In the "Issues to Resolve" doc that I made available early in January, there was an item related to the identifiers rm / roh (item 6.12). I know that the process we agreed on at the JAC meeting was that Milicent Wewerka would first review all of the items in this doc to see if there was anything that should have been given a different recommendation due to my limited knowledge of MARC usage, and that she is still working on that. In the case of rm / roh, however, I'd like to know whether the JAC could resolve this particular issue right away.


The issue is that rm / roh represents the name "Rhaeto-Romance", but it's not clear whether that's intended to refer to the individual language Romansh or the genetic sub-group of Romance that includes the three languages Friulian, Ladin and Romansh. 


Relevant facts:


The genetic sub-group that includes Friulian, Ladin and Romansh is referred to by some as "Rhaeto-Romance" (see, for instance, the second paragraph of the description at


MARC has used roh as a collective for Ladin and Romansh (according to the MARC Code List for Languages, 2003 edn; I note, though, that the doc at cites only Romansh for roh.)


Ladin is different from Ladino, for which there is an ISO 639-2 identifier: lad.


ISO 639-2 has a separate identifier for Friulian: fur.


There are several existing web and software implementations that have used rm for the individual language Romansh, including the Swiss gov't web site (see and MS Windows. Some other examples: 



ISO 639-1 lists the indigenous name represented by rm as "romontsch".


According to the info on the Swiss gov't web site (, the indigenous name for Romansh is "rumantsch" (although the Romansh version of the constitution uses "retorumantsch"); the English name is "Romansh", the German name is "Rätoromanisch", the French name is "romanche", and the Italian name is "romancio".



My recommendation is that the denotation of rm / roh should be the individual language Romansh, and that the English and French names should be changed to "Romansh" and "romanche" respectively. This is motivated particularly by existing usage of the ISO 639-1 identifier, rm. 


This recommendation would entail a change for MARC: they would have to discontinue using roh as a collective that includes Ladin as well as Romansh. An alternative would be to dis-unify rm and roh: rm would be the individual language Romansh, and roh would be a collective "Rhaetian languages"; a new alpha-3 identifier for the individual language Romansh would have to be added and equated with rm.


The other alternative is to say that rm / roh denote a collection, but that would be problematic for existing implementations.







Peter Constable

Globalization Infrastructure and Font Technologies

Microsoft Windows Division