[A long delayed piece. The original is in MS Word; hope it comes through footnotes, macrons and all.]
Dating Nouns (Parma Eldalamberon 21)[1]
Here are a couple of pieces of evidence for dating Nouns late within the 1944‑1951 span identified by the editors, and relative to certain other texts: section (A), genitive singular in ‑n for a date after the Notion Club Papers of 1946, and section (B), pronominal subject prefixes for a date contemporaneous with a stage in revisions to "The Steward and the King" between late 1948 and 1951. The conclusions come with heavy caveats, section (C), which suggest such inquiries are best left till after a given period is as covered as it will be by published materials. But they do hint at the possibilities for a fine mapping of the external developments of Tolkien's glossopoeia that are now becoming available, for further say historical or linguistic study. Recent volumes of Parma Eldalamberon offer other examples, more consequential and robust than the very minor ones here; I hope to return to some of these.
The editors date Nouns before 1952 by the use of Lindar, which was changed to Vanyar in revisions to the Simarillion in 1951 or soon after, and by its placement with two other texts between the pages of a newspaper dated July 2 1952 (PE 21: xvi). These two other texts are the Notes on Final Consonants NFC, composed in 1936 (PE 21: xiii), and Common Quendian Declensions, using paper probably obtained in the spring of 1944 (PE 21: xv). Nouns was placed between them, but conceptually follows Common Quendian Declensions (PE 21: xvi). "Closely based" on Nouns and "probably composed not very long after it" are Notes for Qenya Declensions, placed together with Qenya Grammar from the 1940s and with rough declension tables on printed minutes from a meeting of May 5 1944 (PE 21: xvii, both unpublished). Nouns thus falls into 1944‑1951.
(A) Genitive singular in ‑n
The first hint for a later date in the 1944‑1951 span is the abandonment of the genitive singular in ‑n. This feature characterizes Qenya of the late 1930s: the Etymologies, The Lost Road, and the 1939 version of Frodo's greeting, Eleni silir >> Elen sile lúmesse omentiemman (VI: 324). It also seems to characterize the language until 1946 at least. It remains in the Notion Club Papers of 1945‑6: Lowdham's Soroni númeheruen ettuler (XI: 290n62) and both earlier and later versions of Lowdham's Fragments Ilūvatāren 'of‑God' continuing the text of The Lost Road (IX: 246, 310, V: 47). It is present in drafts of Namárie, possibly in the earliest published version of 1941‑2, e.g. turme márien (VII: 284‑5), but clearly in later versions much closer to the LR form, where we can witness a transition between ‑n in one version to ‑o in a minor revision of it, from Vardan telumar, omanyan, míri Kalakliyan, to Vardo tellumar, omaryo, míri Calacilyo (Tyalie Tyelleliéva 12: 12, 21). The linguistic distance between this group of Namárie drafts and the one of 1941‑2 suggest a considerable lapse of time, matching dating the later one to after the Notion Club Papers of 1945‑6.
Abandonment of this genitive singular seems to be witnessed by the revisions in red ink to the Note on Final Consonants NFC, original dated to 1936, which replace precisely the paragraphs discussing it, while keeping interleaved features such as the genitive plural in ‑ion (PE 21: 59‑61). No mention of the genitive singular ‑n is made where one would expect it in the subsequent Common Quendian Declensions and the later Nouns, where other expressions of genitival relationships are given their due place. Finally, it might be absent in a "grammatical description of the Common Eldarin pronominal elements (probably from the 1940s)" (PE 17: 14): this gives a counterpart of Frodo's greeting as elen sile omentiemman >> omentienwan, revised to Elen síla omentielwan, ‑elman, with the final addition "in full: Elen sīla lúmenn(a) omentielmo", alterations that "took into account changes in the grammar made in other writings from the 1940s and early 1950s". Strikingly, lúmesse, lúmenna is omitted with omentielwan variants, so that ‑n is not genitive but dative (short locative), a change which might be interpreted to indicate that the genitive singular in ‑n was gone.
Taking all this as a linear development – I'll return to caveats about doing so ‑ the terminus ab quem for the rejection of the genitive singular ‑n in NFC would be late 1946, with time to allow for the writing of Common Quendian Declension and then of Nouns.
Addendum: There is similar potential for relative and absolute dating in Sindarin genitives. The – or a – rejection of the Noldorin/Sindarin inflected genitives in ‑a, ‑(i)on, seems to be caught in the act both in the pencil revision to Common Eldarin: Noun Structure (PE 21: 78) sometime after the change of Noldorin to Sindarin (revisions to the Grey Annals in the early 1950s, PE 21: 21, XI: 23‑27), and at the end of an undated note on Roheryn that likewise operates with Sindarin (PE 17 III: 51 s.v. Roheryn, see PE 17 I: 222 s.v. Glorfindel on the document). Inflected genitives are found in dagnir Glaurunga of the Grey Annals (XI: 103, 1951‑2) and the Words, Phrases and Passages entry elenath with elenathon (PE 17 I: 250, ?1960 in PE 17: 4). Later than this seems a discussions Caras Galadon before the second edition change to Galadhon, which shifts the genitive plural ‑on to Silvan Elvish: "but Galad and especially the apparent genitive plural Galadon 'of trees' are not Sindarin" (PE 17 I: 356, I 368 s.v. Galadriel, placed with the discussion of √phan‑ in a draft for The Road Goes Ever On, c. mid 1960s before 1966). Yet curiously, Nan Gondresgion 'Stonewain Valley' (PE 17 I: 256 s.v. Gondor, associated with revision of the Index for 1966) also seems to derive from the same late period, but here ‑ion corresponds to genitives of the counterparts in Old English Stānwægna Dæl and Quenya Nand' Ondolunkava beside Ondolunkanan(do), and it is less natural to see ‑ion here as 'region' or similar meanings than in I: 296 Nan‑Duhirion, I: 319 Eregion, I: Map Rhovanion (q.v. PE 17 under these entries), and potentially in the 2nd edition Caras Galadhon. Inversely, one might have expected a mention of the inflected genitive in the Quenya Notes discussion of Sindarin na‑genitive, listing the options "Aran lintaciryalíva, S ara cîr lim or aran na chîr lim = Aran linta ciryalion" (PE 17 s.v. √ana/nā, 1957). Hard to draw conclusions here.
(B) Pronominal subject prefixes
The other line of evidence for a late date, after 1948, is the concept of pronominal subject prefixes in Eldarin and Quenya.[2] As written, Nouns includes the Common Eldarin CE example, "'I will slay thee, wicked man' kyĕ ndăkŭbā̆nyē, uklainā ndere (for uklaină ndēr, since the latter might be = to nyē, I)". This was altered "in the original ink" to "ni‑ndăkŭbā̆kyē, uklainā ndere (for uklaină ndēr, since the latter might be = to ni‑ I)". "Subsequently Tolkien wrote in the space below this, using a slightly broader‑nibbed pen, another version of the example: ndakubānike, ūmanā neređ. Above and to the right of this he added: nakuvan tye uvana nēra (with a b above the v in uvana), but heavily struck out the last three of these words (so that their reading is not entirely certain)." (PE 21: 64‑5)
Here Quenya nakuvan shows a pronominal subject suffix, as do the added CE ndakubānike and original kyĕ ndăkŭbā̆nyē. However, in between these stages, and close to the time of composition to go by use of the original ink, CE ni‑ndăkŭbā̆kyē is remarkable in its use of pronominal subject prefix.
From the 1944‑1951 period of Nouns, as well as for considerably before and ever after, there is only one point in the corpus where pronominal subject prefixes appear clearly in the Elvish languages, and it is sharply delimited by preceding and following conceptual stages with suffixes: in Quenya phrases in "The Steward and the King".[3]
The textual history of "The Stewart and the King" is as follows: "the preliminary draft ('A') …, though written roughly and rapidly, was changed very little afterwards … A fair copy manuscript 'B' followed, with the chapter number 'LVI' … Of this chapter my father made a third, very fine manuscript 'C', numbering it 'LIV'." (IX: 54) There are two Quenya phrases with pronominal subject affixes.
First, the words of Elendil. They are absent in A. Then, "In B the text of RK was reached at all points in this scene apart from the words of Elendil, repeated by Aragorn when he held up the crown, which take the form: Et Ëarello Endorenna lendien. Símane maruvan, ar hildinyar, kenn' Iluve‑metta! … 'Out of the Great Sea to Middle‑earth have I come. Here will I abide, and my heirs, unto the ending of the world.' In the third manuscript C the words remained the same as in B, apart from tenn' (as in RK) for kenn', but were subsequently changed to Et Ëarello Endorenna nilendie. Sinome nimaruva yo hildinyar tenn' Ambar‑metta!" (XI: 56)
Second, Aragorn's exclamation: "When Aragorn saw the sapling at the edge of the snow he cried, in A, En túvien!, which in B becomes En a túvien! This was retained in C but corrected to En [?in]túviet; on the final (typescript) text of the chapter this was retained, but then erased and Yé! utúvienyes written in its place." (XI: 57) Given nilendie and nimaruva, [?in]túviet arguably reflects nitúviet.[4]
This stage of pronominal subject prefixes is sharply delimited: the final manuscript C of "The Steward and the King" has suffixes, prefixes replace them in the final typescript, but there are changed back to suffixes (Christopher Tolkien's wording does not make it clear if the prefixed forms are found already in a revision to C).[5] This dates them between 1948 and 1953: the Chronology puts the first drafts of these two chapters to 14 August – 14 September 1948, but the fine manuscripts to ?August 1948 – ?end of August 1950 or later: "Some time after he finishes The Lord of the Rings proper and has already written much of its background material, Tolkien makes new, very fine (third) manuscripts of four chapters: 'The Steward and the King', 'Many Partings', 'Homeward Bound', and 'The Scouring of the Shire'." The first complete typescript of LR was completed in October 1949, but revisions to it continue as well as to later typescripts, documented for instance for 30 August – 10 September 1952; the typescript of The Return of the King is delivered in December 1953.
Aside from these Quenya forms, the published corpus of Elvish during, before, and after the writing of The Lord of the Rings documents only suffixal expression of pronominal subjects, with a couple of uncertainties, suggesting that the prefix concept in Nouns is to be brought into relation with the Quenya forms:
From mid to late 1930s, all inflected forms in the Etymologies have suffixes, not only Quenya, but also Noldorin, including Old Noldorin YUR : yurine 'I run', while Notes on Final Consonants NFC from that period notes among the inflections of Eldarin "Suffixion of (ancient) forms of the personal pronouns, combined in various ways with the plural i/r/l elements or dual s/th" (PE 21: 61; my underline). Qenya materials associated with The Lost Road show subject suffixes and object prefixes, tye‑meláne 'I love thee'.
From 1940s, versions of Namárie are a remarkable source because they span the decade. All have suffixes (‑lle, ‑lye). The published versions fall into two major conceptual stages. The first stage consists of the earliest published version from 1941‑2 – likely not the first version, Christopher Tolkien's wording being "My father was working at the same time on the Elvish song, which had reached this form" – and a couple of revisions, and has ettulielle (VII: 284‑5, Tyalie Tyelleliéva 12). The second stage is far closer to the version of the first edition of LR, and has hiruvalye. Two of its closely similar versions transition from the ‑n to the ‑o genitive, dateable to after 1946 (see (A) above; for the versions, see again Tyalie Tyelleliéva 12).[6]
Likewise from the 1940s, the Notion Club Papers of 1945‑6 might show a pronominal subject affix in sakkante of númeheruvi arda sakkante lenéme ilúvatáren 'Lords‑of‑West Earth rent with leave of ?' (XI: 246); and if so, it is a suffix. Antecedents have a singular subject and the bare or singular "impersonal" verbal form, Herunūmen ilu [>> eru >> arda] terhante [>> askante] 'Lord‑of‑West world sunder‑broke' (XI: 310). Yet in texts from the late 1930s to the early 1950s, the i/r‑plural of Q(u)enya does not combine with the singular impersonal form, but either with the plural impersonal in ‑r, as in the Fragments' lantaner turkildi, ëari ullier, or (more rarely and usually revised) with pronominal suffixes, as in Toi … meldielto of the Song of Fíriel 'they … are blessed', Nai elle hiruvalle >> Nai elye hiruva of the n‑genitive >> o‑genitive versions of Namárie, or Átaremma emme avatyaremme ta >> emme avatyarirat cited below.[7] So sakkante might be plural, presumably with the 3p ‑nte of Ciryon's Oath tiruva‑nte‑s 'they will guard it' (UT: 305), though the latter text limits person‑indicating plurals precisely to where the subject does not precede.[8]
From the early 1950s, Ataremma I‑IV, preserved with manuscripts from the 1951 to 1954, translates 'we forgive those' as I emme avatyaremme ta, with avatyaremme ta >> avatyarirat >> avayarir ta >> IIa avatyarir ta >> IIb avatyaritar >> avatyarilta, kept in III, IV. Here avatyaremme shows a pronominal subject suffix, the rest only impersonal number agreement with the independent pronoun (see the editorial commentary VT 43: 20). Only suffixes are also found in Ortírieylanna rucimme 'we fly', in the same ms. as Ataremma III, IV and Aia María II, and in Aia María III, IV firuvamme 'we die' (VT 44).[9]
Pronominal subject prefixes do turn up clearly once before in Q(u)enya, in Early Qenya Grammar EQG from the Leeds period 1920‑1925, perhaps specifically 1923 (PE 14: 37), and possibly in Early Qenya Pronouns (PE 15) between EGQ and the Qenya Lexicon (PE 12). The manuscript version of EQG has the nominative expressed by a prefixed pronoun and the accusative by an unemphatic pronoun immediately following the verb (PE 14: 52‑3, 56), while the typescript has both the nominative and accusative expressed by preverbal proclitics in that order (PE 14: 85‑6). In both versions, 1s is ni and 2s ke. So for the manuscript we can reconstruct Qenya ni‑V‑uva‑ke, very close to Eldarin ni‑ndăkŭbā̆kyē of Nouns. The same order is found in Arctic Mára mesta an ni véla tye ento, ya rato nea of 1929, "Good‑bye until I see you next, and I hope it will be soon" (Father Christmas Letters); but here the 1s and 2s elements are written separately, while just possibly mesta might show 1p me‑ of EQG ('let us be well'?). This stage of pronominal subject prefixes as well is isolated in its time; materials both preceding and following have suffixes, as Quenya Lexicon karin 'I make, do', perilme 'we indeed endure things' (PE 12 s.v. kara, pere), and Quenya Conjugations tulin '*I come' (PE 16).
If we do correlate Nouns with the Quenya prefixed forms of "The Steward and the King", we can date pronominal subject prefixes to 1948‑1951, taking ab quem as the composition of "The Steward and the King", and ad quem the replacement of Lindarin by Vanyarin that did not affect Nouns. Within this period occurred in both sources the shift from suffix to prefix to suffix.
The correlation supposes that the prefixed CE form of Nouns is to be related to the prefixed Quenya forms. Tolkien devised CE as a reconstruction based on features of the Eldarin tongues, and Quenya is the only Eldarin language of the period with evidence of pronominal subject agreement. All earlier Noldorin/Sindarin material has suffixes, e.g. 1930s Etym 3AR: Noldorin gerin 'I hold, have' and Old Noldorin YUR: yurine 'I run'; all later material does too, e.g. early 1950s LR Sindarin linnathon, ónen and Túrin Wrapper agorech (VT 50), or early 1960s cerin paradigm (PE 17: 132). There were subject prefixes in Goldogrin in the 1910s, the antecedent to both the Welsh‑style Noldorin‑Ilkorin/Sindarin branch and the Old English‑style Danian branch and itself also bearing features of Gothic (Goldorin Pronominal Prefixes, PE 13: 97); but they are gone by the earliest fragment of the Welsh style languages, Nebrachar of 1931 (MC: 217). This suggests that prefixes in CE would be there to explain prefixes in Quenya. No more than suggests; even for circum‑LR Quenya, almost every publication has brought features I'd not have guessed at – prefixes tye‑meláne, nimaruva, sekorme, marked nominative langon, inflected prepositions onye and agreeing ones imbi eleni, adjectival derivation on inflection menelessea, már‑ya, honorifics by agglutination of address form carintar(i), shades of the long debate about Indo‑European plural and middle r in ha∙matsir, vanimalda as 'your beautiful', the phonology of Aldudénië. If subject prefixes did turn up in the Sindarin or Danian branches of the Nouns period, they would be in good company.
(C) Caveat
All the above presupposes linear developments; but we know that these are unlikely. Tolkien might have gone back and forth more than once in the 1940s about the genitive and about the placement of pronominal subject affixes.
PE 21 itself documents such vacillations within comparable periods for the marking of subjective and objective cases by vowel lengthening or addition and by a dental occlusive. The latter touches our CE sentence in (B): the original with subject suffix and the revision to subject prefix have uklainā ndere, but the added version reverting to subject suffix has the objective ūmanā neređ. Here đ must prefigure the accusative in d of the complex account in the 1951/2 Common Eldarin: Noun Structure ENS (PE 21: 75‑7). It is not found at all in Notes for Qenya Declension between Nouns and ENS. Yet it appears already, with different developments than in ENS, in Declension of Nouns (PE 21: 62 and 62n59) between NFC and Nouns; antecedent concepts are perhaps to be to be seen in NFC's allative d and in nominative rather than accusative t (PE 21: 57‑8). The néra of the Quenya counterpart itself matches the accusative of the early 1930s Declension of Nouns but no published later material, contrasting notably with the bare accusative of consonant‑stem anar, Isil in the (pre‑)1937 Song of Fíriel (V: 72).
For pronominal affixes, the editors of Quenya Conjugations make this very point when discussing the transition from suffixes in the Quenya Lexicon to prefixes in EQG to suffixes again in Quenya Conjugations: "Tolkien illustrated what he meant when he said, in a note accompanying one of the drafts of the poem Nieninque, that the language 'has often endured grammatical changes especially with verbs'" (PE 16: 116‑7). The consistent use of suffixes in the half dozen or so drafts of Namárie would suggest that much of the 1940s did not see prefixes in Quenya: the 1941‑2 version and its variants, and then a second sequence much closer to the published version, which seems to straddle the post‑1946 abandonment of the n‑genitive. Yet the editors' comment to The Litany of Loretto le∙ana suggests that the 1940s situation is far richer: "The use of pronominal subject prefixes in Quenya is characteristic of the language as it was conceived in the 1940s. This feature was still 'viable' at least as late as 1948 … as shown by the subject prefix ni‑ 'I' in nilendie 'I have come' and nimaruva 'will I abide'." (VT 44: 13)
Nothing for it then other than to wait for the publication of further Eldarin and Quenya material, which we know now to discuss their verbal structure in the period concerned.
Acknowledgments
My profound thanks to the long labours of the editors of Tolkien's linguistic materials.
[1] Sources are cited by the abbreviations used by Vinyar Tengwar VT, Parma Eldalamberon PE, and Tengwestie, q.v. www.elvish.org, or by abbreviations appearing at their first mention, usually those introduced by their editors.
[2] The term "pronominal affix" has two senses. In the sense I use here, affixes are pronominal because they reflect more than one of the grammatical categories that enter into pronouns, here person and number in contrast to number alone: Quenya thus distinguishes pronominal 3p ‑lto, ‑nte, ‑lte, from personless or "impersonal" number‑only ‑r. In the other sense, affixes are pronominal because they play the grammatical role of independent pronouns, reflecting a nominal such as the subject rather than combining with that nominal (save for nominals that can themselves combine with other nominals like doubling emhatic pronouns). The inflections of Breton are such pronominal affixes, in contrast to the inflections of Modern Icelandic that mostly require a nominal, while the inflections of Spanish may but need not combine with one; intermediate are systems where inflections combine with nominals in certain but not other positions, such as the typical system of Middle Welsh where inflections did not combine with nominals save ones preceding the verb in the "abnormal sentence". Quenya has always tended towards the Breton or Welsh type systems, with no agreement with a preceding subject, though even then with variation on agreement with a following subject, as in Nai tiruvantes i hárar mahalmassen mi Númen; but there are exceptions throughout its external history, usually cropping up as an agreeing preverbal subject later revised to a nonagreeing one, as in Nai hiruvalle … Nai elle hiruvalle >> Nai hiruvalye … Nai elye hiruva in revisions of Namárie; examples in revisions of Firiel's Song, Lowdham's Fragments, and Ataremma are mentioned below.
[3] It is tempting to link the transitory emergence of Elvish subject prefixes at this time to the invention and development in 1946 of Adûnaic: agreement of 3rd person subjects for number and gender is attested, does indeed use prefixes (cf. Patrick Wynne and Carl F. Hostetter, "'Verbs, syntax! Hooray!' A preliminary assessment of Adunaic grammar", VT 24: 14‑38). This is a Semitic feature of the language, and Lowdham's Report (IX) attributes the Semitic features of Adûnaic to influence of Khazadian. Names in the tongue of the Dwarves with their characteristic flavour first appear as Khazaddûm and Gabilgathol and Zirak in the c. 1937 Quenta Silmarillion along with the plural Khuzûd for the Dwarves (V: 274/278) – apart from Telchar of the 1930 Quenta (IV: 143) – and during the composition of LR yielded both names and the 1942 phrase Baruk Khazâd! Khazâd ai‑mênu (VI: 200, 458/465‑7, VII: 117/125, 166/174, 457‑8, VIII: 20). Nothing its yet known of its subject agreement.
[4] Possibly related are changes to the Praise in "The Field of Kormallen". The textual history is similar: a first draft 'A' "for most of its length an extraordinary close approach to the final form", followed by "a fair copy manuscript ('B')" (XI: 45) "The form of the 'Praise' in A runs thus (with some punctuation added from the B‑text, which is closely similar): … Kuivië, kuivië! laurea'esselínen ["changed at the time of writing to ankalim'esselínen", editorial note, ‑MR] … In the fair copy B … the Quenya words became Laitalle, laitalle, andave laita! In the second (final) typescript the Quenya words became A laituvar, laituvar, andave laita! This was then changed on the typescript to A laita te, laita te! Andave laituvalme! Thus the form as it appears on the galley proof is: A laita te, laita te Andave laituvalmet … The final text of the 'Praise', as it appears in RK, was typed onto the galley proof." (XI: 46‑7). The forms of laita as originally found in the second typescript lack the pronominal subject suffixes of the preceding and following stages of revision.
[5] If the loss of subject‑suffixed forms of the imperative in the "The Field of Kormallen" indicates the same shift, it did not affect manuscript B but only the second (final) typescript.
[6] There is also the fragment elli yas atintilar >> yassen elli atintillinar >> yasse tintilar in eleni (VT 28:11); it is closer to the second stage than the first, and seems to precede the earliest of its published versions, as these all have yassen tintilar i elení/eleni, matching the first edition of LR.
[7] Contrast the li‑plural stated to agree as a singular, whether conceived of as the general plural of NFC (PE 21: 57) or the partitive plural of Nouns (PE 21: 63) and later.
[8] The editors of the Koivienéni manuscript have suggested a 3p pronominal subject prefix for sealálan as se‑alálan *'they‑grew' (VT 27: 26). Their dating of the manuscript to 1937‑1941 on the basis of the allative tanna and other evidence (VT 27: 8‑9) has since been strengthened by counterparts of the ‑n‑ie ending of erenekkoitanie 'that he might awake them', esen∙ekkoitanie, first in both the Lowdham's fragments' ullier 'should flow', 'they‑should‑pour', now in NFC's mention of CE "certain obscure[?] uninflected impersonal 3sgs, as subjunctive affix jĕ, 3 sg . jē" (PE 21: 61). Striking in the same manuscript is the treatment of the locative as prefix as well as suffix, sekormen written above {se}kormenesse (with ko above {se}), versus otherwise consistent evidence of suffixed locative for the period (the Song of Fíriel númessier, all versions of Frodo's Greeting with lúmesse, all versions of Namárie with Andúniesse, all PE 21 descriptions, etc.; but cf. with a different element Lowdham's Fragments' nu huine (V: 74) >> unuhuine (V: 47, IX: 310) >> nuhuinenna (IX: 246, frontispiece) 'under shadow' (IX frontispiece 'under‑shadow')). Still, the analysis of sealálan remains open. Se might be a locative preverb, as the editors mention, or derive from one, say a perspectival preverb of the German hin‑/her‑ type. Alternatively, it might be a subject prefix of a different sort, one that does not conflict with the coding of a class of pronominal subjects by suffixes, including 1st/2nd person. Specifically, it might indicate "switch reference", that is, that it is picking up the object rather than the subject of the preceding sentence; Tolkien at a much later point had Quenya use different pronouns for the two (VT 49: 14‑5), as many languages do using the personal ‑ demonstrative contrast (Finnish hän – tämä, German er – der, more systematically than English they – these); but Tolkien might have been inspired specifically by Welsh conjunctive pronouns "always set against a noun or pronoun that goes before (or is implied): Dioer, heb ef … A unben, heb ynteu 'By heaven said he … Ah! prince, said the other" where the conjunctive yntey comes from *séndos tou̯os 'this other' (Morris Jones 1913, §159.iii.2, iv.3). It is the behavior of 1st/2nd person subject markers that is pertinent to dating Nouns.
[9] In the same manuscript as At. III, IV, AM II is also found The Litany of Loretto, which does have one case that the editors interpret as a pronominal subject prefix, identifying le as 2s in a le∙ana ocama men >> ana ocama men >> ocama men >> órava men >> órava ómen >> órava ómesse translating 'have mercy on us'. The expression seems open to the interpretation advanced, or to variations on it that rely on the role of imperative a to license le, but also to others such as le as an unknown particle or ana as 'gift > mercy'. The meaning 'have mercy on us' suggests another intriguing possibility: ana as a form of NA 'be' + le as the 2s object proclitic. The verb have in its transitive habeo type frequently comes from intransitive be + object in dative function, that is mihi est, as in modern Breton 2s imperative az pezet 'have!' continuing older particle a + object 2s prefix ‑z + 3s imperative bezet 'let it be to thee' > 'have (thou) it'. The development is found in all the older Keltic languages and fully grammaticalized in Cornish and Breton; see Lewis and Pedersen's A Concise Comparative Celtic Grammar 1937: §340 (3) for Irish, §349 (p. 207) for Welsh, §352 for Cornish, §355 for Breton. Tolkien would have been familiar with it to the extent he was with Cornish and Breton and the history of Brythonic, for it is always raised as a characteristic development, and arguably through some of the sources cited in Lewis and Pedersen for older Welsh and Irish, where it is sporadic; his one known Welsh source reports it, if only briefly: Morris Jones's A Welsh Grammar 1913: §160(4).
No comments:
Post a Comment