Cantonese. The preferred code is yue, which is new addition from
[ISO 639-3]. Before [ISO 639-3] was published (in 2007) and [RFC 5646]
was published (in 2008) to allow them to be used in language tags,
Cantonese was regarded as a dialect of Chinese and encoded zh-yue;
yue is registered in [IANA Lang Subtag] as an extlang tag for this
purpose. Because historically not all systems have supported extlang
tags and because Cantonese is the main language spoken in Hong Kong,
zh-HK has been used to represent Cantonese, even though this properly
means any Chinese dialect localised for Hong Kong.
Catalan_Spn. Presumably this represents the Catalan language
(ca), as spoken in Spain, though [GEDCOM] is silent on this point.
The preferred code is therefore ca-ES. Why Catalan should be singled
out as the only language to be given a regional variant is unclear, as
Catalan is not widely spoken outside Spain. Perhaps the GEDCOM authors
considered Occitan (oc) and dialects of it such as Gascon and
Provençal, all of which are spoken in Southern France, to be forms of
Catalan.
Mandarin. The preferred code is cmn, which like yue for
Cantonese is a new addition from [ISO 639-3]. Previously it was encoded
zh-cmn, where cmn is registered in [IANA Lang Subtag] as an
extlang tag for this purpose, and zh refers to the family of Chinese
dialects or languages. However because Mandarin is the most widely
spoken Chinese language, as well as the official language of China, zh
is often used to refer specifically to Mandarin Chinese.
Norwegian. The preferred code is no, which is the code for the
Norwegian macrolanguage encompassing the two standard written forms of
Norwegian in common use today, Bokmål (nb) and Nynorsk (nn).
Neither is regarded as the official form of Norwegian, but Bokmål is
used by considerably more people people than Nynorsk.
Pahari. This is a group of around 20 languages spoken in Nepal and
northern India. The most widely spoken one is Nepali which has is
assigned its own language code in [GEDCOM]. Once Nepali speakers are
excluded, the majority of non-Nepali Pahari speakers speak one of the
Himachali languages, which collectively have the [ISO 693-5] code him,
which is the preferred code; however there are also significant numbers
of speakers of Kumaoni (kfy) and Garhwali (gbm), which are Pahari
languages that are not in the Himachali group.
Serbo_Croa. Serbo-Croat formerly had the language code sh in [ISO
639-1]. After the break-up of Yugoslavia, this was removed and replaced
with three separate codes: sr for Serbian, hr for Croat and bs for
Bosnian. Although sh no longer appears in [ISO 639-1], it does appear
in [IANA Lang Subtag] which does not remove old codes, and is therefore
the preferred code.