LIETUVIŲ KALBOTYROS KLAUSIMAI, XXXIV (1994)

LIETUVIŲ KALBOS TARMĖS IR JŲ TYRINĖJIMAI

Danguolė MIKULĖNIENĖ

ACCENTUATION OF DECLINABLE DERIVATIVES WITH THE PREFIX *PĀ-

§1. The two prefixes of present-day Lithuanian, -po- and -pa-, are historically related to the preposition po (< *pā). Both may be regarded as allomorphs of one and the same prefix *pā-: pa- corresponds to the weak, and po- to the strong grade of realization (Mikulėnienė, Žeimienė, 1994).

This is evident from the accentuation of declinable derivatives. Derivatives with the prefix po-—both nouns and adjectives—are always accented on the prefix and only according to a fixed-stress paradigm, cf. pódūkra (1), pópietė (1), póžemis (1), pósūris, -ė (1). Derivatives with the prefix pa- are characterized by mobile stress, cf. pameragė (3ᵇ), pagirūs (3ᵇ), pavakarė (3⁴ᵇ), papilnys, -ė (3ᵇ). Various cases of root stress are found: pakalnė (2) and pakilnė (1), pakrántė (1) and pakrántė (2). Root-stressed variants such as pažūris (1), patėvis (1), pasuolė (1) often have prefix-stressed variants—in either fixed or mobile stress—in the dialects, cf. pajūris (1), pãtėvis (1), pasuolė (3ᵇ): pãsūolė.

Thus, the abundance of variants prompts us to seek factors that may have determined the differing accentual development of derivatives with the prefix *pā-.

§2. For the prefix po- to develop from *pā-, the vowel a must have been long. It may later have shortened—*pā- > pa- (Zinkevičius, 1981, p. 186–187). In that case, po- and pa- should represent two chronologically distinct strata of derivatives, with the older allomorph being po-, which is formally closer to the preposition po- (Kazlauskas, 1968, p. 54).

§3. On the other hand, it seems that development in the opposite direction was also possible: as a result of vowel gradation, *pā- could have arisen from *pā- (cf. Fraenkel, 1962, p. 635; Bammesberger, 1992, p. 49 ff.).

The latter development is attested by the prefix pa- in Latvian, Lithuanian, and Prussian, and perhaps by pā, used as a conjunction and preserved in western Lithuanian dialects, cf.: Tiktai pro tas duris tauks pa tauks Slnt; Ka tu visur par daug kišies: antīs pa antīs, vištos pa vištos Jrb; Kakarykū pa kakarykū vis tą patį gied Kv; Jis vis taigi pa taigi NdŽ (also Sb, Skr—see LKŽ IX l; Būga, 1961, p. 132–133). Moreover, the prefix pa- is characteristic not only of the Baltic languages—cf. Slavic pa- alongside the preposition po, and Latin pō- (pō-situs) alongside pōno < *pō-s(i)-nō (Fraenkel, 1962, p. 635).

Thus Lithuanian and Prussian po- (absent in Latvian) should be regarded as more recent—

118

¹. J. Endzelīns regarded the neologisms as contemporaneous with the Prussian prepositions po (Endzelin, 1944, p. 146).

§4. It is plausible that these allomorphs could have arisen at the same time. Their emergence was determined by the prefix’s differing position in relation to the accent. This is how, for example, the variation in the Prussian prepositions (or prefixes) no/na, po/pa, pro/pra is explained (Mažiulis, 1981, p. 253—254).

The contrast between *pā- and *pā- could have arisen from the differing accentual class of the base words. When the accent of the derivative was formed, the root of an oxytone (or mobile-accented) word fell into an unaccented position, so the prefix had to become the strong morpheme *pā-, later turning into the allomorph pó- (pód ukra < *pā-dukr-ā). The unaccented *pā- could have been short (or shortened), while the baritone root retained the derivative’s accent (pasául from < *pa-sāul-īs). Likewise, *pā- must also have shortened after losing its accent for some reason.

Visais atvejais dariniai priklausė pastovaus kirčiavimo paradigmai (Mikulėnienė, 1992, p. 12).

§5. If the prefix is derived from the long allomorph, derivatives with mobile accentuation appear chronologically more recent; conversely, they become more archaic if the possibility of a short *pā- is accepted.

In other words, at least several accentuation patterns for derivatives with the prefix *pā- may have long coexisted in Lithuanian. For example, nouns formed from words with baritone accentuation (B) — sáulė, béržas (Latvian bērzs), rankà, balà... — can have three accentuation patterns in dialects (see Scheme 1): fixed on the root (P₁), fixed on the prefix (P₂), or mobile (K).

Scheme 1

Derivatives formed from words with mobile accentuation (O) — akis, sparnas, kaklas, dukrà... — likewise show three accentuation patterns: they are accented consistently on the prefix pó- (P₁) or pa- (P₂), or may have acquired mobile accentuation (K′).

§6. The antiquity of the accentuation types is demonstrated by hydronyms with the prefix *pā-, cf.:

¹ The sole Latvian example, dialectal pūosaga, seemed doubtful to J. Endzelīns (Endzelīns, 1951, p. 665).

² The variant with the prefix pa- (padukra, LKŽ IX 59) is unaccented.

119

_{P₁ — Pašyvis} Gdž, tributary of the Srautas: Šyvė lake, Gdž; Paskāistis Rk, tributary of the Juodupė: Skāistė lake, PnmR; Pagulbis lake, Ut — cf. Gulbis lake, Mlt, gulbė, gulbis; Pabāmbis Sln — cf. Bamb-enà Smn; Pagėrvis Jz, tributary of the Alsia — cf. Gėrvė adv., Gėrvė Rdn;

_{P₂ — Pādegla} Sd, tributary of the Kiršinas — cf. Dėglė Krtv; Pāprūdžiai Krtn — prūdas, pāprūdžai; Pāupis Dkst, part of Lake Dysnykštis — ūpė, pāupis;

K — Pamolys lake, Mlt — cf. Mólė Skdv, mólis; Padubė Pn, village of Padubė, on the Juoda River: Dūbos village, Pn; Pakojys (also Pākojis) lake, Onš: Kója lake, Onš;

P — Póażvintė Trgn: Ažvintis lake, Trgn;

P — Pādaubis End, tributary of the Letausas — daubà, pādaubis; Pāšilis lake, Ds — šīlas, pāšilis;

K′ — Pašilė Vp, on the Širvinta River — šīlas, pašilė; Pašakys Žmt, tributary of the Siesartis — cf. Šakā personal name, Šakā ‘branching of a river’ and others (Vanagas, 1970, p. 218—222; 1981, p. 236—264).

§7. A comparison of derivatives with fixed accentuation shows that only one type of them (P₁) can be directly associated with *pā-, which had a long vowel; however, this type is attested least frequently among hydronyms—its sole certain example is Póażvintė, which, moreover, differs from the others by having a disyllabic derivational stem.

Elsewhere, the prefix is reconstructed with short (or shortened) *pā- in both stressed (P₂ and P′₂) and unstressed (P₁) positions, and therefore it cannot be regarded as a later development than *pā-.

§8. Thus, the reasons for the diversity of accentuation patterns among derivatives probably lie not only in different diachronic strata; above all, they should be explained by areal (or dialectal) differences.

Thus, the possibility cannot be ruled out that alongside stressed *pā- there may also have been unstressed *pā-, and alongside unstressed *pā-—stressed *pā-; see Scheme 2.

Scheme 2

Examples:

pódunka (1) pāsáulìs (1) pasáulìs (1) pakakīs (3ᵇ) póakis (1) pāspamis (1) pabėržis (1) pabėržė (3ᵇ) Póażvintė (1) Pādegla (1) Pašyvis (1) Pašilė (3ᵇ) Pāšilis (1) Paskāistis (1) Pašakys (3ᵇ)

§9. Areal differences are also reflected in derivatives included in dialect dictionaries. Words of all types are found, for example, in the northeastern Žemaitian Dunininkai dialect; cf.:

_{P₁ — pakójos} (1): kója (1), pakrūmė (1): krūmas (1), pakūgė (1): kūgis (1), palóvė (1): lóva (1), pasáulė (1) ‘1. earth, world; 2. humankind’: sáulė

120

(1); pakūnsis (1): kãnsas (1), pakūnonis (1): klūonas (1), pamuilčiai (1): millai (1), etc.;

P’₂ — pajutris (1), pãjutris (1): pėra (1), pãšonis (1): šonas (1), pãteris, pãtevis (1): lėvas (3)³, etc.;

K — pažįnamės (3ᵇ)⁴: žãnėna (1); pakankys (3ᵇ) ‘skandulys’⁵: kãnas (1);

P’₁ — pódulktu (1), pódulktre (1): cf. duktė (2), duktė (3ᵇ);

P’₂ — pakalkis (1): kãklas (1), pãšalis (1): šalis (1);

K’ — pabarzdys (3ᵇ) ‘beard’: barzda (1) ‘beard’, pakraščys (3ᵇ): kraštas (1), pamuiškys (3ᵇ): miškas (1), parugys (3ᵇ): rugiai (1), etc. (Vilkauskas, 1976).

§10. Apparently, the accentuation system of derivatives does not differ between the Southern Aukštaitian and Eastern Aukštaitian Vilnius subdialects (in A. Salys’s classification, the western and eastern Dzūkian dialects). In these Aukštaitian dialects there are almost no nouns that are consistently accented on the prefix pa-, i.e. of types P’₂ and P’₂; cf.:

P’₁ — pakūnsis (1): kãnsas (1); pakójės, pakójės (1): kója (1), panóse (1): nósis (1), etc. (Naktinienė, Paulauskienė, Vitkauskas, 1988), as well as pakūnsis (1): kauršys (3), pakója (1): kója (1), panóse (1): nósis (1), etc. (Petrauskas, Vidugiris, 1985);

P’₂ — pakalkūnas (1) ‘foothill’: kãlnas (1), pakūpelis (1): kūpelė (1) (Naktinienė, Paulauskienė, Vitkauskas, 1988);

K — pakójyš (3ᵇ) ‘1. foot-end; 2. area around the feet’: kója (1) (Naktinienė, Paulauskienė, Vitkauskas, 1988), and papėrys (3ᵇ): prėva (1) (Petrauskas, Vidugiris, 1985, p. 184);

P’₁ — in both dialects póūksis (1): ūkis (1); pódulktu, pódulktre (1); póšunis (1): saunė (3);

K’ — pakalkėlė (3ᵇ): kãklas (1), pakėlė (3ᵇ): kėlis (1), pakrauklys (3ᵇ)⁶: krauklas (1), etc. (Naktinienė, Paulauskienė, Vitkauskas, 1988), as well as pabarzdė (3ᵇ): barzda (1), pasuergys (3ᵇ): suerna (1), pagurklis (3ᵇ)⁷: gurklys (3), etc. (Petrauskas, Vidugiris, 1985).

The prefix type P’₂ in pakalkūnas indicates that prefix-accentuation patterns in these dialects have probably disappeared for morpho(phono)logical reasons—owing to a change in the derivative’s stem-final element and accentual paradigm (Stundžia, 1981, p. 60); cf. pakalknė (1), pakalkūnė (2), pakalkūnė (3ᵇ), and pakalkūnus (1), pakalkūnis (2), pakalkūnys (3ᵇ) (see LKŽ IX 146–147; LKA I 178–179).

§11. Over time, the accentuation of some base words may also have changed in the dialects (Stundžia, 1981, pp. 86–90). Therefore, for example, the derivative pakūnsis (1)

³ For the accentuation of the base word, see §13.

⁴ Alongside the variant pažįnamui (1) (Vitkauskas, 1976, p. 252).

⁵ The first variant is — pakūnas (1) (Vitkauskas, 1976, p. 235).

⁶ The second variant is — pakrãuklis (1) (Naktinienė, Paulauskienė, Vitkauskas, 1988, p. 239).

⁷ The first variant is — pagūrklis (1) (Petrauskas, Vidugiris, 1985, p. 181).

121

synchronically associated with several base forms that differ in accentuation: káušas (1), káušas (3), kaušỹs (3); see §9–10.

§12. A considerable amount of information on the accentuation of base words and their derivatives is found in old written sources.

For instance, alongside M. Daukša’s nouns that retain the acute accent of the root of the base barytone (P₁) — paʃáulis (paʃáulo, paʃáulo, paʃáulų, paʃáulie), pakãtnes (like antkãtnes), pakãtnis (pakãtnei)⁸, pakózis (Skardžius, 1935, p. 80–81, 109), — the prefix-accented variant (P₂) is also attested — paʃáulis (paʃáulo, paʃáulų).

§13. More than twenty P₁-type derivatives with the endings -ė (of the (i)ē-stems) or -ais (of the (i)ā-stems) are given in F. Kurschat’s dictionary (Kurschat, 1883), cf. pabálke (: bálkis), padūmė (: dūmai), pajūrės (: jūrės, jūrios), pakinkės (: kinka), pakūgė (: kūgis), pakūlsė (: kūlbė, kūlbsis), palówė (: lówa); pajūrai (: jūrios, jūrės), pakáubsis (: káubas), pakrélis (: krélas), pakrūmis (: krūmas), pamūris (: mūras), Papélkiai (: pélkė), paʃáulis (: ʃáulė), paʃiėnis (: ʃiėna)⁹, patiltis (tiltas). The ending differs in pakója (: kója).

Closely related are pažándes (: žándas), pakálnis (: kálnas), paʃtógis (ʃtógas), patėwis (: téwas), pažénklis (žénklas), but the accentual paradigm of the base word remains unclear in the forms recorded by F. Kurschat—the singular (plural) nominative and singular (plural) genitive. Old written sources show that at least some of these were barytones, for example, kãtnas (see footnote 8), ʃtógas (ʃtógu, st'ogais), téwas (T'éwai, T'éwū, T'éwams, T'éwus) (Buch, 1961, p. 14; Kabelka, 1964, p. 211–237; Zinkevičius, 1975, p. 17–18).

§14. F. Kurschat also recorded a few P₂-type nouns accented on the prefix — pãmotė (: mótė, -ẽrs; also cf. moterà, -õs)¹⁰, pãodė (: oda, ūda)¹¹.

Prefix accentuation is probably associated with a mobile accent in the base word: in F. Kurschat’s dictionary, this is the accentuation of some derivatives formed from mobile-accent nouns with circumflex or short roots, cf. pãdugnės, -ių (: dūgnas), pãnages, -ių (: nãgas), pãpetės, -cziū (: petỹs, pėczio); pãauʃis (: auʃis, -iẽs)¹², pãʃparnis (: ʃpãrnas), and the author’s unaccented pabarzdis = *pãbarzdis (: barzdà, -õs), pablaũzdis (: blauzdà, -õs), padiewis (: Diẽwas)¹³.

Thus, the prefix-accented variant paʃáulis in M. Daukša’s writings could be associated with the mobile-paradigm noun ʃaulė (Skardžius,

⁸ In M. Daukša’s writings, kálnas is a barytone word; cf. kálnai (kálnai, kálnų, kálnams, kálnus, kálnop, kálnop) (Skardžius, 1935, p. 30, 132).

⁹ In F. Kurschat’s dictionary, the sign for the acute accent is written above the second component of the diphthong ie — ʃiėna.

¹⁰ M. Daukša consistently accented — mótė (mótė, móters, móterieʃp, móteres, móterų, móterump, móterimus, móterimis) (Skardžius, 1935, p. 127).

¹¹ Neither variant of the base word is accented in the dictionary.

¹² Alongside it, the form paauʃiai, -ių (= *paausys). is given.

¹³ Cf. pãblauzdis, pãdiewis (LKŽ IX 18,48).

122

1935, p. 105). It is entirely plausible that the paradigm of the base word is not an innovation of a single peripheral dialect; cf. saulė (3) (Petrauskas, Vidugiris, 1985, p. 229), Latvian dialectal saūle (oral information from D. Markus and A. Sarkan) ¹⁴, as well as Old Prussian saule (Mažiulis, 1981, p. 14).

§15. Other derivatives with the same root are accented by F. Kurschat according to the mobile-accent paradigm (K̑), for example, padirvė (: dirwà, -õs), pakaklė (: kãklas), pakelė (: kẽlias), pakìškė (: kiškà, -õs) ¹⁶, pakrem̃lė (: krem̃lė, -ės), pãlazdės, -džių (: lazda, -õs), papilwė (: pìtwas), paragė (: rãgas), paſ talė (: ſ. tãlas), paſwardė (: wãrdas), ¹⁷ pažaſtė (: žal̃tė, -ės); pabalỹs (: balà, -õs) ¹⁸, pabulỹs (: bulis, -iẽs), pakaklỹs (: kãklas), pakelỹs (: kẽlias), pakemỹs (: kiẽmas), pakraſtỹs (: krãbſtas), paligỹs (: ligà, -õs), patworỹs (: tworà, -õs).

§16. Nouns derived from circumflex or short-root barytones also exhibit mobile accent, for example, padalgė̃ (: dalgis), pãßwentės, -čiū (: ßwentė, -ės), pawartė, -ės (: wartai) ¹⁹, pagirỹs (: girė, -ės; giria, -ios) ²⁰, palapỹs (: lăpas), pamarỹs (: mãrės, -rių; mãrios, -rių), papū̃ys (: ūpė, -ės).

Thus, mobile accent may have been generalized to all derivatives with circumflex and short roots, irrespective of the accentual paradigm of their base word (cf. Otrębski, 1965, p.20).

§17. Evidence of such generalization has already been observed in the writings of K. Donelaitis; cf. pagirỹs (Pagiri̇̀, pa(g)iri̇̀, pagir’eis) and patworỹs (Patwori̇̀) (Buch, 1961, p. 36; Kabelka, 1964,p.157, 171).

§18. Derivatives formed from words with mobile accent must have been the first to acquire the mobile paradigm. This is also attested by the sole such example in M. Daukša’s writings, pakampė̃ (pãkampę, pãkampes, pakampemės, pakampė̃n) (Skaržius, 1935, p. 109).

The derivative pažemė ‘underground’ in the anonymous catechism of 1605 is recorded with two accents — pažėme[nú (Zinkevičius, 1975, p. 42). If the two accents are not a proofreading error (cf. Girdenis, 1984), then this instrumental plural accentuation supports the hypothesis that a derivative formed with a prefix first took over the accent of the unprefixed word.

For example, according to P. Skardžius, nūopelnas, -ai arose from *nuopelnas, -ai (like pelnas, -ai); cf. M. Daukša’s nūpetnóp, nūpetną, nūpetnái, nūpetnū, nūpetnūmp, nūpetnámus, nūpetnãmus, nūpetnūs, nūpetnáis (Skardžius, 1935,

¹⁴ This variant is not mentioned in earlier sources (see Mülenbach, 1927—1929, p. 772—773; Фортунатов, 1993, p. 206), but it is not new.

¹⁵ Plg. M. Daukšos dirwá (dirwos, dirwoie, dirwoſe; dirwóie) (Skardžius, 1935,p. 85).

¹⁶ The LKŽ (V 884) gives only kiškà (2).

¹⁷ F. Kurschat’s dictionary also contains pažáſtis, -iẽs (: žal̃tis, -iẽs).

¹⁸ Cf. balà (2) and balã (4); see LKŽ (I 579—580).

¹⁹ The nominative form of the derivative is unaccented in F. Kurschat’s dictionary.

²⁰ Žodyne taip pat pateikta giria, -ió‘s (Kuschat, 1883, p. 124).

123

p. 45) ²¹.

The stress that had been in the derivative’s root—as also shown by the genitive plural forms of nouns in the 1605 catechism—palaikyt ‘remains’, priepuolų ‘attacks’ (Zinkevičius, 1975, p. 35)—could later have shifted to the prefix (cf. palaikai (³b): palaikus, priepuolis (1)).

Thus, the strong position of the derivative’s prefix did not develop all at once.

§19. There is no evidence attesting to this type of stress in prefixed derivatives—unless it is F. Kurschat’s padáṅges, -ių (Kurschat, 1883, p. 283); in that case, however, one would have to acknowledge an acute accent in the base word and a mobile-stress paradigm, which may have replaced the earlier barytone paradigm. Cf. padáṅges, -ių < *dangrìs, -aũs: dángų < *dáng-us, -aus (cf. Klimavičius, 1970, pp. 192–193) ²².

Old texts also attest to the two stress patterns of this base noun; cf. M. Daukša’s dagùis (dagáus, dagùia, dagùie, dágus (n. pl.), dágų, dagùimus, dagùlė, dagùję, dagùfzia) and dágus (dágau, dágau̇s, dágne, dágun) (Skardžius, 1935, pp. 123–124), as well as the forms in the anonymous catechism of 1605 dungáus, Dungùis, Dungáus, dungùin, dungùjų, dungùjų, dungùl̇i, Dungùȯl̇i, Dungùȯfra and Dùnguy (loc. sg.), dùngen (Zinkevičius, 1975, p. 25).

Other nouns derived from mobile-stress words with an acute root or a polysyllabic stem are given mobile stress by F. Kurschat (K¹): pagalbùs (: galbù, -òs), pagérklė (: gerklė, -ės); pamuokarỳs (: mákaras), pamuodėnùjs (: uonùdė, -ės).

§20. Several possible hypotheses concerning the intonation and stress paradigm of the base word dangùs call for a fuller discussion of the accentuation of the derivative padáṅgė as well.

On the basis of F. Kurschat’s evidence, acute metatony can already be discerned in this and several other derivatives of the P¹ type with a circumflex stress pattern; cf. padáṅges, -ių (: dangrìs, -aũs) and Pakìrfmà (: Kìrfmà, -òs), palàikytis (: lùtkos, -ų), pavèrfnìs (: uerfnù, -ės), pavìrùs (: vîrùs, -ìs).

Be jų, kaip priegaidžių kaitos reiškinį K. Būga (1959, p. 406–407, 412–413) dar nurodo pakàmpenis (: kampùs), pakràutė (: kraũtas), pakràutės (: kriaũšius); padỹsniai (: Dỹsna, -os), palùbis (: lùba, lùbà).

§21. The circumflex of the base word in these derivatives is usually explained synchronically, by the most widespread or best-preserved stress variant (as in the case of dangùs discussed in §19).

However, other stress patterns are also found. For example, the root in F. Kurschat’s uerf’nù, -ės was probably circumflexed (cf. Kurschat, 1876, p. 186), whereas in M. Daukša’s writings it is acute; cf. uerfnùė (uerfnùes, uerfnùe,

²¹ P. Skardžius relates possible derivatives of these disyllabic nouns solely on the basis of stress and emphasizes that “these, with their second element dangùlis [...], can be used as independent nouns without prefixes [...]; dangùmas with a prefix is primarily derived not from nouns of the same stem, but from the corresponding prefixed verbs” (Skardžius, 1943, p. 454).

²² P. Skardžius apparently regarded the acute intonation of the root as a new secondary phenomenon (Skardžius, 1935, p. 125).

124

werfmeíe, werfmiũ), and may even have had fixed stress — wêr[mes (gen. sg.) (Skardžius, p. 106).

Alongside the more usual forms of basic nouns with mobile stress, kampas (4), krántas (4), liaukà (4), or tūba (4), one should also mention variants that bear fixed stress in dialects: kámpa (1) (LKŽ V 179), kámpas (1) or even kámpė (1) (LKA I 101,103), krántà (1) (LKŽ V 424–425), liáuka (1) (Karaliūnas, 1968, p. 69) ²³, tūbas (1) (Būga, 1959,p.413) ²⁴. Likewise, patámsis (2) should be compared only with tamsà (4); tamsùs, -ì (4) (Vitkauskas, 1976, p. 249, 545), whereas patámsė (1), patámsis (1) should be compared with támse (1), támsa (1) or támsis (1); támsus, -ì (1,3) (Naktinienė, Paulauskienė, Vitkauskas, 1988, p. 251, 395).

§22. The archaism of the derivational stem’s acute accent is beyond doubt — P₁-type nouns such as pakrúmė/pakrúmis, palóvė/palóvis are attested throughout the Lithuanian-speaking area (LKA III 36–37, map no. 32).

The existence of variants with falling intonation in dialects or preserved in old writings allows us to cautiously suppose that many P₁-type derivatives with an acute stem may have retained an older accentuation. This is also likely, since an acute change in intonation is not characteristic of nominal derivational stems (Mikulėnienė, 1989, p. 48) ²⁵.

When the accentual paradigm of the base word changes, the derivative reflects a more archaic accentuation. For example, the Eastern Aukštaitian ingėrklis is probably related not to gerklė (4): ger̃klė nor to gerklė (3): gérklė (3) — an accentually closer variant is * gérkl-ā (> gérkla, -os), cf. 1a. iėdžėrklis (Būga, 1959, p. 413).

§23. The relationship emerging between possible base forms and their derivatives can be represented as

pakámpė (1): kámpa (1) = pakampė̃ (3ᵇ): kaũmpas (4).

Lithuanian nouns pasparnė (3ᵇ), pasparnỹs (3ᵇ): spar̃nas (4) are to be related to Latvian pasparne: la. spàrns (Būga, 1959, p. 407). The reconstructed derivative Latvian paspārne = and Lithuanian * paspárnė would suggest taking * spārn-ā as the base word—as in dialectal kámpa, máiša, šérna (Skardžius, 1968, p. 21); cf. Lithuanian spìrti (Fraenkel, 1965, p. 861).

§24. With circumflex stress on the stem, F. Kurschat gives parañkė (: rankà, -os), paláukis (: laũkas), pakaltùs (: kalnės, -ės), as well as several derivatives characterized by circumflex intonational alternation—pabùrnis (: burnà, -õs), pakalnė (: kálnas).

F. Kurschat explains the nouns pakalnė and pakálnis (see §13) differently: pakalnė — ‘die Niederung, die Gegend am kurischen Haff und den Memelmündungen’; pakálnis — ‘ein Anberg, der Fuss eines Berges’ (Kurschat, 1883, p. 287). Since the two could have been recorded in different dialects, the base word may also have differed in accentuation, cf.

²³ LKŽ (VII 392) lists only liaukà (4).

²⁴ The shift of nouns from one accentual paradigm to another is still taking place in our time (cf. Stundžia, 1984).

²⁵ In the nominal stem, it can also perform the function of marking meaning (iconicity) in the case of žiopla: žioplas (= kváisa: kváiso).

125

pakalnė (1): kálnas, -ai (see footnote 2) = pakalnė: kálnas, -ai²⁶

§25. K. Būga’s examples given earlier (1959, p. 407), palõvė: (: lóva), pašóstė, (: šóstas), pakõjė (: kõja), should be interpreted as analogical dialectal formations or formations that arose in some other way.

§26. The possibility cannot be ruled out that nouns distinguished by circumflex metatony are later remodelings, for example, of derivatives from the mobile paradigm (K¹), which may have arisen through a shift of stress onto the medial syllable of the word; cf. M. Daukša’s bedievùs, -ė̃ (bė́dieuniu, bė́diewu, bediewė́i, bediewė́iu, bė́diewes, bediewiámus, bediewiámus, bė́diewius) and bediẽvis, -ė (bediė́wis, bediė́vių, bediė́viamus) (Skardžius, 1935, p. 131).

Similarly, if *beplaukūs, -ė̃: bėplauki, -ę is taken as the basis for deriving beplaukis, -ė, the root pitch accent of the derivative should change; cf.

bedė́vis: bedievỹs, bė́dievį = hairless: *beplaukūs, bėplauki.

Thus, diachronically, what can be discerned here is not an automatic alternation of the pitch accents of a single morpheme, but a shift of stress characteristic of the entire derivational stem { *bėplauk-} → { *bėplauk-}. This would have brought about the emergence of the circumflex in the root of the derivative: { *bėplauk-} → { beplauk-}.

§27. Apparently, pakalnė (*pàkaln-ē > *pakaln-ē > pakalnė) arose in a similar way. The examples recorded by K. Būga (see §25) show that such changes may have occurred not only in derivatives with an oxytone root, but also in those with a barytone root. From the perspective of word formation, the alternation of the pakõjė (:kõja) type (from *pàkājē) is similar to the alternation of kubìlias (: kūbilas), although the emergence of the circumflex alternation of pitch accents itself is usually explained by a stress shift in the opposite direction (cf. Stang, 1966, p. 145; Stundžia, 1981, pp. 63–65).

§28. M. Daukša’s variants svetimmõteris (ʃwetimmõteri, ʃwetimõterei, ʃwetimmõterius) and svetimmotēris (ʃwetimmotēris) (Skardžius, 1935, p. 130) show that at that time the accentuation of derivatives with metatony was not yet firmly established. Thus, stress shift could still occur in derivatives with the prefix pa- as well. Functionally, it is identical to the shift of stress onto the suffix of a formation during the development of suffixal stress in derivatives (cf. Coccop, 1977, p. 625; Girdenis, 1972, p. 66 ff.). It should therefore be dated to a similar period. That circumflex metatony in nouns with the prefix pa- is a recent phenomenon is also indicated by dialectal evidence: the material collected for the *Atlas of the Lithuanian Language* does not reveal a more continuous distribution area for the derivative pakalnė (LKA I 179). In general, the circumflex alternation of pitch accents is more firmly established in the Western Aukštaitian and Samogitian dialects (Lyberis, 1972, p. 43 ff.). Eastern and Southern Aukštaitians, by contrast, often pronounce derivatives without it—perhaps preserving a more archaic system of derivative accentuation.

§29. Circumflex metatony is sometimes associated not only with a change in the morphological position of the prefix (or stem-final element), but also with a formation’s transition from one part of speech to another (Brugmann, 1906, pp. 112–113; Wijk, 1958, p. 34; Skardžius, 1943, p. 456).

§30. No derivatives with the prefix po- (< *pā-) have been found in the accentuated old texts. Several nouns are recorded in F. Kurschat’s dictionary (pódùkra, pópietis,

26 Apie šių vedinių reikšmės skirtumus vienoje tarmėje plačiau žr. (Stundžia, 1981, p. 61).

126

pólunis). Thus, their archaic character can be established only on the basis of data from other languages, first and foremost Prussian; cf. Poducre, Passons, as well as Pomatre (see Mažiulis, 1981, p. 21).

In some cases, pó- may have arisen within Lithuanian itself, through changes to the stem-final segment and the strengthening of the prefix’s morphonological position (cf. M. Daukša’s nūpėhą, see §18) 27. Thus, nouns ending in -is and -ys are quite justifiably almost treated as accentual variants (Vitkauskas, 1976, p. 11).

The same applies to adjectives, whose dialectal variants reflect the development of their accentuation quite accurately; cf. 28.

pápiktas, -a (1) Šts, Als; pápiktas, -à (3ᵇ) J; papiktys, -ẽ (3ᵇ) Jdr, Vž → pópiktis, -ė (1) Rt, Trgn, Klt; papiktas, -à (4) Als

§31. Thus, several accentual systems of derivatives with the prefix *pā- are still in use in the Lithuanian-speaking area today (see Scheme 3) 29.

Scheme 3

Examples: pasáulis (1) póakis (1) pāsaulis (1) paakys (3ᵇ); patėvis (1) póspamis (1) pātevis (1) pasparnė (3ᵇ); pabėržis (1) [paberžė (3ᵇ)] pāšonis (1) [pāšparnis (1)]

The prefix *pā- and the fixed-stress paradigm seem to be more widespread in the northwestern Samogitian area and are therefore probably historically linked to the Curonians (cf. Salys, p. 287: map). Alongside *pā-, South and East Aukštaitians tend to use pó- uniformly in derivatives, which can provisionally be called the Prussian allomorph. The accentual system of Central Aukštaitian derivatives is difficult to reconstruct because of the mixing of systems, but, judging from M. Daukša’s writings, it is characterized by the prefix *-pā-.

²⁷ Particularly pronounced tendencies toward morphonological alignment are seen in verb-based nouns formed with endings; cf. pāsukos (3ᵇ), pósūkis: pasūko.

²⁸ Istoriškai senesniais laikytini šakninio kirčiavimo variantai (plg. Ambrazas, 1992, p. 17).

²⁹ Dialectal accentual variants influenced by analogy are given in square brackets.

127

§32. Indeed, not all derivatives can be linked to *pā- either: the fixed-prefix-accentuation variant pasáuls would have to be derived from the older form pasaũlis, which retained the accent on the baritone acute root. It would then be difficult to explain the persistence of the mobile-accent paradigm in the system of the derivatives under discussion.

Thus, diachronic analysis of derivatives with the prefix *pā- shows that, by the time they were formed, both allomorphs already existed — *pā- and *pā-, the latter apparently having arisen from the earlier short *pā-.

Sources and Literature

Ambrazas S. Baltų ir slavų kalbų vardžodžių daryba (senosios bendrybės ir skirt ybės) // Baltistica, 1991. T. 27(1). P. 15—34.

Bammesberger A. Zur Herkunft der baltischen Ablautstufe ā // Linguistica Baltica, 1. Warszawa, 1992. P. 49—53.

Brugmann K. Grundriss der vergleichenden Grammatik der indogermanischen Sprachen. T. 2. D. 1. Strassburg, 1906.

Buch T. Die Akzentuierung des Christian Donelaitis / Prace językoznawcze, 25. Wrocław-Warszawa-Kraków, 1961.

Būga K. Die Metatonie im Litauischen und Lettischen // Būga K. Rinktiniai raštai. T. 2. V., 1959. P. 386—483.

Būga K. Lietuvių tauta ir kalba bei jos artimieji giminaičiai // Būga. K. Rinktiniai raštai. T. 3. V., 1961. P. 85—282.

Endzelin J. Altpreussische Grammatik. Riga, 1944.

Endzelīns J. Latviešu valodas gramatika. Rīga, 1951.

Фортунатов Ф. Ф. Об ударении и долготе в балтийских языках. II. Долгота в латышском языке // Periferinės lietuvių kalbos tarmės (Lietuvių kalbotyros klausimai. T. 30). V., 1993. P. 181—213.

Fraenkel E. Litauisches etymologisches Wörterbuch. T. 1, 2. Heidelberg-Göttingen, 1962—1965.

Girdenis A. Lietuvių kalbos vardažodžio priesagų kirčio susiformavimas // Baltistica, 1972. 1 priedas. P. 66—72.

Girdenis A. Dvikirčiai žodžiai M. Daukšos ”Postilėje” // Kalbotyra, 1984. T. 35(1). P. 108—109.

Kabelka J. Kristijono Donelaičio raštų leksika. V., 1964.

Karaliūnas S. Kai kurie baltų ir slavų kalbų seniausiųjų santykių klausimai // Baltų ir slavų kalbų ryšiai (Lietuvių kalbotyros klausimai. T. 10). V., 1968. P. 7—100.

Kazlauskas J. Lietuvių kalbos istorinė gramatika (kirčiavimas, daiktavardis, veiksmažodis). V., 1968.

Klimavičius J. Dėl kai kurių žodžių literatūrinės formos ir kirčio // Iš lietuvių leksikologijos ir leksikografijos (Lietuvių kalbotyros klausimai. T. 12). V., 1970. P. 189—193.

Kurschat Fr. Grammatik der litauischen Sprache. Halle, 1876.

Kurschat Fr. Wörterbuch der litauischen Sprache. Halle, 1883.

Lyberis A. Metatonijos reiškiniai vediniuose // Kalbos kultūra, 1972. Sąs. 22. P. 43—48.

LKA — Lietuvių kalbos atlasas. T. 1—3. V., 1977—1991.

LKŽ — Dictionary of the Lithuanian Language. Vols. 1–2. V., 1968–1969; vols. 3–15. V., 1956–1991.

Mažiulis V. Prūsų kalbos paminklai. V., 1981.

Mikulenienė D. Metatonija morfonologijos požiūriu // Kalbotyra, 1989. T. 40(1). P. 45—51.

Mikulenienė D. The development of the prefix *pā- in derived nouns // Lithuanian Dialects and Their Study. Past and Present / To Antanas Salys (1902 07 21 — 1972 07 31). Conference abstracts. V., 1992. P. 12—13.

128

Mikulėnienė D., Žeizienė J. The accentuation of prefixed nouns from the perspective of morphology // Kalbotyra, 1994. Vol. 42(1). (in press).

K. Mūlenbach. Latviešu valodas vārdnīca. Redigējis, papildinājis, turpinājis J. Endzelīns. T. 1–4. Rīgā, 1927—1929.

Naktinienė G., Paulauskienė A., Vitkauskas V. Druskininkų tarmės žodynas. V., 1988.

Petruskas J., Vidugiris A. Lazūnų tarmės žodynas. V., 1985.

Otrebski J. Grammar of the Lithuanian Language. Vol. 2: The Study of Word Structure. Warszawa, 1965.

Salys A. The Žemaitian Dialects. Part 1. History of the Žemaitian Language Area // Salys A. Raštai. Vol. 4. Lithuanian Dialects. Roma, 1992. P. 141—286.

Skardžius P. Daukšos akcentologija. K., 1935.

Skardžius P. Lietuvių kalbos žodžių daryba. V., 1943.

Skardžius P. Lietuvių kalbos kirčiavimas. Čikaga, 1968.

de Saussure F. Lithuanian accentuation // de Saussure F. Selected Works on Linguistics. M.,1977. P. 620—632.

Stang Chr. S. Comparative Grammar of the Baltic Languages. Oslo-Bergen-Tromsø, 1966.

Stundžia B. On circumflex metatony in derived Lithuanian nouns // Baltistica, 1981. Vol. 17(1). P. 58-65.

Stundžia B. Accentuation variants of disyllabic nouns in Lithuanian // Kalbotyra, 1984. Vol. 35(1). P. 87—90.

Vanagas A. Lietuvos TSR hidronimų daryba. V., 1970.

Vanagas A. Lietuvių hidronimų etimologinis žodynas. V., 1981.

van Wijk N. The Baltic and Slavic Accent and Intonation Systems, 2nd ed. ’s-Gravenhage, 1958.

Vitkauskas V. Šiaurės rytų dunininkų šnektų žodynas. V., 1976.

Zinkevičius Z. Historical Grammar of the Lithuanian Language. Vol. 2. V., 1981.

Zinkevičius Z. From Lithuanian Historical Accentology. The Accentuation of the 1605 Catechism. V., 1975.

THE ACCENTUATION OF DECLINABLE DERIVATIVES WITH THE PREFIX *pā-

Summary

Historically, the preposition po (< *pā-) is associated with two prefixes, po- and pa-, which function as allomorphs in the modern language.

The article discusses the causes of the different accentuation of prefixal derivatives: 1) the development of the prefixes themselves; 2) the differing accentuation of the base words; 3) territorial or dialectal peculiarities of the derivatives.

Alongside derivatives with pa- (to which the derivatives pasáulīs, pósparnīs belong), pā- is reconstructed (pāsaūlis, pāspārnīs). The latter is associated with derivatives having mobile accentuation (paspárnē/paspārnȳs: pāspārnē/pāspārnī).

The nouns in which acute metatony occurs synchronically (pakámpē kàmpas) are treated in the article as variants that preserve the archaic accent in the root (cf. dialectal kám pa).

Circumflex metatony is associated with a change in the accent paradigm of the base (pakálnē: kálnas, -ai = pakalnē: kálnas, -ai). In such derivatives it is regarded as a relatively recent phenomenon, which could have arisen not only as a result of a change in the stem-forming suffix, but also through the transfer of accent from the prefix to the root of the derivative (cf. *pākalnē, > *pakalnē > pakalnē).

129