Janez Orešnik Ljubljana CDU 831.1-022 THE ORIGIN OF THE CLITICNESS OF THE WEST GER.."1.iANIC DEFINITE ARTICLE: THE CASE OF Beowulf 801. 21 The present paper is mainly concerned with the definite ar- ticle of Beowulf. The literature about the West Germanie defini- te article has always a-llotted a prime place among the old Ger- manic texts to this poem: it is undisputably archaic (Krihn 1933: 105, D. Hofmann 1959:189), and suitable for statistical geriera- lisations. Its more than 6 ,OOO hemistichs contain about 300 ·in- stances of the. definite article (Lichtenheld 1872:332). 1 Let me remind you of the long-known distinction between what Kuhn (1933 :4-5) called satzpartikel ahd Satzteilpartik-el. Both terms refer to unaccented or weakly-accented-words: in broad outli·ne, a satzpart-ikel is a syntactic constituent at the clause level (e. g. a pronominal subject), a satzt::ei-lpartikel is a con- stituent within a syntactic constituent at the clause''ievel or below it (e. g. a preposition within a prepositional phrase). Concerning the satzpartikeln of Germanic, Kuhn (1933:8) fornnl:- lated the following Satzpartikelgesetz: "Die satzpartikeln stehen in der'ersten senkung des satzes, in der proklise entweder zu seinem ersten oder zweiten betonten worte." The definite a-rticle is of course a satzteilpartikel. However, as noticed by Kuhn in a cautious footnote (1933:45 fn. 1}, it frequently ("gern") accompanies satzpartikeln in Beowulf. To Kuhn' s observati·on, I add the following specification (the spelling of the illustrations, the punctuation, and the morpho- logical analyses of the words are taken from ed. Klaeber 1950): 1.1 Almost always when the definite article accompanies satzpartikeln in Beowulf1 the definite article immediately fol- lows the satzpartikeln, and the sequence satz·partikeln + defini- te article stanas-clause-initially. Example. 66b (georne hyrdon,) o.t.Q. p-ae.t seo geogoe geweox (the Satzpartikeln oel.El pat and the inunediately -following definite 383 article seo before the first lift). 1.2 In 24 instances, the definite article is found imme- diately after a clause-initial sequence of words consisting of a finite lexical ("full") verb .(accompanied by its own satzteil- partikeln, if any) and of Satzpartikeln. Example. 1O9a ne gefeah he t>aere fi:hf1e, ( ac . he hine feor for- wr~c) (the satzpartikel he and. the definite article .P9f~ after the fin~te verb). 1.1-2 will here be referred to as the.satzpartikel-usage. of the definite.article. According to my count,.the.satzpartikel-:usage bolds for about 70 % of th_e instances of the .definite article in Beowu.lf, and is 1the most widespread usage of the definite article in the poem. Tt must also be .the oldest usage of the definite article in Beowulf, for all .other.usages of the definite article (cove- ring about 30 % of all instances) can be explained as the rela- xation of the constraints .(cme or two .at a tirne) that govern the Satzpartikel-usage. The constraints are, 2.1 The definite article immediately.follows the satz.par- tikeln. (This is the .. basic constraint:, from which the remaining ones follow as consequences.) 2.2 The definite article is not used with non-first accen~ ted noun phrases of the clause (becaus·e no noun phrase can pre- c.ede the satzpartikeln in the clause, nor can anything interve~ ne between the Satzpartikeln and the definite article). 2 1 ~ 3 There is only one defini te article per clause (because there is only one Satzpartikel(cluster) per clause). 2.4 The definite article precedes the·first·lift of the he- mistic;::h (because the sa.tzpartikeln prec:ede the first lift of the hemistich) • In studying the relaxations from the above constraints, it is useful to treat separately · th.e .definite articles occurring in prepositional phrases (3.1),:and the remaining definite ar- ticles ( 3. 2) ; a few more exceptional .cases ha ve been relega ted to (3.3). Several instances are listed sub more than one item. 3.1 The de:f;inite article in prepositional phrases. The re- laxation of constraint (2.1): 384 3.1„1 Only the preposition intervenes between the Satz- partikeln and the definite article. 19 instances. E. g. 192b (lap ond longsum), pe on .Qa leode becom + 270a 1215b 1484a 1485b 1612a 1629a 1665a 1694~ 1780a 1868a 2028a 2406a 2468a 2528a 266la 2690a 2788a 2860a. 3.1.2 The prepositional phrase begins a hemistich which is not clause-initial. 43'instances, mostly also breaking con- straint (2.2), because the first accented noun phrase stands in a preceding hemistich. E. g. 125a mid pare W2.lfylle (wica neosan) + 326b 425a 443a 617b 647a 695a 713b 824a 838a 919b 1016b 1073b 1082b 1191b 1199a 1280a 142la 1614a 1635a 1638a 1639a 1956b 198la 1984b 2010a 2039b 2083a 2139a 2197a 2232a 2298a 2366a 2374a 2405b 2465a 2520a 2560a 2786a 2835a 2856.a 2905a 3109a. 3.1.3 The prepositional phrase is clause-initial, but the clause does not begin with Satzpartikeln. 5 instances, e. g. 1052a on t>2re medubence (mapdum gesealde) + 1030a lllOa 2669a + (involving a hemistich-initial finite full verb) 2690a~ 3.2 Definite articles~in noun phrases outside prepositio- nal phrases. The relaxa tion of constraint .( 2. 1) : 3.2.1 The noun phrase begins a hemistich which is not clause-initial. 14 instances, mostly also breaking constraint (2.2) (because the first accented·noun phrase of the clause stands in a preceding hemistich). E. g. 9b (oe p~t himzghwylc) para ymbsittendra+ 416a 707a 792a 1057a 1105a 1309a 1406a 1614b 1685a 2382a 2823a 2887a 3122b. 3.2.2 The noun phrase is clause-initial, but the clause does not begin with satzpartikeln. 15 instances, e. g. 107b (in Caines cynne -) pone cwealm gewr~c + 363a 639a 80lb 1202a + (involving a hemistich-initial finite full verb) 3lla 712a 762a 1563a 1570a 1982a 2239b 2329a 2977a 299la. 3.3 The remaining cases. 9 clauses contain two definite articles each, thus breaking constraint (2.3), e. g. 639a Bam wffe pa word (wel licodon) + 109-10 646-7 712-3 1190-1 1614 1638~9 2587-8 2999. Constraint (2.4) is broken by 5 instances, e. g. 301.la meltan mid pam modigan, (ac pi:r is maama hord).+ llOa + (the following instances all involving sele parn hean) 385 713b 919b !984b; constraint (2.4) is also broken by 7 instan- ces containing the postposed definite article, e. g. 2334b (ealond utan;) eoraweard ao.ne + 2007b 2588a 2734b (?) 29~9b 308lb~ - The cases defined sub (1.2), involving a hemistich- initial finite full verb + satzpartikeln + definite article, are an ancient variant of constraint .( 2. 4) , and ar.e here not considered a violation thereof on a par with the relaxations enumerated in the .present. paragraph·. I now return to the Satzpartikel-usage of the definite ar- ticle in Beowulf, The satzp~±tikel-usage of the definite article has nothing inevitable about it~ seeing that the usage of the definite ar~ ticle in Beowulf is. optional anyhow (,Barnouw 190:2), witness, for instance, the rarity of the otherwise expected definite ar- ticles in A-, D-, and E-hemistichs .. ,(The satzpartikeln were abo- ve all used in the "light" B- andC-hemistichs, cf. Hinderschiedt 1979; the same holds true of the Satipartikel-usage Qf the defi- nite article.) Therefore, the abundant satzpartikel-usage requ- ires. an explanation. However, the latter is not the object of the present paper, and I confine myself to the following remarks to illustrate what use of the definite·. article was so similar to the use of the satzpartikeln as to justify the article's place next to them. Hodler(l954:29-32) P.osited, among th.eearliest uses of the definite article, what he called der satzverkniipfen- de Artikel (i.·e. article whose special function it is to join clauses together), noticing that it can stand only before the subject.of .discourse., normally before the subject noun phrase. (My addition: as many as 50 % of all definite articles in Beowulf introduce subject noun phrases.) The most recent treatment of the Germanic definitearticle, Moskalskaja·l977, considers (253- 4) the early definite article to be very probably a syntactic particle accompanying nouns, true enough, but playing the role of a "suprasyntactic" means in the logical and communicative structure of the text, most often signalling the theme of the claus.e. (Hadler. and Moskalskaja did .. not study the position of the definite article. with respect to satzpartikeln, but arrived at their respective · views investigating the semantics · of the d.e- 386 finite article, especially in Gothic.) I now cometo the mostly neglected problem as to how the definite article (whose source, a demonstrative pronoun, re- ceived the normal word accent) became clitical. I believe that it was the satzpartikel-usage that triggered-what the definite article, being a function word, presumably aspired to anyhow1 namely almost permanent unaccentedness. Given the already mentioned archaicness of Beowulf, it can be assumed that the usage of the definite article in the poem reflects the earliest usage of the Old English definite article that can be observed (Lichtenheld 1872:332-3). Hence my postu- late that the satzpartikel-usage is the oldest usage of the de- finite article in Beowulf implies that the satzpartikel-usage is the very earliest accessible usage of the Old English defi- ni te article. For a tirne it was probably even the only usage of the definite article in Old English. In those days, the defini- te article ALWAYS stood next to the satzpartikeln, and in the permanent vicinity of those light elements, I believe, lost its word accent through assimilation to the lack/weakness of word accent in the accompanying satzpartikeln (obligatory accent sandhi). Later, the unaccented definite article freed itself gradually from the obligatory presence of the satzpartikeln, most probably because the article's character of satzteilparti- kel asserted itself more and more. (The article's unaccentedness lost its sentence-phonetic motivation.) (Judging by the statistics sub (3) above, the relaxation of constraint (2.1) was quicker with definite articles in preposi- tional phrases than with other definite aiticles. This may be due to the fact that, in prepositional phrases, the definite article was protected by the preposition, i. e. by a clitical word, also after the relaxation of constraint -c2.1), so that the relaxation was felt to be less of a departure from the ori- ginal state of affairs with prepositional phrases than with ot- her constituents containing the definite article.) Beowulf seems to be the only preserved Old English text in which the ancient Satzpartikel-usage is so prominent. (Compare Barnouw 1902 for an analysis of the definite article in a num- 387 ber of other Old English poems.) In Old High German, the Hilde- brandslied, although like Beowulf in the respect under discus- sion, is a much less convincing case, containing as it does no more than five instances of the definite article. At any·rate, I believe that the definite article of the West Germanic langu- ages other than Old English became clitical in the same way. On the other hand, I do not wish to preclude the possibility that a part of the definite articles, of English or of any other West Germanic language, became clitical in some alternative way, say because they were of a different origin than the bulk of the de- f ini te articles. If I am right in assuming that the development was as des- cribed above, function·words can become clitics in obligatory contact with unaccented or weakly accented words. Other words .. than the definite article may also have perman- ently lost their word accent if and when the syntax brought them into contact with the satzpartikeln, for instance the pronominal subjects (Orešnik 1984). These interesting cases have been left out of account here. REFERENCES BARNOUW, ADRIAAN JACOB. Textkritische untersuchungen nach dem gebrauch des bestimmten artikels und des schwachen ad- j ekti vs in der altenglischen poesie. Leiden: Brill. 1902. GUXMAN, M. M., ed. Istorikotipologičeskaja morfologija germanskix jazykov. Fonomorfologija, paradigmatika, kategorija i- meni. Moskva: Nauka. 1977. HINDERSCHIEDT, INGEBORG. Zur Heliandmetrik. German Language & Literature Monographs, vol. 8. Amstei~am: John Benja- mins. 1979. HODLER, WERNER, Grundzilge einer .germanischen Artikellehre. Hei- delberg: Winter. 1954. HOFMANN, DIETRICH. Die altsachsische Bibelepik ein Ableger der angelsachsischen geistlichen Epik? ZDA 89, 1958/59, 173-90. KLAEBER, FR., ed. Beowulf and the Fight at Finnsbur . g. Third edi- tion with.first and second supplements. Boston: Heath. 1950. KUHN, HANS. Zur wortstellung und -betonung im altgermanischen. PBB 57, 1933, 1-109. 388 LICHTENHELD, A. Das schwache adjektiv im ags. ZDA 16, 1872, 324-93. MOSKALSKAJA, O. I. Stanovlenie kategorii opredelennosti/neopre- delennosti. Artikl. In: Guxman ed. 1977, 238-86. OREŠNIK, JANEZ. Clitics are linguistic signs of excellent quali- ty, or tne origin of obligatory pronominal subjects in Germanic languages. Wiener Linguistische Gazette. Sup- plement Beiheft 3 (= Discussion Papers for the Fifth International Phonology Meeting), 1984, 185-8. My thanks are due to Miss Margaret G. Davis for the correc- tion of my English. - The printing process used for the pre- sent number of the journal has made it necessary to disre- gard the usual rules for dividing words at the end of a li- ne. Povzetek ODKOD NASLONSKOST ZAHODNOGERMANSKEGA DOLOČNEGA ČLENA - PRIMER Beowulf 70 % vseh določnih členov v staroangleški pesnitvi Beowulf stoji tik ob "stavčnih" breznaglasnicah, čeprav določni člen ne sodi mednje. Ta stava določnega člena je najstarejša sploh ugo- tovljiva stava določnega člena v Beowulfu in s tem v stari an- gleščini (kajti vse druge stave določnega člena v Beowulfu so razložljive kot spreprostitev najstarejše). Pojasnjuje, kako je staroangleški (in verjetno širše zahodnogermanski) določni člen postal naslonka: v obveznem stiku s "stavčnimi" breznaglasnica- mi. 389