Ksenija Leban Filozofska fakulteta, Ljubljana A Survey of Internationalisms between Slovene and English V članku avtorica obravnava internacionalizme med slovenščino in angleščino, pri čemer se na- slanja na korpus 895 leksikalnih parov, ki se začnejo na L, in pokaže, da lahko le v omejenem številu preučevanih parov prepoznamo prave prijatelje, medtem ko v preostalih primerih nastane oblika lažnega prijateljstva. The paper discusses the phenomenon of internationalisms between Slovene and English on the basis of a corpus comprising 895 lexical pairs beginning with the letter L, and shows that only a limited number of the analysed pairs can be regarded as true friends, while others result in false friendship of different types. 1. Introduction Internationalisms are "words which are used internationally" (lvir 1988: 93). The formerly prevalent opinion that these internationally used words represent the part of the lexicon that can be used by foreign-Ianguage users without reserve, for they not only share more or less the same form but also have the same meaning, has long be en replaced with translators' and interpreters' warnings against these translation traps for the unwary or false friends as most of them are of ten re- ferred to. As Ivir points out, and the present analysis proves, totaloverlap between in- ternationalisms that have entered two languages is only one of the possibilities. More over , only a limited number of internationalisms belong to true friends, and can be used by foreign-Ianguage users without falling into the trap of committing interlingual errors. Most of the internationalisms, however, result in false friend- ship, either on the semantic, morphological, phonological and/or orthographical level. True friends are, strictly speaking, only those pairs of words of common origin with which totaloverlap on all the above-mentioned levels can be established. How- ever, they also include pairs of words with which differences could be accounted for through language systems. Thus, for example, all the analysed Slovene lexemes end- ing in -acija, -ocija or -ucija correspond to English lexemes ending in -ation, -otion and -ution respectively (laktacija v. lactation, lokomocija v. locomotion and lokucija v. locution); all the analysed Slovene lexemes ending in -izacija correspond to English lexemes ending in -isation/-ization (legalizacija v. legalisation); and all the analysed Slovene verbs ending in -irati correspond to English verbs, formed through conversion from their respective nouns (lamentirati v. lament), while all the ana- lysed Slovene verbs ending in -izirati correspond to English verbs ending in -ise/-ize (literarizirati v. literalise/literalize). Similarly, some language-system spelling patterns may be observed such as, for example, Slovene -kt-, -gv- and -vk- corresponding to the English -ct-, -gu- and -uk- respectively (laktacija v. lactation, lingvističen v. lin- guistic, levkom v. leucoma). K. Leban, A Survey of Internationalisms between Slovene and English 43 Internationalisms not resulting in true friends form the false-friend group. False friends between Slovene and English! can be described as pairs of words of non-native origin which have preserved their foreign-Iooking and/or -sounding form in the Slovene language, and perhaps in the English language as well, and in which the misleading relationship between meaning and/or form may induce interlingual errors. False friends included in the sample were studied strictly from the synchronic point of view. This means that only the differences existing today were taken into account. Therefore, with pairs of different morphological structure, it did not matter whether a particular lexeme had been adopted into one of the languages concerned with its affix or not, or if the affix in question was productive in the relevant lan- guage or not. The analysis was concerned exclusively with present differences be- tween the lexemes. Thus lantan v. lanthanum, for example, belongs to the category of morphological false friends a1though the suffix -um was never productive in English and the English language borrowed the lexeme together with its original suffix. All the lexical pairs included in the corpus were analysed in detail to show the differences and similarities between the lexemes. Thus, on the basis of the definit- ions in Slovene and English monolingual dictionaries, possible differences in mean- ing between the representatives of lexical pairs were established first. All the pairs characterised by differences in meaning were then classified as semantic false friends. With the remaining pairs, the exception always being the so-called zero-equ- ivalent false friends, only differences in the morphological structure, pronunciation and orthography were looked for. Pairs characterised by totaloverlap both in mean- ing and in form were then classified as true friends, while pairs characterised by morphological, phonological and orthographical differences were classified as morph- ological, phonological and orthographical false friends respectively. When differences in form proved to be systematic, and could therefore be attributed to language sys- tems, lexical pairs were moved into the true-friend category. II. Internationalisms - A Corpus Analysis Internationalisms beginning with the letter I3 in Slovene and English were ex- amined with a view to establish how significant the phenomenon of false friendship really is. The Slovene corpus consists of internationalisms beginning with the letter L found in the only general monolingual dictionary of the Slovene language, the Slo- ! 1 believe the same elaboration could be applied to other language pairs, e.g., Croatian and English, perhaps even to any Slavonic language and English. Although my opinion is based on the similarities that 1 could notice when consulting literature on false friends between several Slavonic languages and English, it has, to my personal knowledge, not been supported by any comparative analysis. 2 Edward L. Thorndike, on the basis of a study of the lexicon, divided the English alphabet into 105 approximately equal units, called blocks. The letter L occupies four blocks or al most 4 per cent of the English vocabulary. Since a considerable number of lexical items of foreign origin begin with this letter, it is assumed that internationalisms beginning with the letter L might be considered a representative sample. 44 Slovenski jezik - Slovene Linguistic Studies 3 (2001) var slovenskega knjižnega jezika (1970-91), supplemented with the lexical items oc- curring in the Slovar tujk (1982) and the Leksikon Cankarjeve založbe (1994). The corpus, composed of 895 Slovene lexical items, was then completed with the English counterparts, provided the English language had any. Once the corpus was finished, each Slovene lexeme was compared to its English counterpart in order to establish whether any differences, regarding the form and/or meaning, may be observed. This was done by comparing the informa- tion about the respective lexemes in the above-mentioned dictionaries and encyc1o- paedia for the Slovene language and in the Collins English Dictionary (1994), The New Shorter Oxford Dictionary (1993) and the Oxford English Dictionary (1989)3 for the English language. Whenever the information occurring in the above-men- tioned lexicographica1 works proved insufficient or even questionable, other diction- aries were consulted, e.g., the Longman Dictionary of English Language and Cul- ture (1992), the Tehniški metalurški slovar (1995) and the Veliki moderni poslovni slovar (1997). On the basis of the dictionary information,4 it could be established whether the lexical pairs differed on the semantic, morphological, phonological and/or ortho- graphical leveis. When no divergenc es cou1d be found or when they proved to reflect systematic differences between the two languages, lexical pairs were c1assified as true friends. Often, lexical pairs differed on more than one level. A hierarchy of false friends was therefore developed. This part1y reflects the conc1usions of other ana1yses and definitions of false friends (Crystal 1987, Golobič 1988 and 1989, Granger and Swallow 1988, Hayward and Moulin 1984, Ivir 1968 and 1988, Limon 1983, Malone 1982, Partington 1993, Topalova 1997, Van Roey 1990 and WeIna 1977), and the re- su1ts of the comparison of the lexical items inc1uded in the corpus. While analysing the above-mentioned lexemes, it could be observed that, with certain types of false friends, differences may occur on various leveis. Semantic fa1se friends, for examp1e, may differ in meaning, morphology, pronunciation and spelling. With morphological fa1se friends, the meanings of the lexemes are the same - the lexical pairs wou1d differ in their morphological structure. Of ten , morphological differences would be 3 All the above-mentioned English dictionaries are general monolingual dictionaries. The de- cision to use the Collins English Dictionary as the basis for the comparison with the information on the Slovene lexical items provided by the Slovar slovenskega knjižnega jezika, which was dic- tated by the fact that the Collins English Dictionary is considered to be one of the best diction- aries of the collegiate-size category and is constantly being revised, proved to have some disad- vantages. Ideally, two unabridged dictionaries should serve as the basis of the analysis. Unfortun- ately, the only general monolingual dictionary of Slovene falls into the desk-size category. There- tore the amount of information included in the dictionary is by definition smaller than the amount found in a collegiate dictionary. It is possible that the results of the analysis and the re- spective conclusions would be different if two dictionaries of the same size had been consulted. 4 Relying on dictionary information has certain drawbacks. First, the information included in the dictionary is not always reliable or up-to-date. Second, it is concerned only with a limited number of lemma properties, e.g., the pronunciation, meaning, grammatical behaviour, etc., while other aspects, pragmatical, for example, are usually neglected. Nevertheless, dictionaries represent the most objective sources and are comparable among each other. That is why dictionaries were con s ulted in order to discover the differences between the lexemes. K. Leban, A Survey of Internationalisms between Slovene and English 45 accompanied by phonological and orthographical divergences as well. Finally, while with phonological false friends differences in pronunciation may be supported also by different spelling, divergences among orthographical false friends are restricted to orthography. Zero-equivalent false friends represent aseparate category, for, with this type of lexemes, false friendship is not caused by semantic, morphological, phonological or orthographical differences, but by the non-existence of asimilar counterpart. A possible hierarchy of internationalisms would then consist of true friends and zero-equivalent false friends representing the two opposite ends of the scale, with other types of false friends occurring in-between. Because they are treacherous, semantic false friends would appear closest to the zero-equivalent false-friend cate- gory. Semantic false friends would then be followed by morphological and phonologi- cal false friends, with orthographical false friends occurring closest to true friends. Figure 1. A hierarchy of internationalisms true friends orthographical false friends phonological false friends morphological false friends semantic false friends zero-equivalent false friends On the basis of the differences between the analysed lexemes and the hierarchy proposed, false friends were classified as semantic, morphological, phonological, orthographical and zero-equivalent false friends respectively. False friends with which differences occurred on more than one level were classified according to the most treacherous of the divergenc es established within the lexical pair. Thus, seman- tic false friends which also differ in morphological structure, pronunciation and/or spelling are treated as semantic false friends, morphological false friends with treacherous pronunclatlon and/or orthography belong to the morphological false-friend category, and phonological false friends with which differences extend to spelling as well are referred to as phonological false friends. Analysing the corpus and classifying true and false friends do not always pro- duce the same resu1ts. First of all, the classification very much depends on the accuracy and refinement intended by the analyst.5 Second, the unstable meaning 5 In her work, Magdalena Pregelj (1995) analyses Slovene internationalisms beginning with the letter A and compares them to the English ones. She claims to have found only 300 internat- ionalisms, of which 200 were true friends. Unfortunately, however, she does not mention anything about the compilation of the corpus. Therefore it is practically impossible to speculate on the dis- crepancies between her analysis and mine. At first 1 thought that her analysis only comprised the lexemes of foreign origin in the Slovar tujk (1982), but the number of the lexemes beginning with 46 Slovenski jezik - Slovene Linguistic Studies 3 (2001) with its fuzzy edges is very of ten hard to classify, especially if the individual classes are not as clear-cut as they should be. Third, the analysis of individual pairs is based on the information found in dictionaries. This poses several problems, the most serious ones being their reliability and comparability. As already mentioned, there exists only one general monolingual dictionary of Slovene, the compilation of which took the lexicographical team twenty years. This means that the dictionary does not register new words, new meanings and changes of meanings. Furthermore, the definitions and the illustrative material were taken from citation files and were not based on a corpus of Slovene as it is standard practice in the English -speaking countries nowadays. The analysis was also hindered by the size of the Slovar sloven- skega knjižnega jezika, comparable to English desk dictionaries, while the diction- aries used for the analysis of the English lexemes, in order to obtain as much in- formation as possible, belong to the categories of the so-called collegiate and una- bridged dictionaries. And fourth, language is a living form and each individual's contribution to it might result in an acknowledged change. III. Internationalisms Beginning with the Letter L - A Classification According to Ivir (1988), there are "three possible types of semantic relation- ship between an internationalism in one language and its counterpart in another language: full overlap, partial overlap, and no overlap" (lvir 1988: 96). When comparing the Slovene lexemes with their English counterparts in order to establish whether any differences, in form and/or meaning, may be observed, on- ly 84 lexical pairs, which equals 9.38 per cent of the corpus, have been classified as true friends.6 With the exception of liJtboy, lineation, localism, logograph and lotion, true friends seem to be restricted to subject fields, e.g., music,? religion (labarum v. labarum, lama v. lama, limbo v. limbo) , biochemistry and chemistry (lipid v. lipid, lipoid v. lipoid, lupulin v. lupulin), or are rarely used (lavacija v. lavation, letargija v. lethargy, lunacija v. lunation). See Figure 2. the letter A in the above-mentioned dictionary exceeds by far the number given by Pregelj. An- other objection to her analysis might be that her comparison and classification are based on the differences found when consulting the Slovar slovenskega knjižnega jezika for the Slovene lan- guage and Oxford Advanced Learner's Dictionary (1989) for the English language (ef. Pregelj 1995: 15). While the former is a monolingual dictionary for native speakers, the latter is intended for foreign-language learners. Since they cater for different dictionary users, the type and amount of information is different as well (ef. Hartmann 1983, 1992, Cowie 1987 and 1990, Landau 1989, Svensen 1993). The results of Pregelj's analysis would probably be different if she had consulted two monolingual dictionaries for native speakers. 6 The results can be supported al so by the findings of Alan Partington. In his 1993 article, Partington concludes that "there is tentative evidence to suggest, then, that the number of wholly reliable true friends between even closely related languages is probably fewer than is generally imagined" (Partington 1993: 109). ? Out of 84 true friends, 12 (14.3 %) are restricted to music (lamento, lamentoso, larghetto, largo, leggero, lentamente, lento). 8 The semantic analysis did not comprise zero-equivalent false friends for the simple reason that the English language lacks the counterpart with which the Slovene internationalisms could be compared. Since 115 zero-equivalent false friends were found in the corpus, the analysis was carried out on a sample of 780 Slovene internationalisms and their English counterparts. K. Leh~n .. A Survey of rnrer!l~li{)Il:lli"m" octween Sloveli!;: aml Engli'h 47 Figure 2. Slovene~English true friends beginning witb tbe letter L 6% Omu~le .medicine Dzoo!ogy Dbotany .culture !Jlaw IIngulstlcs Dre-lIglon • eurrency .art o blochem !slry El geogr!lphy .ma1hs .devlces .labrlc8 . measures ~An~lomy Dmater!al Dphllosophy Dphyslcs o psychoan81ys Is DIormal or IIter~ry r:lolher A ca Teful exam i nat ion of 1 he sem:lntic contenI of t he Slovene-Em: 1 is h 1exica 1 pairs~ s110wed that, on the concep!Ual level (and mayhe on other seman'lic levels as well), 45 pairs of lcx:cmes do no! overlap at all, whi le 134 pairs of inlcrnmionalisms overlap only paTlially. Out of thesc 134 cases of rarlial fa lse friendship, the Slovene lc:dcal itemis narrower in meaning Ihan ils Eogli~h counterpart (LI < L2) in 7X instances (laburističen \". 'ahollrILahour, lic(,nca v. licence and !iker v. fiqucur), there are 37 c L2) (lel'it v, Levire, limonada v. {emonm/e, !ull/innl v, /uminaf), whi!e 19 lexical pairs are at the same lime hroadcr and narrowcr in mean- ing (Ll <> L2), e.g .. lazaret v. lazaret lolla za retf{aZOrl'{{c , limu:ina v. limausine and lo:'a v. loge. See Fif!lITC 3. 48 Slovenski jezik - Slovene Lingui lic Studies 3 (2001) Figure 3. Slovene-English conceptual false friends beginning with the letter L total false friends 25% partlal false friends (L 1 <> L2) 11% partlal false friends (L 1 > l2) 21% partial false friends (L 1 < L2) 43% El! partial fals e fr ie nds (l1 < L2) • partial .alse friends (l1 > L2) O partial 'alse frie nds (L 1 <> l2) Olotal false friends of ten divergeoces on the cooceptual lewl are accompanied by differencea on the stylistic, collocational and/or connotational levd as well, e.g., lIlborlrati v. labcnIr. Jak v. Uu:quer aDd li/t v. liiI. Figure 4. SIovene-BqJiah .emandc falH trienda beJinniJll with the letter L connotational false friends 1% collocational false friends 6% stylislic false friends 16% • conceptual false friends collocatlonal false friends • styiistic false friends connotational 'alse friends conceptual 'alse friends n% K. Leban. A Survey of lnlernalionalisms bctween Slovene and Engli h 49 Apart from the above-mentioned semantie false friends, differences in tbe stylistic meaning resu/ted in false friendship in 37 additional cases, 14 lexical pairs had different ranges of coLlocations, while 2 more pairs differed on the connolalion- al level. In other words, 232 cases of semamic false friendship were idemified . The remaining 466 pairs of internationalisms were then compared in order to establish whether any misleading differcnces in form could be observed. Differences in Ihe morphoJogical structure result in false friendship in 236 additional cases. The corpus further consists of 184 examples of phonological false friend and 44 pairs of orthographical false friends. See Figure 5. zero-equivalent false friends 13% orthog raphical false friends 5% phonologlcal false friends 21% true friends 9% m orphological false friends 26% semantlc false friends 26% ~true friends .semantic false friends o morphologlcal false friends o phonologlcal false friends • orthographical false friends EJ zero-equivalent false friends Similarly lo true friends, morphological, pbonoJogieal and orlhographicaJ false friends do not present any problem semantical1y. The errors are due to a cenain similarity in form, which leads us to believe that total correspondence may exist. Most trans]ators, lexicologisls and lexicographers tend lo ignore ralse friendship ca used by dissimilarity in form . Yet, the fact thal 51.7% of internationalisms begin- ning with the letter L eilher have a different morphoJogical structure or are pro- nounced andlor spelled differcntly shows that lbis group should not be neglected. The analysis of morphological mlie frieDcls 1howedthat. oli the bitsis 'of the differences in form, most of the internationalilml may be dividei4 intothrec groops: intemationalisms with cUfferent suffixes (137 pait.~ which amOUntB to 58% of m.otphoIogical falae friends), intemationaliim.1 provided. with a suffix in slovene and corresponding to Jexemes witbout any suffix in English (63 pain, which amounts to 26.7% of morphoIogical falae friends) and intematiana1i1lD1 provided with a suffix in English and corresponding to lenmeJ without any soffix in slovene (23 pairs, which amounta to 9.7S% of morpho1ogical falae friends), see Figure 6. 50 Slovc:n,ki jezi" - Slovenc Linguislic Studie_ 3 (2001) Figure 6. Slovene-English morphological fa1se friends beginning with the letter L .Qooup 1 (dlfferent .um-.) • Group 2 (sufftx ln L 1 Y. no eufllx in L2) C Gl'0UfI3 (no suttlx ln L 1 v. sufftx in L2) O ~p 4 (no preflx ln L 1 Y. pratla: ln L2) • Group 5 (mlac:ellMeoua) Although there is a very restri:ted, yet existent and of ten quoted. group of treacberous ad't'el'bl. e.g .• t\lf!1JlUaino v. ~ntUIJlly, evidentno v. evtderrrly. etc., morpholo&ital false frieDdB in the corpus of Slo'ICnc~ English intemationalisms be- ginning with the letter L do not iJxtude adverb5. Wbi.Ie the noun is tbe prevailing pan of speecb in Group 1 and the only part of speecb in Group 3, most of the eXamPIeS in Group 2 are adjectives (71.4%). Very often, the differences between slovene and &gUsh can be explained as languages' preference for different word-formation processes (cf, Klinar 1996: 149-233). Cenain word.formatioo PatterD$ seem to be predictab1e. when analysing inter- nationalisms with different auffixes, for examp.e, il cruld. be establisbed that all the slovene w.rbs eOdiDg in -izirati com:spouded to English ...erbs ending in -isel-ir.e. e.g., rtJiciVratiJlai$aft Y. la1dstIIaJCiz.e. iegalir.irofi Y. legalj~!legaJW! and lekJJka1Wrati se v. Iai~. WbiIe time Cllding in -irali teIKIed to com:spood to English. verbs c:onverted from the correspcmding nouns, or to those furnisbed with the zero morpbeme, e.g., lmnDIlimli "lJJmenJ.licencirtzd v.liCM.3e and lokavtirat; v.lock out. The same degree « predjctability was established with Slovenc nouns ending in -acija, .ociJa and -udja whicll all correaponded to English nouns cnding in -tJtiOll, ~itm and -«tion (jakUu:ija v. lactation, loIwmocija v. locomotion and lole.elja v. lo- cutio1l). with Sl