The diphthongs ei, ou usually simplified to long ī, ū in Quenya. This change was not universal, however, and these diphthongs occasionally developed into other sounds, but the conditions under which the variant developments occurred as well as the resulting sounds gradually evolved throughout Tolkien’s life.

In the Early Qenya of the 1910s and 20s the common phonetic development was ei, ou > ai, au, such as:

• ᴱ√EIKA > ᴱQ. aika “broad, vast” (QL/29).
• ᴱ✶sleiwa > ᴱQ. laiwa “pale” (PE13/149).
• ᴱ√LOURI > ᴱQ. laure “gold” (QL/51).
• ᴱ✶souna > ᴱQ. sauna “clean” (PE13/148).

One exception is ᴱ✶nēgittĕ > ᴱQ. neite “wax” (GL/60); this is the only appearance of ei in an Early Qenya word, and the diphthong may have been preserved because it was derived from a long ēi.

In the Compative Tables of phonetic developments from the 1930s, Tolkien introduced ī, ū as alternate developments for ei, ou, so that ei > ī or ai and ou > ū or au (PE19/25), but Tolkien did not describe the conditions under which the various sound changes occurred. Looking at contemporaneous examples in The Etymologies, we see the following phonetic developments for (undeleted) entries:

• ᴹ✶neı̯ti- > ᴹQ. níte “moist, dewy” (Ety/NEI).
• ᴹ√NEI̯ > ᴹQ. níre “tear” (Ety/NEI).
• ᴹ✶weı̯rē > ᴹQ. Vaire “moist, dewy” (Ety/WEY).
• ᴹ√NOWO > nau̯þe > ᴹQ. nause “imagination” (Ety/NOWO).

ᴹQ. aule “invention” might be another example of ow > au, since it seems it could be derived from either ᴹ√GAWA or ᴹ√GOWO (Ety/GAWA). A possible later example is Q. Sauron which in one place seems to have been derived from *þowson > þauson (PE17/68).

Deleted entries show both ei > ī (Ety/GEY) and ei > ai (Ety/EY). For weı̯rē > Vaire Tolkien said this sound change was “owing to wei > wai”. Likewise, in the somewhat earlier Declension of Nouns [Version 4] from the early 1930s, Tolkien said that “ei > ai after ı̯”, giving the instrumental plural form of nie as niainen [nie + inen] to illustrate this phonetic development (PE21/6). These two examples seem to show yei, wei > yai, wai, but that does not explain ᴹ√NOWO > nause.

In the Outline of Phonetic Development (OP1) from the 1940s Tolkien introduced a new paradigm:

ei remained in AQ, but became in PQt. ē when it bore the PQ stress. But in unstressed syllables (according to PQt. accent), and in final syllables other than monosyllabic words it became [ī̆] ... Similarly ou remained in AQ; but became in PQt. where still stressed ō. Otherwise it became ū̆ (OP1: PE19/53).

A similar development is described in a deleted note associated with the first version of Quenya Personal Pronouns (QPP1) from the late 1940s:

stress[ed] ei in most syll[able]s > é. unstress[ed] > ī - ou > ó / ū (PE23/87 note #85).

According to these rules, ei, ou > ī, ū when unstressed and > ē, ō when stressed, and these sound changes occurred in Parmaquesta [PQ]. Probable examples of this would be monosyllabic ᴹQ. tō “wool” < ᴹ√TOW (Ety/TOW) and ᴹQ. réte “female” [< *reiti-] derived from ᴹ√RĒ/ERE (PE23/87 note #85), with the second example revised from ríte. Tolkien described similar rules in the Outline of Phonology (OP2) from the 1950s:

ei remained in AQ, but became ē in PQ. This development occurred only in stressed syllables (according to the accentual system observable in Parmaquesta). In final syllables other than stressed monosyllables, and in unstressed medial syllables it became ī ... ou similarly remained in AQ, but became ō in PQ. Similarly ou > ū in unstressed (PQ) syllables (OP2: PE19/106).

However, actual examples from the 1950s and 60s show ei, ou > ī, ū pretty consistently, even in positions where the diphthong would be stressed:

• ✶meinā > Q. mína “desiring to start, eager to go” (VT39/11).
• √keye > Q. kíta- “suppose” (VT49/19).
• √LOW > lounē̆ > Q. lūn “pool” (PE17/137).

In a table of diphthongal developments from the late 1960s, Tolkien said:

CE: ei > Q. í (ē) ... [CE] ou > [Q.] ū (ō) ... ē after ı̯, y; ō after u̯, w (VT48/7).

Thus it seems that by the late 1960s (like the early 1930s), the special developments for ei, ou are once again conditioned for those diphthongs appearing after the semi-vowels y, w. These specialized developments were also mentioned in OP1 (PE19/53), as discussed in the entry on how [jei], [wou] became [jē], [wō]. This alternate pattern can be seen in a 1969 discussion of the phonetic developments for the suffix -itë:

íte (from iti, eiti), éte (from eiti after i, y) (PE22/155).

As an example of this alternate development Tolkien gave √SKEY [> *skeyeite] > Q. xiéte “passing, impermanent”, though this development could also be the result of eye > ie. To summarize the typical developments in each conceptual period:

• 1910s-1920s: ei, ou > ai, au.
• early-to-mid 1930s: ei, ou > ī, ū but sometimes ai, au (after y, w?).
• mid 1930s to early 1950s: ei, ou > ī, ū when unstressed but ē, ō when stressed.
• mid 1950s to 1960s: ei, ou > ī, ū but ei > ē after y and ou > ō after w.

Note that it is possible that the rules whereby “ei, ou > ē, ō when stressed” and “ei, ou > ē, ō when after y, w” were simultaneously valid at various points (as they seem to be in OP1), but lack of examples make this difficult to determine.

See Q. [ei], [ou] generally became [ī], [ū] for discussion.

Exception: neite