence that there's a socio-linguistic reality out there that we're coding--i.e., we shouldn't code something just because some party would _like_ that to be a reality or likes to _believe_ it is a reality.
From: ISO 639 Joint Advisory Committee [mailto:[log in to unmask]] On Behalf Of ISO639-3
Sent: Thursday, July 07, 2011 2:23 PM
To: [log in to unmask]
Subject: Fwd: Is Berber eligible for a an 639 code as a macrolanguage?
Dear Rebecca, et al.,
The code [ber] is currently a collective designator under ISO639-2 for the smaller Berber dialects which are not otherwise coded. There are three individual ISO 639-3 languages in her list of Berber languages.
Please comment on his request for making a macrolanguage which includes the three languages plus the collective. If this is deemed to be a suitable candidate, I will request the appropriate forms to make the proposal.
Thank you for your help with this.
ISO 639-3 RA
7500 W. Camp Wisdom Rd.
Dallas, TX 75236
--- the forwarded message follows ---