According to Christopher Tolkien (SD/439), “Lowdham’s Report on the Adunaic Language” from the 1940s (SD/413-440) was the last substantial work his father wrote on the subject of Adûnaic. As pointed out by Andreas Moehn (LGtAG), this does not mean Tolkien made no further changes to the language in the decades that followed. There are a number of later Adûnaic names and words that are not consistent with the grammatical and phonetic rules presented in Lowdham’s Report. The discussion here uses the term Late Adûnaic to describe these later variations.

In Lowdham’s Report, Tolkien said that the vowels [ē] and [ō] could only be long in Adûnaic (SD/423). There are a number of later examples in which short e and o appear, however: Zimraphel (UT/224), Belzagar (UT/222) and roth with its derivatives nadroth, obroth, Rothinzil (PM/376). The last of these words appeared in earlier writings with a long ô, Rôthinzil (SD/403), which is a strong indication Tolkien changed his mind about short vowels in Adûnaic.

Another phonetic difference appears in the later name Agathurush (UT/263). The sound “sh” [ʃ] is not a phoneme used in Adûnaic according to Lowdham’s Report (SD/418). From The Lord of the Rings appendices, we know that this sound appears in the daughter language of Adûnaic, Westron (LotR/1120), along with “ch” [tʃ]. Elsewhere, there is a note (PE19/75) indicating that “ch” was used for the Númenórean pronunciation of Quenya ty [tʲ], so perhaps Tolkien introduced these two sounds into Adûnaic.

The words nadroth “hind-track” and obroth “fore-cutting” appear to have prepositional elements as prefixes. In Lowdham’s Report, prepositions were usually suffixed elements (SD/435). Finally, the objective case was quite prevalent in compound names in Lowdham’s Report, but is largely absent in later names. The only counter-example, Balkumagân (PM/151), was coined in the early 1950s.

There isn’t enough information to know what, if any, concrete changes Tolkien made to Adûnaic in his later writings. The only way to construct a remotely consistent Adûnaic grammar is to use the rules outlined for the Middle Adûnaic in Lowdham’s Report. This requires us to ignore or explain away the changes in Late Adûnaic. Here are some (entirely invented) explanations for these inconsistencies:

• Phonetic inconsistencies (sh, ch, short e and o) could be explained by assuming these later forms are partially Westronized, no longer entirely consistent with the phonetics rules of Classical Adûnaic.
• In the case of nadroth and obroth (PM/376), these words appeared in a document exploring the possibility that the element -ros in the name S. Elros was derived from the Beorian language, so perhaps these words are also of Beorian origin. Beorian was an older “cousin” of Adûnaic rather than a direct ancestor language, so it could have different phonology and syntax. If these prepositional elements were used in “proper” Adûnaic, they might appear in the normal Adûnaic order: *rôthnad and *rôthôb.

These inventions could be used to reconcile later Adûnaic words with the phonetics and grammar of Lowdham’s Report, though it is extremely unlikely that they represent Tolkien’s actual thinking on the subject.