It is well known that Quenya only allowed five final consonants: n, t, l, r, s, and the cluster nt (Let/425; PE19/104; PE22/62; VT42/7). That means any other consonants and consonant clusters had to be modified when final. The number of possible combinations is large, but these changes can be reduced to three relatively simple rules:

1. Final consonant clusters reduced.
2. Remaining final spirants were altered.
3. Remaining final consonants became dentals.

This entry discusses the first change, reduction of final clusters. For the most part, the last consonant was lost, as in -nd > -n or -st > -s, but there were some exceptions to this rule. The exact developments (and the list of valid final combinations) changed somewhat over time but was surprisingly stable.

Conceptual Development: In the 1910s, the number of valid final clusters in Early Qenya was somewhat larger. In addition to -nt, the following combinations also appear in the Qenya Lexicon:

• nk: ᴱQ. lenk “limb” (QL/53).
• lk: ᴱQ. elk “spear” (QL/35).
• rk: ᴱQ. nark “spiteful remark, snap of a dog” (QL/64).
• lt: ᴱQ. lilt “dance” (QL/55).
• rt: ᴱQ. virt “servant, slave” (QL/102).

These additional final combinations were listed in Qenya Phonology, though lk, rk, lt, rt were primarily limited to monosyllables. All other combinations were simplified finally:

Finally these groups as described on this page 21, with which cf. 20a, 19, and especially 16, became further simplified. Only l, r, s, n, k, t, nt, nk and, rarer and usually only in monosyllables, lt, lk, rt, rk were possible. -ss, ks, ps, ts, st all became -s (but -s became -r). nd, mb, ng were very common (from nřə, etc.) and gave n, n, n. Combinations + l however produced vocalic -l (PE12/26).

There is a table on the following page, which indicates that for the most part, the last consonant of the pair was lost, e.g. -nq > -n, -lp > -l, etc. Exceptions include the s-clusters ps, ks, ts listed above, which became -s instead: ᴱQ. tyus (tyuks-) “cud” (QL/50), ᴱQ. oaris (oarits-) “mermaid” (QL/70). Another set of exceptions (not given in Qenya Phonology but appearing in the Qenya Lexicon) are words ending in kt, pt which produced -t:

• ᴱQ. mut (mukt-) “dirt, filth” (QL/63).
• ᴱQ. let (lept-) “finger” (QL/53).

Tolkien revisited the phonetic development of final clusters in the Qenya Phonology of the 1910s. A large number of final clusters were possible because of the loss of short final ă, ĕ, ŏ:

STAGE II. The vowels, e, a, o, vanished in final syllables even when followed by a consonant — but not when followed by a group or long consonant (i.e. when long by position) ... A large number of new consonants originally medial now came to stand at the end of words. These, together with the old original groups, were drastically reduced. The result, already in Old Qenya, was that only the simple dentals l, r, s, n, t and the dental group nt could stand finally (PE14/68-69).

Tolkien then proceeded to list the reductions for a large set of combinations, most of which demonstrate the loss of the final consonant: nz > -n, rm > -r, etc., though many of these also show dentalization, as in ng > -n and mp > -nt. Tolkien lists the same set of exceptions given above: ps, ks, ts > -s (PE14/69) and ht, pt > -t (PE14/70), the latter reflecting the new sound change introduced in the 1920s whereby kt became ht [xt]. Tolkien does add a “?” indicating he was unsure whether kt > -t first or whether kt > ht > -t.

In the Outline of Phonetic Development (OP1) from the 1940s, the section on final consonant is incomplete, and does not address final clusters (OP1: PE19/52, 66-67). Tolkien did return to this topic in the Outline of Phonology (OP2) from the 1950s:

This list of “permissible finals”: n, r, l, s, t and nt remained constant in Quenya speech-feeling. Weak vowels were most readily lost when their disappearance left simple n, r, l, s, t finally. And when vowels were lost after other consonants or after groups, as happened in later AQ and PQ under certain conditions (e.g. especially in long compound words), these consonants or groups were changed or reduced to one of these permitted dentals: m > n, and stops > t; d > r; th, þ > t.

Consonantal groups were simplified. Mostly by dropping all but the first constituent: as nd > n; and modifying that if necessary: as mb, ñg > m, ñ > n. But all combinations containing s (ps, ts, ks, ss, st) were reduced to s; and those containing ht > t.

The last may often actually go back to a period when kt- still existed as such and in fact show kt- > k > t. In PQ t also occurs for pt (OP2: PE19/104).

This is essentially the same as the system from the 1920s, with the same set of special exceptions ps, ks, ts > -s and ht (kt), pt > -t. There is even the same uncertainty over whether cluster reduction occurred first (kt > -k > -t) or kt > ht occurred first. Thus Tolkien’s treatment of final consonant clusters is remarkably consistent. Other than the elimination of a few allowable final clusters from the 1910s (nk, lk, rk, lt, rt) he stuck with the system for final cluster reductions introduced in the 1910s throughout his life:

• Final ps, ks, ts > -s.
• Final ht (kt), pt > -t.
• Otherwise the last consonant of the final pair was lost.

The only other major conceptual change is that Quenya gradually became more resistent to final vowel losses that would produce “unpleasant clusters”. In addition, in the 1950s Tolkien does hint that the cluster -nt was mostly lost: “and nt in classical PQ (TQ n) (PE19/104)”. Indeed, in later writings this cluster’s use seems to be limited almost entirely to the dual dative inflection.