According to Christopher Tolkien (SD/356), Adûnaic first appeared in the draft of “The Drowning of Anadûnê” (discussed on SD/331-413), his father’s second iteration on the story of the fall of Númenor. This tale was a companion to two other related documents, the story “The Notion Club Papers” (SD/145-327) and “Lowdham’s Report on the Adunaic Language” (SD/413-440), the latter being a partial description of the phonetics and grammar of Adûnaic. These works were composed in the mid-1940s, during various breaks J.R.R. Tolkien took from working on The Lord of the Rings.

As suggested by Andreas Moehn (LGtAG), Adunaic went through three phases of conceptual development, which he labeled k-Adûnaic, d-Adûnaic and a-Adûnaic (after three distinct words for “Earth” in each phase: kamāt, dâira and aban). The discussion here labels the phases: Draft Adûnaic, Middle Adûnaic and Late Adûnaic.

The first phase, Draft Adûnaic, is reflected in the drafts of the three works mentioned above. In this phase, Adûnaic was a European-like inflected language with five noun cases (Normal, Subjective, Genitive, Dative and Instrumental). Christopher Tolkien published some fragments of grammar from this phase (SD/438-439), and more information can be gleaned from Adûnaic words and phrases appearing in the drafts of the two stories.

As J.R.R. Tolkien revised the stories, he made dramatic changes to its grammar. The final versions of the 1940s stories reflect this Middle Adûnaic phase, in which the “Semitic” character of the language emerged (SD/415). This was likely due to the influence of Dwarvish on Adûnaic (SD/414), which Tolkien also described as “Semitic” in nature (PE17/85). The noun case system was simplified (to Normal, Subjective and Objective cases) and made rather non-European in its use. Changes to the verb system are harder to analyze, since verb conjugation was not fully described.

The bulk of the information on Middle Adunaic grammar comes from Lowdham’s Report, which details the phonetic development of Adûnaic and most of its noun declensions. It was abandoned before Tolkien described other parts of speech, with only a few notes on verbs remaining (SD/439). In “The Notion Club Papers”, there is a lengthy text (SD/246-7) referred to here as the Lament of Akallabêth. Draft versions of the Lament also exist (SD/311-2), and comparing the draft to finished forms of the Lament gives useful insights to the conceptual changes from Draft to Middle Adûnaic. Andreas Moehn wrote a detailed analysis of these changes (LGtAG).

Christopher Tolkien stated that his father did no further work on Adûnaic after setting aside Lowdham’s Report (SD/439). However, (as pointed out by Andreas Moehn, LGtAG), J.R.R. Tolkien’s later Adûnaic examples are not entirely consistent with Adûnaic as described in Lowdham’s Report. There are enough differences in later Adûnaic names and words in the published Lord of the Rings and Silmarillion to believe there was some further conceptual development in Adûnaic, which the discussion here collectively labels Late Adûnaic.

Draft and Middle Adûnaic were contemporaneous with Middle Quenya and Noldorin, while Late Adûnaic was contemporaneous with Late Quenya and Sindarin. We have so few examples of Late Adûnaic that it is difficult to say much meaningful about this conceptual phase. See the entry on conceptual-changes-in-late-Adûnaic for the details.

Rather than separating the conceptual phases into different languages like the various periods of Quenya and Sindarin, the discussion here lumps all of them together as “Adûnaic”. Furthermore, almost everything we can say about Adûnaic grammar is from the Middle Adûnaic as described in Lowdham’s Report. This was not Tolkien final vision of the language, but it is the best we can do with the information available. Since Christopher Tolkien said that his father abandoned work on Adûnaic at that point, it is unlikely that there is some major new work on that language that remains unpublished.