LEKSIKOS TYRINĖJIMAI

LIETUVOS TSR MOKSLŲ AKADEMIJA

SEMANTICS AND ETYMOLOGY

SIMAS KARALIŪNAS

1. Some Questions of Semantics

Human beings live and act in a sea of linguistic phenomena. Language, through the totality of its meanings, constitutes an environment vital to human existence itself and is one of the components of human semiotic behavior. “Le sens apparaît toujours comme une donnée immédiate: cela suffit à l'homme pour vivre et agir dans un monde signifiant”, asserts one of the most prominent semioticians of our time, A. Greimas¹. The extent to which human thought and behavior are conditioned by language was demonstrated by the Sapir–Whorf hypothesis and L. Weisgerber’s linguistic relativism. The theory of “semantic fields,” through which L. Weisgerber developed the above-mentioned principle of linguistic relativism, was important not only for the study of the relations language–thought and language–culture, but also provided semanticists and lexicologists with a structural method. Consequently, instead of the two problems addressed by classical semantics—the classification of meanings and the discovery of the laws of semantic change—the problems of descriptive semantics and the elucidation of semantic structure have come to the fore in recent decades. In answer to the question of how far meanings can form a structure, L. Hjelmslev pointed out that they may be regarded as forming a structure for two reasons: 1) because partial meanings (significations particulières) depend on the variants of enumeration, which are logically derived from possible relations envisaged in describing form; 2) because general “meanings,” or sememes, depend on values (de valeurs), which likewise arise from form and define possible correlations². Despite the fact that word meanings belong to that area of language which is most difficult to systematize and formalize, that it is often unclear where one segment of the content plane ends and another begins, and that subjectivism and arbitrariness therefore find fertile ground in this area, it nevertheless seems possible to study the lexicon and meanings using more precise methods of structural linguistics as well. Semantics is enabled to do so by F. de Saussure’s brilliant conception of the two planes of language—the content plane and the expression plane (plan du contenu et plan de l’expression)—and by their subsequently established isomorphism, which makes it possible “à concevoir la structure sémantique comme une articulation de l’univers sémantique en unités significations minimales (=ou sèmes), correspondant aux traits distinctifs du plan de l’expression (=ou phèmes)“³.

On the basis of these cardinal principles of general linguistic theory, the task is to obtain clear information about the number and nature of minimal semantic elements (seman-

¹ A. J. Greimas, Du sens. Essais sémiotiques, Paris, 1970, p. 39.

² L. Hjelmslev, Pour une sémantique structurale, in: Essais linguistiques (=Travaux du cercle linguistique de Copenhague, vol. XII), Copenhague, 1959, p. 108.

³ A. J. Greimas, op. cit., p. 40.

5

tic primitives) and their relationships, while describing at least a limited number of them at present; the structure of a semantic theory is postulated, and efforts are made to create a semantic theory based on the principles of generative grammar.

Whatever theory of structural semantics we may construct, the study of structural semantics, it seems, depends primarily on the structural study of the lexicon. A synchronic description of a language’s (for example, Lithuanian’s) vocabulary (ideally, such a description would be perfect) should provide at least two kinds of information: an inventory of simple and complex lexical units, and the structural models into which those units enter. The inventory should include: 1) the morphs that exist in the language (of course, absolute completeness cannot be achieved), 2) all words (conventional combinations of morphs), and 3) all phrases (fixed groups of specific words that differ from a mere sum of words). Each item in the inventory should be identified in some way, by indicating the function of a morph or word, giving a synonym or an equivalent expression.

The second type of information, as mentioned, would consist of an indication of the structural models, or paradigms, into which lexical units enter. Structural paradigms can be established by presenting the lexicon in the form of a “network.” What this means can be seen from the fragment of an inventory given here, which consists of lexical items (phonological and actual) having the structure CV in the first syllable, followed by the sequence of phonemes (ras).

Consonants Vowels a e i u ā ē ī ū ė o etc. k karas keras kiras kuras karas keras kyras kūras kėras koras g garas geras giras guras garas geras gyras gūras gėras goras t taras teras tiras turas taras teras tyras tūras tėras toras d daras deras diras duras daras deras dyras dūras dėras doras p paras peras piras puras paras peras pyras pūras pėras poras b baras beras biras buras baras beras byras būras bėras boras etc.

Such a presentation of the lexical “network” would show: 1) a cell for every mathematically possible phoneme combination, 2) a sphere of cells left unfilled because of phonotactic constraints (for instance, kāras, kėras, gāras, gėras, etc. are automatically excluded, since a short stressed word-initial vowel is impossible in Lithuanian nouns, but cf. pėras “a title of this kind”) and 3) cells that are in fact filled by morphs and words and are not⁷. Such a simple mathematical list of phoneme combinations would soon become enormously large and unwieldy. However, by applying phonotactic rules, it

⁴ F. P. Dineen, The Structure of Semantic Elements, Actes du xᵉ congrès international des linguistes, II, Bucarest, 1970, p. 589–593.

⁵ Jerry A. Fodor, Jerrold J. Katz, The Structure of a Semantic Theory, Language, 39, 1963, p. 170–210.

⁶ S. Abraham, I. Evseev and V. Serban, The Bases of a Structural Semantics, Actes du xᵉ congrès…, II, p. 609–611.

⁷ F. G. Cassidy, A Descriptive Approach to the Lexicon, in: Approaches in Linguistic Methodology ed. by I. Rauch and Ch. T. Scott, Madison, Milwaukee, and London, 1967, p. 9–15.

6

would decrease considerably if all unfilled cells were eliminated. Even so reduced, it would still remain quite large; nevertheless, this method of lexical investigation may be useful, since it would provide a complete inventory of the language’s morphemes and their combinations, reveal structural relations within the vocabulary, and help identify trends in the development and expansion of the lexicon. Computers would greatly facilitate the processing work. In the view of F. Cassidy, the author of this method of lexical investigation, an ideal description of the lexicon and its structural study are now “within the realm of possibility”⁸.

A structural description of the lexicon would, one may suppose, also serve semantics. It would then be possible to ask whether expression-plane segments belonging to one structural paradigm, or to phonologically similar paradigms, sometimes manifest content segments belonging to one or adjacent domains of sensory experience and thought. A structural description of the lexicon might also allow more precise investigation of such complex semantic phenomena as polysemy and its relationship to homonymy, which, as is well known, alongside synonymy, word motivation, and word character, are the principal subject matter of descriptive semantics.

Although logicians regard polysemy as an obstacle to communication and thought, it is a powerful means of linguistic economy, since it greatly eases the burden on our memory. The multiple meanings of words arise for several reasons. Contextual variation in a sememe stems from the simple fact that in different utterances a word never has exactly the same meaning. For example, the word knyga has a concrete meaning, and it may seem that there is little room for variation in its meaning. Yet a book means one thing to its author, another to its publisher or typesetter, and still another to a bookseller. It means one thing to a book collector, another to an antiquarian, still another to a bibliographer or librarian, and something quite different to a reader. Depending on the context and situation in which the word is used, it can—and usually does—have different shades of meaning. Some of these shades are ephemeral; others may develop into stable nuances of meaning which, over time, as the distance between them grows, may become distinct meanings of the word. Metonymy and metaphor also play a major role here. For example, akis, in addition to its primary meaning, “organ of sight,” has 21 meanings that arose through the transfer of a name based on various associations: “1. a mesh opening in a net; 2. a loop in knitting; 3. a loop, hole in a heddle; 4. a snare for catching hares, foxes, and other small animals; 5. a small hole cut in the ice or a place that has not frozen over; an ice hole; 6. an open patch in a bog; a small area of water in a marsh or a lake that is becoming overgrown, etc.”

Since the functional load of a word’s meaning cannot be measured, it is impossible to say how much the meanings of a single word differ from one another, or how much they differ from the meanings of another word. In other words, we cannot know where polysemy ends and homonymy begins. Here the lexicographer must rely purely on instinct. It is true that criteria for distinguishing polysemy from homonymy have been proposed, but they are only auxiliary and cannot solve the problem. For example, it is thought that if two words of identical form are used in one and the same sentence without their repetition being noticed, the speaker perceives no semantic relationship between them. For instance, French pas ‘step’ and pas ‘not’ occur in a single sentence: Cela ne se trouve pas dans le pas d’un cheval “You don’t find this every day.” Thus, in present-day French these are two different words, although both derive from Latin passus.

⁸ F. G. Cassidy, op. cit., p. 14.

7

‘step’. It is also argued that, supposedly, no self-respecting poet would rhyme two words that he considers to have the same meaning⁹. In this connection, one might propose another criterion for distinguishing polysemy from homonymy. If one allomorph manifests the same or similar content segments as another allomorph, the content segments they manifest can be considered to belong to one word (rather than to two homonymous words). For example, kuópti has the following meanings: “1. to clean, make clean; tend; 2. to harvest, gather a crop; 3. to take honey, harvest honey; 4. to clean flax of shives, break it; 5. to heap up, dig, scoop; 6. to bury; 7. to hide; 8. to care for; 9. to plunder, seize, grab; steal; 10. to eat heartily, eat a lot, slurp, devour, shovel in; 11. to run, go quickly, dash” (see LKŽ VI 916). A group of very similar meanings is also found under verbs 1 kópti, 2 kópti|kópti, 3 kóp-til|kópti (see LKŽ VI 351–4). If so, the verbs kópti|kópti and kuópti are evidently allomorphs, since, moreover, the alternation uo: o is entirely regular in Lithuanian. Thus, instead of three homonyms, we will probably have one polysemous verb, kópti|kópti: “1. to clean, tidy; 2. to harvest, gather (grain from the field); 3. to take honey (from a hive, tree hollow); 4. to scoop, shovel, load manure onto a cart with a pitchfork, muck out; 5. to hoe up and mound (potatoes, vegetables…); dig into a heap, make a mound, mound up; scatter; 6. to bury; 7. to hide; 8. to take, plunder, seize, grab; 9. to eat greedily, shovel food in; 10. to go, travel” (alongside the homonym kópti “to climb (up, down, over something…)”), since its relationship to the verb kópti|kópti “to clean, tidy; to harvest, gather (grain from the field), etc.” is not apparent on the synchronic plane. Moreover, the following phenomenon observed in Lithuanian semantics may also be relevant here: the sememes ‘(greedily) eat, devour, swallow, wolf down’, ‘go, travel, run’, ‘beat, strike, hit’, and ‘steal’ often occur alongside various other sememes. In other words, a common Lithuanian verb is used metaphorically to denote the actions mentioned; compare, for example, dróžti “to cut, sharpen, smooth with a knife or another sharp tool; beat, strike, lash, give a blow; eat, wolf down, drink; go, run, ride, travel fast, scrape, etc.” LKŽ II² 747. Apparently, Latvian I žaút ‘to dry, hang out to dry’ and II žaut ‘to strike, deliver a heavy blow, etc.’ ME IV 792–3 are, not only diachronically but also synchronically, one and the same word (cf. Lithuanian džiáuti ‘to hang, put out to dry; steal; strike, hit, etc.’ LKŽ II² 1016); see below for this. Similarly, for example, Lithuanian káuti ‘to go, travel’ LKŽ V 450 is hardly a different verb from káuti ‘to beat, kill; fight; forge (in a smithy)’ (cf. the semantic relations found in dróžti). It is true that, in such cases, we must still consider how lexicographical practice and synchronic semantics should proceed: perhaps it is indeed more appropriate to distinguish two or several words and treat them as homonyms, since in many cases the meanings are so far apart that only careful etymological analysis makes it possible to trace their original connection and thereby establish polysemy (see below).

In examining questions of descriptive (synchronic) semantics, we have arrived at matters of diachronic semantics. Semantic analysis, as is well known, alongside phonological and derivational analysis, is one of the principal aspects of etymology. In investigating the phenomena of polysemy and homonymy diachronically, and in explaining the etymological relationships between lexemes, great importance attaches to a principle formulated by B. Jėger: “… several widely divergent meanings (A, B, C …) meet in one

⁹ S. Ullmann, Semantics. An Introduction to the Science of Meaning, Oxford, 1962, p. 179.

8

If the phonetic form is taken together, this fact must be established. If this series of meanings A, B, C… occurs elsewhere, there can scarcely be any further doubt that a semantic relationship exists between the meanings A, B, and C, even if it is not initially recognizable”¹⁰. B. Jēgers demonstrated the validity of this principle throughout the work cited here, the finest diachronic study of Baltic semantics. This principle presupposes that, in the content plane of a polysemous word, meanings have arisen organically from one another, and that they represent a regular semantic development. Most cases undoubtedly show just that. However, the coexistence of two content segments in the content plane of a single word, once established, need not indicate regular development of the segments. For a word may have borrowed its content segment from another word. This most often happens when words have a similar phonetic form. Here we encounter cases of so-called “pseudo-semantic development.” Specifically, the fact that French essuyer means: 1. “to wipe” and 2. “to endure, bear” does not show that these two content segments it manifests arose from one another, since the verb acquired the segment “to endure, bear” from essayer “to try, attempt; to strive,” which in the sixteenth century also meant “to experience, bear, endure”¹¹. Latv. vuobit means both “to draw the lips together to laugh” and “to peel, shell (e.g., nuts, peas)” ME IV 677. These content segments are too far apart for parallels to be found in other languages and thus to confirm their diachronic relationship. There is no doubt that vuobit acquired the content segment “to peel, shell” from the rhyming word luōbit “to peel, pick out, shell.” We see a complex semantic infiltration in the case of Lithuanian plúoštas and plūoštas: plúoštas “1. textile raw material, thin and strong fibers; 2. raft, ferry” must have acquired its first meaning (cf. the allomorph pláustas, which lacks this meaning) from plūoštas “textile raw material, thin and strong fibers; (of hay, straw, hair, etc.) a bunch, handful, small armful; bouquet, bunch; strip, streak; ferry, raft,” while plūoštas, which is etymologically related to the verb plėšti, apparently borrowed the meaning “ferry, raft” from plúoštas. Their very similar phonetic forms (differing only in the phonemes š and s) apparently caused their content planes to begin to converge as well. Thus, before asserting that the relationship between any two segments in a single content plane is the result of regular semantic development, one must—as far as possible—determine whether it is instead a case of “pseudo-semantic development.”

2. The etymology of Lithuanian džiaugtis

Lithuanian džiaugti “to delight,” džiaugtis “to feel joy, to enjoy oneself” LKŽ II² 1014 has Latvian counterparts džaugt “to delight,” -tiēs “to enjoy oneself,” gáugtîes “to rejoice, exult; hope, expect; promise, be about to; complain; caress, fondle” ME I 694, “to snuggle up” EH I 425. Although these Lithuanian and Latvian cognates were noted long ago, Lithuanian džiaugti(s) and its Latvian relatives still lack a reliable etymology, apart from one or another unsuccessful attempt¹².

¹⁰ B. Jēgers, Verkannte Bedeutungsverwandtschaften baltischer Wörter, K Z, 80, 1966, p. 10–11.

¹¹ J. Orr, Words and Sounds in English and French, Oxford, 1953, p. 157 tt.

¹² E. Fraenkel, Litauisches etymologisches Wörterbuch, Heidelberg-Göttingen, 1962, p. 116–7.

9

H. Hirt¹³ connected Lithuanian džiaugiuosi with Latin gaudēo and Greek γηθέω “I rejoice,” and, assuming that there was a consonant metathesis here, derived it from *gaudžiuosi. The older form of the Latin and Greek words appeared to be *gāu-edh-ējō (or *gāu-adh-ējō)¹⁴. Thus, bearing in mind the conservative character of Lithuanian, we would expect *goved- (or *govad-), not *gaud-. Moreover, it is easy to show, on the basis of other words with the same root in Lithuanian and Latvian (see below), that there was no metathesis here.

P. Persson’s¹⁵ attempt was also unsuccessful: he compared Latvian gaugtiês with Old High German goukolōn, gouggolōn “to fool around, play tricks,” and Latvian gaūbt² with Middle High German gief “fool.” First, as can be seen, the meanings of the words being compared are markedly distant, and—most importantly—the etymological initial consonant of the Latvian words’ root is *d- (not *g-; see below).

Before proceeding to the etymological explanation, it should be noted that the genetic identification of Lithuanian džiaugti(s) and Latvian džaugt(ies), gauties, which seems fairly obvious, encounters phonetic difficulties. Lithuanian džiaugti(s) regularly continues the older root *deug-: the diphthong eu regularly became jau, and the sequence *dja- changed to dž’a (cf. Lithuanian biaurùs, Latvian blaũrs from *beur-). The reflexes of the sequences *dj, *tj before back vowels in Latvian are ž, ž¹⁶. Such sounds are precisely what we find in the words belonging to the root under discussion; cf. Latvian ap-žaugties “to wind around (another tree) as it grows,” apžaugties kam ap kaklu “to wind oneself around someone’s neck” EH I 128, žaudīties (present -guôs, preterite -dzījuos) “to impose oneself; to be affectionately around someone, to caress, embrace, kiss (said chiefly of small children); to caress one another” ME IV 790, žauguoties “the same” ibid., žaugas “caress; feeling, sensation” ibid., žaugule “sweetheart, mistress” ibid., žaugzmas “feeling, sensation” ibid. (cf. Lithuanian džiaūgsmas “a feeling of great satisfaction or pleasure, merriment, happiness” LKŽ II² 1013). These Latvian lexemes should be regarded as regular continuants of the root *deug-.

Because we find dž- instead of ž-, J. Endzelīns (EH I 563) was inclined to suspect that Latvian džaugt(iês), džauga were Lithuanian loans. In fact, these forms are not used in living Latvian; they are known only from older Latvian dictionaries (Lange’s “Lettisch-deutsches Lexikon,” 1773, Stender’s “Lettisches Lexikon,” 1789, Ulmann’s “Lettisch-deutsches Wörterbuch,” 1872). However, neither the presence of dž- instead of the regular ž- nor the fact that these forms are not used in living speech seems sufficient grounds for regarding them as Lithuanian borrowings. There are also other cases in which Latvian has dž, č where Lithuanian has dž, č; cf., for example, Latvian džàut, čuôp², čàpt, čùpt, čupinêt, čuožātiês, čūzinât and Lithuanian džiauti, čiuopti, čiopti, čiūpti, čiupinėti, čiuožti, čiūzinėti. It would be too bold to declare all of these Lithuanian borrowings. All the more so since the reflexes dž, č of the old sequences *dj, *tj are also attested in Latvian itself (in some of the Semigallian dialects)¹⁷. Forms with dž, č from these dialects may have entered the old Latvian dictionaries or spread to other dialects.

¹³ H. Hirt, BB, 24, 1899, p. 280: “Aus dem lit. stellt sich zu γηθέω und gaudeo džiaugiuos aus *gaudžiuos”.

¹⁴ A. Walde- J. B. Hofmann, Lateinisches etymologisches Wörterbuch, I, Heidelberg, 1936, p. 584.

¹⁵ P. Persson, Beiträge zur indogermanischen Wortforschung, Uppsala, 1912, p. 59 tt.

¹⁶ J. Endzelīns, Latviešu valodas gramatika, Rīgā, 1951, p. 177 tt.

¹⁷ M. Rudzīte, Latviešu dialektoloģija, Rīgā, 1964, p. 91.

10

Phonetic development *deug- > lat. žaug- an intermediate stage *d’aug- is probably reflected in the form g’augtiės. Since there is evidence allowing the reconstruction in Curonian of d’, t’, < *d̑j, *t̑j¹⁸, g’augtiės may be considered to have entered Latvian from Curonian¹⁹, since in Latvian palatal d’, t’ in loanwords are reflected as g’, k’. The verb gaugtiės is indeed attested in the western part of Latvia (Dunika). Thus the root *deug-, meaning “to rejoice” or something close to it, is also attested in the third Baltic language—Curonian—in which, it may be assumed, the diphthong *eu had likewise already become jau.

In Lithuanian, the zero-grade variant of the root džiaug- < *deug- is found in džiūgti ‘to begin rejoicing’ LKŽ II² 1031, džiūga ‘joy’ ib. 1029, džiūgà, džiūgas ‘someone who is quick to rejoice at something’ ib. 1029, džiūgùs ‘one who soon grows joyful, rejoices, is merry’ ib. 1031, džiūgauti ‘to rejoice greatly’ ib. 1029, etc. The presence of the affricate dž before a back vowel shows that these forms probably arose on the basis of the full-grade root variant džiaug-, on the model šaukti: šūkti = džiaugti: x. This is confirmed by the fact that they have no Latvian counterparts.

The Latvian forms džiaugti(s), ap-žaugtiės, džaugt(ies), g’augtiės may help clarify the further connections. Latv. žaugātiės² ‘to climb or go without thinking; (of cows) to leave the herd and run into the grain, to jump over ditches’, žauga² ‘a cow that repeatedly leaves the herd and goes into the grain, jumps over ditches, etc.; one who goes where it is forbidden’, is related to the root-final -k- form žauka ‘a high-spirited person; someone who climbs here and there; a prostitute, etc.’ ME IV 790.

That the basis of these latter Latvian words is Latv. žauties, žaunuōs, žāvuōs ‘to climb; to rise high; to frolic, be exuberant; to run; to go quickly despite all obstacles; to run heedlessly at or over something; to go somewhere for no reason; to fall, tumble’ ME IV 793, ‘to go somewhere headlong, to shove; to go somewhere secretly; to climb; to hurry, run; to crawl through something’ EH II 817, is quite clear. Compare Latv. guovis pirmuo dienu ganuos žaugājas… ‘they jump onto one another’s backs’ ME IV 790 s. v. žaugātiės², where the semantic connection between the verb žaugātiės² and žauties is plainly evident. Latv. žauties corresponds in Lithuanian to džiūti, džiūna, džiovė ‘to run, go quickly; to do something energetically’ LKŽ II² 1016. In terms of meaning, the segment ‘to fall, tumble’ in Latv. žauties stands somewhat apart. However, that this is a contextual variant, probably arising from metaphorical use of the verb, is clearly shown by Lithuanian Kad džiovė (griuvo) aukštienykas, net kojos užsiriėtė Žiežmariai (see LKŽ. II² 1016 s. v. džiūti). The relatives of Latv. žauties and Lithuanian džiūti in other Indo-European languages are Skt. dhavate ‘runs, flows’, Gk. ϑέω ‘I run’, ϑοός ‘swift’ (IE *dheu- ‘to run, flow’²⁰).

From the standpoint of derivation, the important fact here is that the Baltic root *dēu- > *djāu- ‘to run, go quickly; to flow, etc.’ was expanded by the formant -g-. It should therefore be noted that in Latvian this formant is also found in lexemes meaning ‘to dry’ or containing related semantic segments. Compare Latv. žugt, žugstu, žūgu = žūt ‘to dry’ ME IV 835, žugls ‘lean’ ib. 829 (for the connection between the semantic segments, compare Lithuanian džiūti ‘to become dry, dry out; to grow thin, diminish, etc.’ LKŽ II² 1035). Thus, the fact that the lexemes žaugātiės² ‘to climb or go without thinking; to leave the herd and run into the grain, to jump (over ditches),

¹⁸ J. Endzelin, Über die Nationalität und Sprache der Kuren, FUF, 12, 1912, p. 59—72; K. Būga, RR, III, 1961, p. 172—4; V. Kiparsky, Die Kurenfrage, Helsinki, 1939, p. 253, tt.

¹⁹ This was already done by K. Būga, RR III 174, and J. Endzelīns, ME I 694 (the latter also allows for the possibility of borrowing from Lithuanian dialects).

²⁰ J. Pokorny, Indogermanisches etymologisches Wörterbuch, Bern, 1959, p. 259.

11

‘to jump’ and žūgt ‘to dry’ are formally related (they share the formant -g-!), which also permits us to seek a semantic connection between them—in other words, to suspect that they are etymologically identical. Lithuanian džiáuti ‘to hang or put out to dry; to dry; to steal; to chop; to strike’ LKŽ II² 1016 (for the relations among its semantic segments, see above), džiūti ‘to become dry, dry out; to wither, languish, perish (of plants); to grow thin, diminish’ ib. 1035, Latv. žaut ‘to dry, to hang up to dry’, žaut ‘to strike, deal a heavy blow; to pour; to pour, rain heavily’ ME IV 792—3 (for the fact that we have here one polysemous lexeme, see above), žūt ‘to dry’ ib. 838, are known to have several etymological hypotheses²¹. E. Berneker²² compared them with Skt. dunóti ‘burns’, davah ‘forest’ (< ‘burning’). J. Endzelin²³, assuming the original meaning ‘to dry out, wither, stiffen’, connected these Baltic words with OHG tawalōn ‘to disappear, die’, Old Norse dó ‘became stiff (of joints)’, OHG touwen, Old English dōian ‘to die’, Gothic dauþs ‘dead’. This connection encounters semantic and phonological difficulties (the Germanic words have au in the root). More recently, the Germanic words have been etymologized differently²⁴. On the basis that the sememes ‘to shine, gleam, glow’, ‘to heat, burn’, and ‘to dry’ are related (compare, for example, Lat. sūdum ‘clear weather; clear sky’, sūdus ‘dry; cloudless, clear’), F. Specht²⁵ hypothesized that the ancient Indo-European name for the sky, diēus ‘radiant, gleaming sky’ (Skt. dyáuh ‘sky, day’ and ‘god’, Gk. Ζεύς, Lat. diēs ‘day’), was still contained in Lithuanian džiáuti and Latvian žaut. Unfortunately, in view of the Baltic evidence presented below, to which modern methods of linguistic analysis give priority, we are compelled to abandon this interesting hypothesis²⁶.

The established derivational relationship between Latv. žūgt ‘to dry’ (: žūt ‘to dry’) and žaugāties² ‘to climb or go without thinking; to leave the herd and run into the grain, to jump over ditches’ (: žauties ‘to frolic, be exuberant; to run; to go quickly despite all obstacles, etc.’) (they both have the formant -g-), as already noted, also allows us to seek a connection in their semantic domains. That the sememe ‘to dry, become dry’, whose intransitive–transitive opposition has ‘to hang, put out, cause to dry’ as its second member, and the sememe ‘to run, flow, pour’ may be historically related is first apparent from Lithuanian šėkti ‘to diminish, dry up, subside (of water); to disappear, dwindle, grow thin’, Latv. sikt (present sieku = Lithuanian senkù) ‘to run dry, fall (of water), dry up’, and Lat. sicāre ‘to dry, dry out; to drain; to drink up; to milk, suck out’. This assumption is further supported by the fact that words sharing a root with Baltic šėkti, sikt include not only Gk. (aorist) ἐ-σχ-ετο ‘ran out, was exhausted, dried up’ and Skt. á-sakrah ‘inexhaustible, unspent’, but also—and above all—Avestan hišku- ‘dry’ (from the reduplicated formation si-sk-us), Middle Irish sesc, Welsh hysb, Breton hesp ‘dry’ (from *sisk̑uo-)²⁷. Lat. siccus ‘dry-

²¹ See E. Fraenkel, LEW, p. 117.

²² E. Berneker, IF, 10, 1899, p. 158.

²³ K. Mühlenbachs Lettisch-deutsches Wörterbuch, IV, p. 793.

²⁴ J. Pokorny, IEW, p. 260.

²⁵ Fr. Specht, Der indogermanische Himmelsgott im Baltisch-Slavischen, KZ, 69, 1/2, 1948, p. 115—123.

²⁶ Compare also M. Mayrhofer’s comment: “Specht’s proposal of a diēu*- ‘to radiate, shine’—which is supposed to have yielded Lithuanian džiáuti ‘to hang up to dry’ and džiovà ‘drought’ via ‘to burn’—remains uncertain.” — M. Mayrhofer, Kurzgefaßtes etymologisches Wörterbuch des Altindischen, II, Heidelberg, 1963, p. 71.

²⁷ J. Pokorny, IEW, p. 894; M. Mayrhofer, op. cit., I, p. 64.

12

“sas” (from *sikkos with expressive gemination) and Av. hiku- “dry” are cognate with O.Ind. sécate, siñcāti, Av. hiñcaiti, hičaiti “pours out, spills”, OHG sihan “to strain” (Ger. seihen), OHG, OSax., OEng. sigan “to fall in drops, subside, flow”²⁸ (PIE *seik̑- “ausgießen, seihen, rinnen, träufeln” Pokorny IEW 893)²⁹. It. asciutto “dry” continues Lat. exsūctus, which is the past participle of Lat. exsūgo “I suck out, draw out” (: sūgo “I suck; absorb; soak up”). On the basis of these clear facts, the historical relationship between the sememes “to dry, become dry” and “to run, flow, pour” may be considered proven³⁰. And since “to run, flow” is used of both water and a person, cf., for example, a child runs and water runs, and the relationship between tekėti “to flow”: takas “path”, it thus also becomes clear that Lith. džiauti “to hang, put out to dry; steal; cut, strike”, Latv. žaut (žaut) “trocknen, zum Trocknen aushängen; schlagen, einen starken Schlag versetzen”, Lith. džiūti, Latv. žūt “trocknen” and Lith. džiauti “to run, walk quickly; do something energetically”, Latv. žauties “klettern; hoch steigen; tollen, ausgelassen sein; laufen; ungeachtet aller Hindernisse schnell gehen; neapduomīgi kam virsū jeb pāri skriet; unnüt., wohin gehen etc.” belong to one and the same root, especially since content segments “to run, flow” are also manifested in the very root under discussion, cf. Latv. žaut “giessen, stark regnen” ME IV 793. Moreover, the historical relationship established here between the sememes “to dry, become dry” and “to run, flow, pour” also shows that the latter sememe, or its alloseme “to run, hurry, walk quickly”, is primary. Having established this, we may return to the etymology of džiaugt(i)s, ap-žaugties, džiaugt(ies), gāugties.

In the content plane of the verbs examined here, the following principal groups of content segments stand out: 1) “to have joy, enjoy oneself; (sich) ergötzen, sich freuen, jubeln” (džiaugt(i)s, džiaugt(ies), gāugties), 2) “hoffen, erwarten; versprechen, im Begriff sein; sich beklagen” (gāugties); 3) “mylināties, liebkosen; glausties” (gāugties), “einander liebkosen; liebevoll um einen herumsein, küssen, umarmen, sich aufdrängen” (žaud zīties, žauguoties); 4) “wachsend (einen andern Baum) umwinden” (ap-žaugties).

The segments “sich aufdrängen” and “wachsend (einen andern Baum) umwinden” are fairly close. The genetic relationship of the sememe “to rejoice” to the content segments found in Latv. žaugāties² “die Herde verlassend (ins Getreide) laufen, (über Gräben) springen (von Kühen gesagt) etc.” and žauties “klettern; hoch steigen; tollen, ausgelassen sein; laufen etc.” can likewise be demonstrated fairly easily; cf. OHG frao, frō, frawēr “cheerful, joyful” (Ger. froh): OScand. frār “quick, swift, sudden, impetuous”, ME frow “quick, hasty” (from Germ. frawa-), O.Ind. prā́vate “jumps, leaps”; English exult “to rejoice, exult”, which formerly also meant “to leap for joy (to leap upward, especially for joy)”. This English verb continues Lat. ex(s)ultō “I leap, jump; seethe, rage, surge; delight in,”

²⁸ Cf. A. Walde–J. B. Hofmann, Lateinisches etymologisches Wörterbuch, II, Heidelberg, 1954, p. 533.

²⁹ Cf. also what J. Pokorny says: “daneben eine Wz. seik̑- ‘trocken’, die wohl über ‚abrinnen, versiegen = austrocknen‘ mit seik̑- ‚ausgießen‘ zu vereinigen ist”, — J. Pokorny, IEW, p. 894.

³⁰ Cf.: “A few (i.e. lexemes denoting ‘dry’) are from the notion of ‘extract the juice’,” — C. D. Buck, A Dictionary of Selected Synonyms in the Principal Indo-European Languages, Chicago, Illinois, 1949, p. 1076.

13

“I rejoice, etc.”; Lat. exsiliō “I leap out; jump up, etc.” is a frequentative. Its base is salio “I leap, jump”, cognate with Gr. ἄλλομαι “I leap”³¹.

The emergence of the other segments in the semantic field of “to rejoice” is more or less understandable: for “mylināties, liebkosen; glausties” cf. Lat. gaudeō “I rejoice, feel satisfaction” and gaudere aliquid “to love, like something”; for “hoffen, erwarten; versprechen, im Begriff sein; sich beklagen” cf. Latv. priecāties “sich freuen”, priekš “die Freude”, O.Ind. prināti “delights, makes glad; urges, incites, rouses; tastes, tries, makes use of; enjoys, admires, takes pleasure in”, Av. frīnaīti “desires, yearns for, wants, wishes”, Goth. frijōn “to love”, O.Slav. prijaja, prijati “to foresee, anticipate, consider beforehand; to be prudent, cautious; to care; to make an effort”.

Thus, there is good reason to think that the lexemes meaning “to rejoice” — džiaugti(s), ap-žaugtiēs, džaugt(ies), g’augtiēs — most probably derive from the base verb džiauti “to run, walk quickly; do something energetically”, žauties “klettern, hoch steigen; tollen, ausgelassen sein, laufen; ungeachtet aller Hindernisse schnell gehen; neapdairingi kam viršų jeb pāri skriet; giessen, stark regnen; schlagen, einen starken Schlag versetzen etc.” Lith. džiaūgti(s) has a short diphthong au (< eu) in the root, whereas džiauti, žauti show that there was a long diphthong ēu; cf. also pres. džiauna/džiaū-ja, žāuju/žāunu, pret. džiōvē, žāvu. However, the fact that alongside the root with the long diphthong there was also an allomorph with a short diphthong is shown by Latv. žavēt “trocknen, räuchern” ME IV 794 and žugls “mager” ib. 829.

3. Etymology of Lat. džaubt, g’aibt²

Besides ap-žaugtiēs, džaugt(ies), g’augtiēs, Latvian also uses lexemes with a root-final -b- to express the sememe “sich freuen”: g’aibju, g’aubu, g’aubt² “ergötzen, jubeln”, -ties “sich freuen, jubeln; hoffen, erwarten; versprechen, im Begriff sein; sich beklagen” ME I 694, gaubt “achten auf, sorgen für, besorgen (gādat)” EH I 425, g’aubas “das Wehklagen” ME I 694, g’aubigs “ergötzlich” EH I 425, džaubt “sich sorgen, Sorge tragen”, džaubtiēs “sich ergötzen; etwas zu tun versprechen” EH I 365. Here, as in the case of the root džaug-/g’aug-, dž-/g’- also alternate at the beginning of the root. Forms with dž- are used in the Augšzemnieki dialects (Bērzgale, Pilda, Zasa), while forms with g’- are concentrated more in the southwestern part of the Augšzemnieki dialect, near the Lithuanian border, also extending into the area of the central dialect (Bēzaune, Blīdienē, Kaldabruņa, Krustpils, Liepkalne, Nereta, Saikava, Sauka, Stūri, Sunākste, Susēja, Zasa). As can be seen, the geographical distribution of the forms under consideration tells us nothing definite about their origin. The forms with g’- apparently entered Latvian from the old Baltic dialects (perhaps also from Lithuanian), in which the clusters *dj, *tj may have developed into *d’, *t’. In any case, what matters here is that the alternation dž-/g’- presupposes an etymological *dj- at the beginning of the root. Thus, the lexemes under consideration have a root that earlier appeared as *djaub- < *deub-, differing from the root *d jaug- < *deug- only in its root-final consonant.

The content plane of the lexemes with the roots *deug- and *deub- is so similar, and their phonetic form so alike, that the idea of a genetic identity between these roots readily suggests itself. And indeed, precisely this view—that g’aubtiēs is a late remodeling of the verb g’augties, or a variant form involving root-final consonant variation

³¹ A. Walde-J. B. Hofmann, op. cit., II, p. 468.

14

variation, was also proposed.³² Cases of sporadic alternation between the consonants b/g and p/k can be found in both Lithuanian and Latvian; cf. Lith. bābas “horned cake, cake (in children’s speech)” LKŽ I² 538: bāgas “ditto” ib. 554; brýnas “without property, without work, bare” LKŽ I² 1050: grýnas “…naked; bare; empty; poor, destitute” LKŽ III 606–7; blúoksnis “willow” LKŽ I² 952: glúoksnis LKŽ III 441, glúosnis “a tree of the willow family; osier (Salix)” ib. 442; Latv. plùons² “threshing floor, wooden floor” EH II 305: kluõns “clay floor; threshing floor” ME II 238; praulêt “to become worm-eaten” ME III 379: kraulêt² “to rot” EH I 643; žup­snis “as much as one can grasp with 3 (or 2) fingers, a little, etc.” ME IV 832 (cf. Lith. žiùpsnis): žuksnis “a handful, as much as one can hold in one’s arms, etc.” ib. 829. However, as can be seen, these consonants alternate mainly at the beginning of a word and next to the sonorant consonants l, r, whereas the roots *deug-, *deub- have them at the end of the root. A phonetic derivation of these roots from one another is contradicted by the fact that 1) according to the data in K. Mühlenbach and J. Endzelīns’s dictionary, the root *deub- has no attested allomorph *žaub- corresponding to the allomorph -žaug-ties, and 2) segments appear in the content plane of the lexemes with the root *deub- (cf. “sich sorgen, Sorge tragen”, “sich beklagen; das Wehklagen”) that are absent from the lexemes with the root *deug-.

On the analogy of džiaùgti(s), ap-žaugtiẽs, džaugti(ies), gaugtiẽs, the simplest assumption would, of course, be that džàubt(ies), gàubt(ies) too are derived from žauties “to climb, ascend; frolic, be exuberant, run; go quickly regardless of all obstacles, etc.” However, the above-mentioned fact prevents this assumption: no attested allomorph *žaub- corresponding to the allomorph -žaugtiẽs has been found, with *žaub- identical to the root of žauties.

It cannot be accidental that the verbs džáubt-, bju-, -bu “to strangle” EH II 365 and gàubt “to slaughter, strangle” ME I 694 also have the same root-initial consonants. The connection between the sememes “to worry, take care” and “to strangle” can be compared with Latv. sirds man žnaudzəs ME IV 824 s. v. žnaùgt (: žnaùgt “to strangle, gag, bind tightly, squeeze hard; wring out”, Lith. žniaùgti “to press with a press, clamp”); with the semantic relations between English anxious “worried, uneasy; eagerly seeking, desirous”, Latin anxius “fearful, timid; uneasy, anxious, concerned; gloomy, morose, angry; frightening, tormenting” and Latin angō “I squeeze, press; oppress, strangle”; and with the fact that English worry “to worry, be concerned; to torment, annoy, etc.” continues Middle English worien, wirien “to seize by the throat; anger, irritate” and Old English wyrgan “to strangle, suffocate.” Thus, the genetic relationship between these sememes can hardly be doubted.

In both expression and content, Latv. džàubt “to worry, take care” is very close to Lith. džiaūbia, džiaūbė, džiaūbti “to walk with one’s nose down” (Gervėčiai) LKŽ II² 1012. The identification of the Lithuanian counterpart definitively confirms the idea advanced here: that Latv. džàubt(ies) “to take delight, etc.; to worry, etc.” and gàubt(ies) “to take delight, rejoice, etc.” have an etymologically different root from the lexemes ap-žaugtiẽs, džaugti(ies), gaugtiẽs.

Lith. džiūbti “to fall, plunge to the ground (with one’s full weight)” LKŽ II² 1029 is undoubtedly related to Lith. džiaūbti, as are džiūbinti “to beat, thrash; strike, chop,” džiūbyti “to peck, strike with the beak; beat, thrash; strike, chop, etc.” ib. 1028, and džiūboti “to strike, chop” ib.

³² J. Endzelīns, ME, I, p. 694 s. v. gàubt²; E. Fraenkel, LEW, p. 117.

15

1029, džiaubénti ‘to walk with one’s head bowed’ ib. 1028. Although these zero-grade lexemes were created relatively late on the basis of džiaubti—which explains why they have the affricate dž before u and why no Latvian cognates are found for them—they could nevertheless have been created while džiaubti still retained its older semantic segments, and thus they may reflect the older meaning of the root džiaub-/džiub- (‘fall, tumble to the ground’, ‘strike, beat; chop, hack’, etc.).

Lith. džiaubti, Latv. gaubt(ies), džaubt(ies) represent an allomorph with jo, the normal root vocalism e: *deub- characteristic of thematic verbs. Their relatives, representing the allomorph *doub-, may include Lith. daūbti ‘to hollow out, scoop; to strike, beat’ LKŽ II² 310, daubà ‘valley, hollow, ravine; hole, cave, cavern, nook; den’ ib. 309, daūbas ‘hole, cave, cavern, nook; den’, ‘hanger-on, sloven, idler’ ib. 310, Latv. daūba ‘the ravine’ ME I 443, Pr. pa-daubis ‘valley’ Elb. 30, and forms with zero-grade root vocalism: Lith. dūbti ‘to become hollow, develop a hollow, bend; sink, subside, plunge; split, decay’ LKŽ II² 785, Latv. dubt ‘sink in, cave in, become hollow’ ME I 509. They are all continuants of the IE root *dheu-b- ‘deep, hollow’ Pokorny IEW 267.

Important for the etymological identification of the lexemes mentioned is the semantic connection between daūbti ‘to hollow out, scoop; to strike, beat’ and džiaubti ‘to walk with one’s head bowed’, gaubt ‘to pay attention to, care for, attend to (gādāt)’, and džaubt ‘to worry, take care’, a connection clearly illustrated, for example, by Lith. rūpėti ‘to be a concern’, Latv. rūpēt ‘to cause concern, trouble’, Lith. raūpti ‘to be concerned, grieve’: raūpti ‘to dig, hollow out’, raupýti ‘to bore’ alongside Serbo-Croatian rūpa ‘hole, pit’, Old Scandinavian rauf ‘hole’³³.

The genetic relationship between the sememes ‘to worry, take care’ and ‘to strangle’ can be demonstrated fairly easily (see above), whereas the historical relationship between the sememes ‘to worry, take care’ and ‘to rejoice, exult, delight’ must be said to remain rather unclear. Alongside the sememe ‘to worry, take care, attend to’, which, as already mentioned, developed on the basis of the sememes ‘to hollow out, scoop, bore’ and ‘to strike, beat, slaughter, strangle’ (cf. Latv. džaubt(ies) ‘to strangle’, gaūbt(ies) ‘to slaughter, strangle’), the sememe ‘to rejoice, exult, delight’ could have arisen in two ways. One may suppose that semantic polarization played some role here, based on the semantic universal ‘to become, turn into the opposite of what one is’; compare, for example, the well-known Lith. brėkšti ‘to dawn, grow light’: Samogitian ‘to grow dark’; smagūs ‘fast, vigorous; cozy, pleasant; cheerful, in good spirits’: Samogitian ‘heavy’, Latv. smags ‘heavy, burdensome’; Lith. viltis ‘to hope’: Latv. vilties ‘to deceive oneself, be mistaken’; and finally the cognate pair šilti: šalti, as well as Old Church Slavonic kon-bco ‘end’: na-čona, na-četi ‘begin’ (from *kin-), etc. That this could have been the case is suggested by Old Church Slavonic radъ ‘one who rejoices’: Old Russian roditi ‘to worry, make a great effort’. In addition, the emergence in the semantic range of džaubt(ies), gaūbt(ies) of the sememe ‘to rejoice, exult, delight’ alongside ‘to worry, take care, attend to’ may also have been aided by džiaugt(ies), gaūgt(ies): being quite similar in form, they may also have influenced the semantic range (see above on this phenomenon), especially since the semantic ranges of these two verbs had some points of contact; cf. gaūgties ‘to hope,’

³³ For semantic motivation, compare: ‘Lith. rūpi “concern takes hold” is the “stative” form corresponding to the active raūpti “to pick at”; in other words, the impersonal verb rūpi expresses the state produced in a person by someone’s continual “picking” (scraping) — raupimas’, — K. Būga, RR, II, p. 545.

16

‘to expect; to promise, be about to; to complain’ and gàubtiēs ‘to hope, expect; to promise, be about to; to complain’. Their semantic segments are intermediate between the sememes ‘to rejoice, exult, delight’ and ‘to worry, take care, attend to’.

SEMANTICS AND ETYMOLOGY

S. KARALIŪNAS

Summary

Possibilities of a structural approach in semantics and lexicon are discussed. The brilliant idea of F. de Saussure about two plans of language — the plan of expression and the plan of content—and their isomorphism makes it possible to apply structural methods in the field of semantics and lexicon. The grid presentation of lexical data proposed by F. G. Cassidy may be of great value in the investigation of Lithuanian lexicon. Beyond any doubt, structural study of a lexicon would in many ways further semantic research. It would, in all probability, provide descriptive semantics with stricter methods in studying such complex phenomena as polysemy and homonymy.

The principle established by B. Jēgers allows in many cases to discover polysemy in diachrony where it is generally considered to be homonymy from the sinchronic point of view. In this connexion it must be borne in mind however that there may be some cases of „pseudo-semantic“ development. E.g. because of their phonemic similarity Lith. plūostas „1. fibre, filament; 2. ferry, raft“ seems to have borrowed the first meaning from plūōstas „fibre; filament; tuft“.

In accordance with this principle, the etymologies of Lith. džiaugtis “be glad, rejoice” and Latv. džàubt “grieve, worry”, gàubt² “delight, rejoice, exult” are given. Lith. džiaugtis “be glad, rejoice” and its Latvian equivalents apžaugties “twine round another tree by growing”, džiaugt(iēs) “delight, please; take delight; be amused”, gàugtiēs “be glad, be joyful, rejoice, exult; hope, expect, await etc.” must be derivationally connected with Latv. žaugātiēs² “letting a herd run (into corn), leaping over ditches (when said of cows) etc.” and with Latv. žautiēs “climb, mount, ascend; romp, frolic; run etc.”, and Lith. džiáuti “run, hurry etc.” The semantic development from “run, hurry, leap etc.” to “be glad, rejoice, exult” is clearly seen in E exult “rejoice greatly; be jubilant or triumphant; (obsolete) leap upward, especially for joy”, or in OHG frao, frō, frawēr “joyful, jubilant”, ME frow “quick, rapid, fast” etc. It is worth pointing out that Lith. džiūti, Latv. žūt “dry, grow dry”, and Lith. džiáuti, Latv. žaut “hang up (for drying)” have proved to belong to the same root as Lith. džiáuti “run, hurry etc.” and Latv. žautiēs “climb; mount, ascend; romp, frolic; run etc.”, since the sememe “dry, grow dry” is to be understood as “flow out, escape, run out; drip out”.

Latv. džàubt “grieve, worry”, gàubt² “rejoice, exult” are most likely related to Latv. dzāibt “strangle, throttle; choke”, gàubt “butcher, slaughter; strangle”; for their semantic coherence, compare the development of E anxious and worry. Since dž- and ǵ- are dialectal reflexes of the consonant cluster *dj-, the Lithuanian counterpart of Latv. dzāibt, gàubt is džiáubti “go, walk in a state of sadness”. It is possible that

2. Leksikos tyrinėjimai

17

these verbs belong to the root attested in Lith. daũbti “hollow out; hit, kill”, daubà “ravine”, Latv. daũba “id.”, Lith. dùbti “become hollow, sunken”, Latv. dubt “sink in; cave in”. The semantic analogy is provided by Lith. rūpėti “be anxious, be concerned”, raũpti “id.”, raĩpti “dig, hollow out”, Serbo-Croat. rȕpa “pit”. In the shift to “rejoice, exult”, the sememe “grieve, worry, be anxious, be concerned” may have followed the model of the semantic universal “change the state of being X” or, more likely, may have been influenced by the words džiaugt(iės), gáugt(iės) because of their phonemic similarity.