Sauron Defeated (73 page)

The original consonants w and Y were weak (consonantal forms of the vowels v and i). Medially they disappeared prehistorically before the vowels v and I respectively. But initially they were strengthened, becoming more spirantal (though w remained bilabial); so that the initial combinations WU and YI remained. The same strengthening occurred between vowels (where w and v had not been lost). After consonants both w and Y remained weaker, like English w and Y. Before consonants and finally they were vocalized and usually combined with the preceding vowels to form diphthongs (see the Vowels). [Footnote 8]

The sound > [see Footnote 1] had no sign in Adunaic script, except in the archaic inscription referred to above [page 418

and Footnote 6]. Presumably it disappeared very early. It cannot be determined whether it had ever been used medially as a base-forming consonant. Probably not.

3 became weakened, until in the classical period (parallel with the softening of the voiceless equivalent H to the breath-H) it merged with the adjacent vowels. This softening of the back spirants may be ascribed to Avallonian influence.

Initially 3 disappeared. Medially between vowels it disappeared also, and contractions often resulted (always in the case of like vowels, A3A to A); U3 + vowels became UW-, and 13

+ vowel became n-. Finally, or before a consonant, 3 became merged with the preceding vowel, which if short was consequently lengthened; as A3DA to ADA.

Assimilations in contact.

As noted above, these were only sparingly made, owing to the strong consciousness of the basic consonantal pattern in Adunaic. And even those assimilations most commonly made in actual speech are seldom represented in writing, except in the comparatively rare cases where the structure of the word was no longer recognized.

The nasals offer, however, a surprising exception to this conservative tendency, both in writing and speech. This is all the more remarkable, since the combinations MP, NT, NK seem not only easy to us, but are highly favoured in Avallonian. They were disliked in Adunaic, and tended to be changed even at the contact point of distinct words in composition: as Amatthani from AMAN + THANI 'the realm of Aman'.

The dental nasal N was in speech assimilated in position to following consonants of other series. It thus became M before P, Ph, B, and M; though notably NW remained unchanged (NW is a favoured combination in Avallonian); and 9 before K, Kh, c, H, 3.

Where the nasal still remained a nasal, as in MB, NG, this change of position is often disregarded in writing.

After these changes in position the combinations of Nasal +

Voiceless consonant all suffered change. In the combinations MP, MPh, NT, NTh, NK, NKh the nasal was first unvoiced, and then denasalized, the resulting combinations being PP, PPh, TT, TTh, KK, KKh. These changes were recognized as a rule in writing, though a diacritic was usually placed above the I, T, or x that resulted from a nasal; the evidence of the audible forms seems to show that this sign was etymological and grammatical, not phonetic. In old formations N + H became 9H and then HH

(phonetically XX, long back voiceless spirant); but in contacts made after the weakening of H to breath-H, or remodelled after the event, NH remained and is heard as a voiceless NN with breath off-glide. NS became TS.

Since M did not become assimilated in position to following consonants there were the combinations MT, MTh, MK, MKh, MS, and MH. Parallel with the development described above these became PT, PTh, PK, PKh, PS, but no example of P-H for M-H is found. In the few cases of contact of M + H MH is written, and (as in the case of NH) a voiceless MM is heard.

Where the following consonant was voiced the changes are few (other than the changes in position described above). 3 after N or the infixed homorganic 9 does not disappear but becomes nasalized yielding 99, which became NG (phonetically 9G). NR, NL

tended to become RR, LL, but usually with the retention of nasality (transferred to the preceding vowel), in speech; the change is not as a rule represented in writing, though such spellings as NRR, NLL are found. M3 became, in accordance with the general tendency of 3 to be assimilated to a preceding voiced sound, MM. MW became in speech MM (colloquially a preceding labial usually absorbs a following w), but this change is usually not shown in spelling.

Other assimilations are rarer and less remarkable. In speech there was a tendency for consonants in contact to be assimilated in the matter of voice; but this tendency is less strong than in, say, English, and is mostly disregarded in writing. Thus we usually find Sapda from Base SAPAD, and Asdi from Base ASAD, where sabda and azda may be spoken (though the z in such a form is only partly voiced and is not the same as the strongly buzzed sound of a basic z).

The aspirates Ph, Th, Kh have naturally a strong unvoicing tendency on the sounds that follow, and transfer their aspira-tion or audible breath off-glide to the end of the group. Thus Ph

+ n, or T, or Th became PTh (or strictly PhTh). Thus from Base SAPHAD is derived * saphdan 'wise-man, wizard', becoming later sapthan (phonetically, as described above, safpan). But such combinations are not very common, and in perspicuous forms (such, for example, as arise in verbal or noun inflexion, or in casual composition) were liable to be remodelled, especially after the change of the aspirates to spirants; thus usaphda 'he understood' for usaptha.

The continuants W, Y; L, R, Z are pronounced voiceless after the aspirates, but otherwise suffer no change. They are also unvoiced after s and H. Before H and s the continuants L, R, Z

were unvoiced, but w and v had already become vowels (U and I). M, N were unvoiced after the aspirates (while these remained as such), but not after other sounds; after the later developed spirants F, p, X the unvoicing of M, N was only partial.

After voiceless sounds 3 while it still remained an audible consonant became H. After voiced sounds it was assimilated to these, so that for instance B3, D3 became BB, DD. As noted above N3, 93, became 99 and then NG.

After voiced sounds H was not voiced but tended to unvoice the preceding consonant. Similarly where it preceded a voiced continuant (as in HR, HM, HZ, etc.); but before B, D, G it tended to become voiced, that is to become the same as 3, and so to disappear, being merged in the preceding vowel.

The Adunaic Vowels.

Adunaic originally possessed only the three primary vowels: A, I, U; and the two basic diphthongs AI, AU.

Each Base possessed one of these vowels: A, t, v as one of its essential components; this I call the CV (Characteristic Vowel).

The normal place of the CV was between the first and second basic consonant: thus NAK-, KUL B.

The 2-consonant Bases could also add the CV at the end; and the 3-consonant Bases could add it before the last radical: NAKA, KULUB. These forms with two basic vowels may be called the Full forms of the Base.

Various other forms or modifications occurred.

(i) Prefixion of the CV: ANAK, UKULB, IGIML.

(ii) Suffixion of the CV in 3-consonant Bases: KULBU, GIMLI.

(iii) Suppression of the CV in its normal place, in which case it must be present in some other place: -NKA, -KLUB, -GMIL.

This 'suppression' of the normal CV can only occur in 2-consonant Bases where it is also suffixed. It also requires that the CV shall be prefixed: ANKA, UKLUB, IGMIL; or (more rarely) that some other formative prefix ending in a vowel shall be present: DA-NKA, DA-KLUB DA-GMIL.

These modifications are seldom combined: that is, a basic form does not usually have the CV repeated more than twice (as UKULBU, KULUBU); though such a form as UKULB could not originally stand in Adunaic as a word, some other vowel than the CV was taken as the ending (as UKULBA).

One of the vowels of a basic stem must be either the CV or one of its normal modifications (described below); but the second vowel of the 'Full form' need not be the CV, but may be any one of the primary vowels (or their modification). Thus NAKA - NAKI, NAKU; KULUB - KULAB, KULIB. The prefixed vowel (as distinct from a separate formative prefix) must always be the CV; but the suffixed vowel may also vary: so KULBA, KULBI; GIMLA, GIMLU. [Footnote 9]

Every primary vowel A, t, v can show one of the following modifications:

(i) Lengthening: A, I, U.

(ii) Fortification or A-infixion: A, AI, AU.

(iii) N-infixion: AN, IN, UN. [Footnote 10]

In the older language over-long vowels were recognized, and marked with a special sign, in my transcription represented by ".

These occurred: (i) as an actual basic modification: chiefly in 2-consonant Bases, and in any case only before the last basic consonant; (ii) as the product of the contraction of vowels, where one of the merged vowels was already long. Thus Base ZIR 'love, desire' produces both zir and zir; and also zaira and zair 'yearning'.

Similar forms were sometimes produced by Bases with medial W, Y and lengthened CV: as Base DAWAR produces * daw'r and so daur 'gloom'; zayan 'land' produces plural * zayin and so zain.

Except in the oldest texts and 'heard' forms the diphthongs ai, au have become monophthongized to long (open) e and o respectively. The long diphthongs remained unchanged, and are usually heard, whatever their origin, as diphthongs with a long vowel as the first element, and a shorter one (always t or u) as the second element; though this second element is rather longer and clearer than in a normal diphthong: the intonation is

'rising-falling'.

The only source of e, o in Adunaic is the older diphthongs ai, au. The language consequently possesses no short e or o.

Avallonian e and o are usually represented by i and u, respectively; though sometimes (especially in unstressed syllables before r, or where the Adunaic system favours it) both appear as a. In the earlier loans from Avallonian, presumably before the monophthongization of ai, au, Avallonian e and o appear as t and u respectively; but later they appear as e and o.

Contact of vowels.

This can be produced (i) by the loss of a medial consonant, especially 3; (ii) in suffixion, especially in the addition of the inflexional elements: i, u, a, at, im, etc.

If one or both of the components is long then the product is a long diphthong or an over-long vowel.

v contracts with U; I with I; and A with A.

After v a glide consonant w is developed (so u - a, u - t to uwa, uwi), as described above. Similarly after t a Y is developed (soi-a,i-u to iya, iyu)

Earlier Adunaic also possessed the long diphthongs: OI, OU, and EI, EU. These were all contraction products, and EU was rare. In the classical period OI (and EU) remained; but OU

became the over-long simple vowel o, and similarly h became E.

These diphthongs were mainly found in inflexional syllables, where they appear to be produced by adding such inflexional elements as -i, -u direct to the uninflected form (come to be regarded as the stem) instead of to the etymological stem.

Thus the plural of mano 'spirit', from * manaw-, or * manau, is manoi.

But similar forms can also be produced basically. Thus a Base KUY can produce by 'fortification' kauy- to koy, koi. A Base KIW

can produce by 'fortification' kaiw- to kew, keu. It is possible that the inflexional forms are also, at least partly, of similar origin. If the plural inflexion was in fact originally YI not I (as it seems to be, because Y was lost before I medially) then the development would be so: manaw, manau + yi to manoyi to manoi; and similarly izray, izrai + yi to izreyi to izrei to izre.

By the processes (i) of N-infixion, and consonant doubling; and (ii) of varying the position of the CV, and modifying it; and varying the vowels of the subordinate syllables, the Adunaic Bases, and especially those of 3-consonant form, were capable of an enormous number of derivative forms, without recourse to prefixion or suffixion. Naturally no single Base shows more than a few of the possible variations. In any case, any given derivative never shows two of the one kind of variation at the same time; for this purpose w-infixion and consonant doubling count as one kind of process; and Lengthening and A-fortification count as another. Alteration in the position of the CV, and variation of the subordinate vowels, can be combined with any other derivative process.

Even with these limitations such Bases as KULUB and GIMIL

can for example develop the following variants (among other possible forms):

KULBU, -A -I; KULAB, KULIB, KULUB; UKLUB - Kulbo, -a, -e,

-u, -F; kolab, kolib, kolub, kulob, kuleb, kulab, kulub, kulib; uklob, uklub

Kullub, -ib, -ab (with variants showing -ub, ib, ab, eb, ob); kulubba, kulubbi, kulabbu, kulabba, kulabbi, kulibbu, kulibbi, kulibba; kulumba (also kulimba, kulamba, etc., though N-infixion is usually found with the CV preceding the nasal); uklumba; etc.

GIMLI, -A, -U; GIMAL GIMIL, GIMUL; IGMIL with parallel variations, such as GEMIL, GIMEL, IGMEL, GIMMIL, GIMILLA, etc.

The apparent gradations produced by these changes are: Basic A: a - a - a

Basic I: i - i - i; e - ai