In Sindarin and Noldorin, the first consonant of a pair of voiced stops became a vowel, but the exact phonetic development evolved between the two conceptual periods. The clearest description of the Sindarin change is in the Outline of Phonology (OP2) from the early 1950s:

In Sindarin on the other hand [d+d >] zd > d with lengthening of preceding vowel so early that these newer long vowels followed the same development as the original long vowels (ā > ǭ/au; ē, ī > ī; ō, ū > ū). The intervocalic d then > đ (dh) as usual. bd > u̯d > u̯ð and gd > ʒd > ı̯d > ı̯ð: the diphthongs so formed also followed the development of original diphthongs (aı̯ > ae; iı̯ > ī; eı̯ > ī; oı̯, uı̯ > ui) (PE19/91).

A number of examples followed this note, such as:

• ✶khabdā [> *χaudā] > S. haudh “funeral mound” (PE19/91).
• ✶reddā > rezdā [> *rēdā] > S. rîdh “sown field” (PE19/91).
• ✶negdē [> *neɣdē > *neidē > *nīde] > S. nîdh “juice” (PE19/91).
• ✶snagdē [> *snaɣdē > *hnaiðe > *hnaeð] > S. naeð “wound” (PE19/91).

In the case of [d+d], this combination had already become [zd] in Primitive Elvish, and this [z] vanished while lengthening the preceding vowel. As noted above, [g] first became voiced velar spirant [ɣ] (or ʒ as Tolkien generally wrote it) and then [i] to form a diphthong with the preceding vowel. The labial stop [b] probably developed similarly: first becoming the voiced labial spirant [β] and then the vowel [u] forming a diphthong.

Note that there was another similar spirantalization and vocalization of [g] > [ɣ] > [i] before liquids and nasals ([r], [l], [m], [n]). It is tempting to assume the two changes were part of the same process. They must have occurred at distinct stages, however, as indicated by the development of the diphthong [ei]. From the earlier vocalization before voiced stops, [ei] became [ī] as did the original primitive diphthong [ei], but in the later vocalization before liquids and nasals, [ei] became [ai] (in final syllables) as was the case wherever this diphthong developed after the Old Sindarin period. Compare:

• ✶negdē [> *neɣdē > *neidē > *nīde] > S. nîdh “juice” (PE19/91).
• ✶keglē [> *keɣle > *keile] > S. cail “fence” (UT/282).

All of the examples Tolkien provided are of voiced stops followed by [d], and it is unclear how combinations with a following [b] or [g] would have developed; such combinations would in any case have been quite rare. I think it is likely that [b] and [g] would have vocalized to [u], [i] before any kind of voiced stop, as above. In the case of [db] and [dg], however, we know that [zb], [zg] ultimately became [ðβ], [ðɣ] (PE19/101), so it seems likely that d would similarly have developed into ð before b, g and ultimately the following stop would also have spirantalized and vocalized as with the development of z + b, g. Since we have no explicit examples, it is hard to say for sure.

Conceptual Development: There isn’t enough information for us to determine how combinations of voiced stops developed in the Gnomish and Early Noldorin of the 1910s and 1920s. By the 1930s, however, there are enough examples in The Etymologies to deduce similar patterns of stop-vocalization in the Noldorin period. For example:

• ᴹ✶labdā > N. laudh “gluttonous eating” (PE19/45).
• ᴹ✶libda > N. glúð “soap” (Ety/LIB²).
• ᴹ✶negdē > N. nīdh “honeycomb” (EtyAC/NEG).
• ᴹ✶khagdā > N. hauð “mound, grave, tomb” (Ety/KHAG).

The first two examples indicate that the Noldorin and Sindarin developments were similar, but the last two shows that the development of [gd] was at least somewhat different. Compare ᴹ✶khagda > N. hauð to the Sindarin development from above:

• ✶snagdē [> *snaɣdē > *hnaiðe > *hnaeð] > S. naeð “wound” (PE19/91).

In Sindarin the development was [ɣ] > [i] so that [agd] > [aɣd] > [aið] > [aeð]. I am of the opinion that in Old Noldorin, the [g] instead vanished while lengthening the preceding vowel before [d], so perhaps:

• egd > ēd: [negdē] > [nēdē] > [nīdē] > [nīð]; where later [ē] > [ī]
• agd > ād: [kʰagdā] > [kʰādā] > [xǭda] > [hauð]; where later [ā] > [ǭ] > [au]

There is a deleted form nēdh beside nīdh “honeycomb” (EtyAC/NEG), which lends some weight to this vowel-lengthening theory. However, we don’t have enough examples to know for certain.

In draft notes for the Outline of Phonology from the early 1950s, Tolkien first had khagdā becoming Noldorin-style {haeð} before revising it to hauð and updating its primitive form to khabdā to be consistent with the Sindarin phonetic rules given above (PE19/91-92, note #110). Perhaps Tolkien was still vacillating on the exact phonetic development over a decade later.

See this entry for a discussion on the conceptual development of the phonetics changes for [zd] in Noldorin vs. Sindarin.

GS/§4.24 GS/§4.25 GS/§4.26 WGHC/@@@