Language is the main means whereby people Pecan trea erty Ree esa cree cater tres Seemann Perna o saya etoce yas need to examine the problem from both inter- national and intranational points of view. In the former case, we are dealing with the difficulties Pegg aee ca cua Pee eC RCO cracls proposed to alleviate them. In the latter case, we need to consider the consequences for mutual understanding that stem from the existence of so Pe ee ee aon Tae au A Geest ogee tao) PSone rear ec een eet Rute oan have been proposed to reduce or eliminate the Pee ner ne CROC Sun Un and interpreting ~ a field whose future will Pete eee ge eer Pee ae atte La auniliary language (such as Esperanto) or to simplify an existing language for international use (such as Basic English). A third solution is to Roce eer et tae tor a ed SOO rat teenie Tae seems to be happening to English. A fourth is to ETAT tsa fuotonnaoye lel foster the growth of multilingualism in individuals and societies, either through the natural course of EO ue a One eta Diora the promotion of special educational procedures for teaching languages. Each of these approaches is given separate discussion and illustration. A focus is provided by the branch of sociolinguistcs known as ‘language planning’ ~an area that has attracted increasing attention in recent years. ‘Within a language, the co-existence of many specialized varieties presents a further area of enquiry. Fields such as science, medicine, Jaws religion, and mass communications have developed styles of language that require careful study if they areto be understood. Popular Pte rete Cr ene aoc erent ern aL na een eer oC ota Sererrem teectur eich of linguistic features they display. Partx then concludes by considering theimpact of the “information er ol cas errr tol linguistic data, only a fraction of which can its ere rece aaa prc feta for increasing difficulties of organ Peon56- THE LANGUAGE BARRIER tage can be a barrier to com- de by al who travel, study, ie activity is. tourism, business, or data dis- The discovery tha ang munication is quickly ma govern, or sell, Whether the esearch, government, policing, b semination, the lack of 2 common language can Severely impede progress and can halt it alrogerher “Common language’ here usually means ae lan jage; but the same point applies in principle to any eee ee itu salere orgie witine single language. ‘They don't talk the same language” has a major metaphorical meaning alongside its literal one ($63), ‘Although communication problems of this kind ‘must happen thousands of times each day, very few become public knowledge. Publicity comes only when a failure to communicate has major consequences, such as strikes, lost orders, legal problems, o fatal acci- dents—even, at times, war. reported instance of com- munication failure took place in 1970, when several ‘Americans ate a species of poisonous mushroom. No remedy was known, and two of the people died within days. A radio report of the case was heard by a chemist who knew of a treatment thar had been successfully used in 1959 and published in 1963. Why had the ‘American doctors not heard oft, seven years later? Pre- sumably because the report of the treatment had been published only in journals written in European lan- guages ~ other than English. (After D. A. E. Shephard, 1973.) Several comparable cases have been reported. But isolated examples do not give an impression of the size of the problem — something that can come only from CRITICAL LANGUAGES Weis chiUs Abeion Edna Kio owa Sec ee meee res, 9) 1a wa ee eee ee ees uot eae lshedalstof 169. Armerian Fifan Korean Nyanja Tamil Eeeccmete a, eae one nee theUS Govern. Aymara French Keo Oromo Telugu pt ene ‘Azerbaijani Fulani Kumauni Panjabi Te a tobe critical’, in Bahasa. Ga Kurdi "apiament habs thesersetnt” indonesia Ganda (oh manor? puaaas knowledge of Balinese Gbaya Lamani Persiat Tere them wouldpro- Baluchi Georgian Latvian Polish ae mote important Bamileke German Lithuanian Poly a Semcicresarch Bashir Greek luba Poruguese Foe ‘orsecurityinter- Basa (Kru) Guarani lacedonian. “ dyad Sisotonatons! feiguson Surah Medwere’” Rope Ta oreconomickind. Bemba Haitian Creole oer wee Summit Soe fotincreoe Mesh fomnin jengali fausa Male ia irae fool ce oe a ees | Berber Hebrew: ‘Malayalam Rundi a eile! Se, Re Menta” hy oh a oe Mardekan Rwanda Urdu ington Postcalled Burmese ‘ban (Malay) arath ait vest patties Su salon Mar sn Nee guepeileracy, rm yeshots Maya Serbo-Croatian Visayan aaselteney. Cano ‘ no fe wae Segoe Sn io Minaglabes 208 Youur ei ee wine Sirhaie es cetemtothon cha aa hep St Yin es Ske Jopnee rin sees eae Danish Kamba Mote Somali Yucatec sovimn Ban ‘en Merdoio SoMa, oe sine Oa Kirmese tal a ee Miiaae Bath ae ee Spann _ ‘kan mae ach Melanesian Sunderese az Nepali ‘Swahili ev sage Thee build zt in in Bruel ‘does the fore md barrier exstsome te the offices ofthe gant Union nS gee retin araconese ViceinBrassel whee Many agencies and nn ai , 000 iter = cortehMmihenee tt ‘with 400 fultime sa about beast he become thelerge ing sevceinthewol 1985 theta inaeae with merbersate arc ‘official anguagesinthee work andperistn, lation pobleminetegs theoretical fang, ais inthe Stangigtig Of the ear 1980 eu tions were aed aaa, ver 3 millon wrtady IRbimpotibe fhe: interpreter forall ais orto prowde nan Coverageon allocaon interpretation no roe all 11 languages weretobe provided, it would requrest least 33 interpreters|iingte esestemo! ten ‘reters per booth whens thansinlonguapesseel ‘ata meeting). Pragati tions thereforehavetove found, Greist eats system terse Portugues interpret ‘example, English mghtbe ssanintermediay #205 ‘Also, everyetfon émase? find interpreterswho et owe tee cr four ina But thereisawaysston of trained interpret Morera eoposed reauctonotthe ni ‘working languagest® Sietasin he UN) tt suggestions are high te ‘versial. No County MO comorablesbon a2 islinguistics i Tmprblemreret tobe confronted: 3 berhipgrons ey cal maximum fora languages he ga gle relay anova pet ‘become increasing, sive, See furthes 5or avoidance of foreign-language af be er different communicative st obit and conta peaking, scientific wold, for he Fok and document 7 pformacion agencies have shown Fi a foreign-language 1 rial is ever con. ltd (One nalysis of over 60,000 British ‘pe field of science and technology io were for foreign-language peri- 1976). Studies of the sources consulted in ead pone only 3 A, Bow (citation analysis) leads toa sim- 2 of foreign-language sources is eas 10%. Likewise, in the non- s lire! orld, there is alsoa marked reliance jaro jroo we found io Fpgibpet material ~ though here the special nalish must be considered (sometimes of of ne er half the requests made or sources nln 4g fo 4 several ways of getting around the Foreign anguage bare erent success «but none is simple, nor has any as yer cease the mumberand availablity of translating ind interpreting services (957). +, Delop an auxiliary language that everyone will understand ($58). 5, Develop an existing language as a world language thateveryone will understand ($59). 4 Posie increased motivation and opporcunity to Jean foreign languages ($$60, 62) THE BUSINESS WORLD The language barrier presents itself in stark form to fumswho wish to market their products in other coun- rts, British and American industry in particular has insect decades often been criticized for its linguistic inlrty for its assumption that foreign buyers will behappy to communicate in English, and that aware- sof other languages is therefore nota priority. In the 194, over two-thirds of British firms dealing, with ‘on Engish-speaking customers were using English ‘routgoing correspondence; many had their sales lic Siuronly in English; and as many as 40% employed ‘Sere quaifed to communicate in the customers "is similar problem was identified in other ‘grou ue And ona ee Sgpunuiss were by no means exempt — though the is vie i of English as an ciaeeaee language He ey oPen to the charge of insularity. Sveti sean ind Publicity given to ¢his problem inn ng ee to have ready improved the st ing yp! taining schemes have promoted an Preuistic and cultural awareness, ~ for cy ep eames of industrial visits abroad, em "of once * Of unior managers, oF the appoint: Senate atts. Many frms developed theie ata eae ie international velephone ser ‘one example ftom Britain, Rowntree cumple, ey a Mackintosh began to publish their do languages (English, French, Conus Baa and Xtosa}, Some firma ru partite foes eg language with which theyare mostinvloedsoonepen duce their own glossaries of technical terms, to enaine consinency when mai being arsed csaow much more readily appreciated that marketing efforts take account of the linguistic needs of the customer, or to look after one’s own linguistic interests abroad (in such areas as patentingand trade-mars contol) The changes in awareness have beer: most marked in English-speaking countries, where the realization has gradually davened that by no means everyone in the world knows English well enough © negotiate in it (659), and that this is especially a problem when’ English is noc an offical language of public adminis- tration, as in most parts of the Far East, Russia, Eastern Europe, the Arabic world, Latin America, and French- speaking Aftica(p. 357), Even in cass where Forign cpstomers can speak English quite wall itis often for- gotten thac they may notbeable to understand ctothe guited level ~bearingin mind thergionaland socal variation which permeates speech (Part 1) and which van ease major problems ofistening comprehension Tn securing understanding, how ‘we peacoat juscas importante appeats a8 how they sPeakto Adasot orignstadens ‘etl Language ston Cambridge Teeate oe So0such schoo a Btn ovshle anata these ae oe rea ‘izedby profeoralbod inthe el 980, hey wee Sealing ove! 00 Scene ech you og erating an aualtumoner ‘Stover 300 lion Many ‘chook tvoughouhe wer now nto couse geared othe special neds ofp groups of oe fessor ante, Inver or nt57. TRANSLATING AND INTERPRETING with a foreign-language barrier, tral term used for all tasks where the meaning of cxpresions in one language (che ‘source’ language) is ried into the meaning of another (the ‘target’ lan- guage) whether che medium is spoken, writen, or Signed. In specific professional contexts, however, a distinction is drawn beewecn people who work with the spoken or signed language interpreter), and those who work with the written language (translator). There are certain tasks thar blur this distinction, as when source speech is turned into rarger writing (for example, in monitoring foreign-language broadcasts, or in writing sub-ttes for forega films), But usually thetwo rolesare seen as quite distinct, and iis unusual to find one person who is equally happy with both ‘occupations. Some writers on translation, indeed, con- sider the interpreting ask to be more suitable for extro- vert personalities, and the translating task for introverts! TRANSLATING Icis sometimes said thac there is no task more complex than translation ~a claim that can be readily believed when all the variables involved are taken into account. Translators not only need to know their source lan. guage well; they must also have a thorough under- standing of the feld of knowledge covered by the source text, and of any social, cultural, or emotional connotations that need tobe specified inthe targetlan- guage ifthe intended effect is to be conveyed. The same special awareness needs to be present for the tar. ge language, so that points of special phrasing, com. temporary fashions or taboos in expresion, local ( regional) expectations, and so on, can all be taken ints account. On the whole, translators work into their mother tongue (or language of habitual us), to ensene schat sounds a6 natural as posible ~ though THE PROBLEM OF TRANSLAT] The aim of translation is to oN rovide semantic equivac lence between source, e pr cue and tng language Thisis what n from other kinds of lin. pone tes such a8 adapting, précis writing, ad stracting, However, there are many problen® fit den within this apparencly simple statement, all to do with what standards of ‘equivalence’ should be expected and accepted. Exact equivalence is of course impossible: no trans- lator could provide translation that was a perfect par- allel to the source text, in such respects as chythm, sound symbolism (§30), puns, and cultural allusions, Such a parallel is not even possible when paraphrasing within a single language: chere is always some loss of information. On the other hand, there are many kinds of inexact equivalence, any of which can be successful ata certain level of practical functioning. Ic therefore follows that there is no such thing asa ‘best’ translation. The suc- cess ofa translation depends on the purpose for which ic was made, which in turn reflects the needs of the people for whom it was made. An inelegant, rough- and-ready translation of a letter can suffice to inform a firm of the nature of an enquiry. A translation of asci- entific article requires careful attention to meaning, but lied attention to aesthetic form. The provision of a dubbed film script will warrant scrupulous care over the synchronization of lip movements, often at the expense of content (p, 396). Literary work requires a sensitive consideration of form as well as content, and may prompt several translations, each of which emphases a different aspect of the original. Its easy LEVELS of TRANSLATIoy + iar (or oan al eine orpheme igre 208 Tere hen iomeng onsare neg Wsraining cats ang Heslewiart che * Literal rasan ne, ‘uisticstructueof-ke Source tertsfoloued te, ormalizedaccortngte, "les ofthe target ena I's raining catsandos Ui pleut des chatsetde chiens, + Free transaton They Quisticstructureofte source language signa, and an equivalents fous based on themeaningir veys: Is raining cats and dos pleut averse. ine practicer __17 TRANSLATING AND INTERPREpIE ight be ‘best’ for one set of circum- pay their tran Pe par what Ey unsuitable for another. Japanese) regard fey nord. Othe ost Mi ybeentl vee Alston fie {RPaNCS) regard ic asa major ince ut athe ost kinds of ana have ison right. The question ofstarus ha sling in seca fri ange of pesibiliis. In Hane specially in Europe, where des ney uch po tf he emphasis emir on accuracy tr isrockting especialy ag a tae face ae bjt at eau for intuc: Since he theca MEU. pool 4 much ciemifcrescarch (921). In esional ransatorhas come nea he pr od maak on mporantfrlirary materi nied Some ald ls phoney 2 be eg sna erg the emotional as Well asthe orn goveramentor comme cmsatonagens fis Tot ehe work, and on mainaining they provide an in-house eens neaon te eas alin equivalence ($12). Fibnegraphic large cadre of freelance uanslawoe ae We 84 el talons aim to pay fll attention to from home. The fld now her tly Noting cele grunis ofthe auhorsand te ecpi- examinations caer sructue see se seal ole im acount diferences baween nisaions —suchasthe Amercan Tiegh alee cs and fet angage, as when Cheistian religious con, the Translator’ Guildof ihe (Ba ces sourtan OIE ne Middle Eascare ‘tanslated’ into Linguists, and the Rédérarion Ieee osttteoF r based int eis ras of Cental ica or moderncd th e910). And there are various kinds of li fnc on, where the aim i to convey the struc- ear ofthe orginal text, often in a. quite literal ane emphasiing. such features as archaisms, aera and levels of formality (§$8, 10). Most sao oF cone, are mixtures ofthese theoretical syeveltng the complex reality of language in use, here‘pure varieties are conspicuously absent. PROFESSIONALISM. fansnors aim to produce a text that is as faithful to ginal as circumstances require or permit, and yet that rads as if it were written originally in the target language. They aim to be ‘invisible people’ — rransfer- fngcontent without drawing attention to the co! cable aristic and technical skills involved in the yoces The complexity ofthe task is apparent, bucits inporanceis offen underestimated, and its practition- © seca status and legal rights undervalued. Some ‘cunts view translation asa menial, clerical cask, and HOW FAST CAN TRANSLATORS WORK? Then sata on arate anaes it dpends- upon such ste raat experience, their familiarity temesibict matter, and whether they dictate the cnoshaveotype t themselves, the difculy of 2c xia factor a tuansator may beable to 1) ta tate Pe hour or popular writing, but beteenag teal material. The linguistic relationship saga auc targetanguage sas relevent lay teens ehtahae structural and cultural patterns Wil (late than those which are widely ctferent. ‘retangectenar aston canals afectspeed Sage ares ate amost twice a fast in working from i they are from Bi serting continues, sma ng nd eee ith fami subject mat imof producing a polished” .erimator 9 "poi She zoom anger tansator outputvary trom {te abo 2000 words a day, with most people (rg 0°-Esiatesof the gain in speed inSStorhumen ea letion (P- 352) are not yet clear: gol the machine-trarsated “en presen "%°4€Ount, and this can sll be con- ee taken i an incervationale des Tra- ducteurs. The European Union of Asses Translation Companies was formed in 1994, ‘The number of translations made i certainly on the increase, fuelled bya growing number ofspecialized mul tilingual publications (such as journals that publish edi- tions in more than one language or provide issues specially devoted to translations of foreign scientific material). Asa consequence, several cntral organizations: have now developed to coordinate information about the availability of wanslationsand wo facilitate theicaccesbil- iy, once they are made notably the International Tans- lations Centre at Delft (The Netherlands) and the National Translations Center in Chicago. In chisdomain, ‘of course, the advent of computational techniques of informacion storage and retrieval has been a blessing = although one that is not yet as widely share as ic might be, In a 1971 study; over 90% of a sample of academic staff had never used any translation indexes; and other reports indicate that perhapsas many as 80% of scientists ‘are not even aware of their existence (J. A. Large, 1983) Great claims have been made for translation, Tt has been called the key to intemational understanding, The Japanese see it as a key co learning, Western Europe, ithas been said, ‘owes its civilization to ans: == ° THE earn OPERATION The tears fuletanetererspett osha he orelgdanguage bani operation (38) Some cs sexing uy oes tan Browkea domain [ae ace caren chica saree ete enie eee cere severe omen “oven igtlraae ete, eee oration effect was to suggest that. oe inner iene + Intone languages (529), it ‘is almost impossible to adapt cee ‘meaning Inonetatin Amer Grrtonlangusge sac sequence, the missionaries: found that a sungtranslation — about ‘sinners’ was in fact ators’ (L. G. Kelly, 1979). Ie is all probably so. grave! pte mina igugeed aT — aguante aoe lik Geman Fa sera VCHOLOCY geen tae DIS 5207 _ 18 etakenfromancarh sre noish 82552 go" Fennec Seman 277 ig 30 re Bae English authors! Hest ae emma waa _* ‘ores peta! falth Tet rset PaYSKS 9 86.2 59 ah ESee TING THE pe WAYS OF TEST ALITY OFA TRANSLATION Foch Q sbanenslarion, one translates # txt In back-ansla are mnpuage Bs a different transl Be tek nto Ay and there ing cs ns the Bcext bac sting tei e the original A text. Ifthe rexts are ns vinong evidence that the original a pyaaton was of high quality (though nor incontro- idence, because the second translaror might en a eee Pre ledge resins speakers of language B are fested about the content ofthe translation (e.g. using a ‘iestionnaite), and the same questions are put to speakers ofA. IFthe results correspond, the translation must be efficient. © In performance testing, speakers of language B are asked to carty out actions based on the text (eg. in a repair manual), as are speakers of A. The results can then be compared to determine translation efficiency. This isa very time-consuming process, however, and requires expert supervision if clear results are to be obsained A good back-translation... Original: Leaks occurring beyond relief valve could ‘cause some indication of low oil pressure. Back-tronslaion: If oil is leaking at the outside of the pressure relief valve, it can activate the warning of oil low pressure, anda bad one Grginal:Toubleshooting precautions. ‘ranslation: Prevencions while repairing, (R, W. Brislin, 1976, p.10.) ee TWAIN'S BACK-TRANSLATION Sens sro ee aac muneerey ecg saan at seceanerments romana mososceee ee =e teense feller touched ae ~ -TRANSLITERATION When the source language is written in a different script from the target language ($33), itis often necessary to providea manslitention ofan original word rather than a translation ~ something commonly done with the names ofpeople places, institutions, and inventions. Here, each Character of the source language is converted into a cha acter of the target language; for example, Russian ‘anvranx ‘companion, satellite’ becomes sputnik Transliteration needs to be distinguished from mman- scription, in which the sounds of the source words are conveyed by lecers in the target language. For example, an English transcription of Soviet premier Gorbachev’ name would have to be Gorbachoff wo reflec the way iis pronounced in Russian, This approach is often unavoidable with languages that use partial alphabetic scripts (e.g. Arabic, p, 204), where transliteration would be very difficuls, or logographic scripts (eg. Japanese Aang, p. 202), where it would be impossible Both approaches have their problems. With transcrip- tions, the target equivalents are likely co differ when the words are converted into different languages, asin English Tebaikowsky; Dutch Tyiaokowstaj, Hungarian Ciajkovsef Wich cranslterations, there is often the problem of there being insuffcienc symbols inthe target language (so that dlacrtcs have tobe added) or too many symbols (in which case arbitrary choices have to be made). Asbitrariness is ‘most noticeable when there is no close correspondence berween the sounds ofthe source and target languages. Asaresult, there are often several transliteration systems available, as has happened with Chinese (p. 314) and Russian (where, for example, the familiar form Khrushchev could appear as Xruséey, or in several other mp fe absene ofan ncemationdly aged scheme, i ty difficult to trace terms and names in international indexes, A Russian name beginning with might be transliterated into ia, ja, or ye with nance tettieva problemsunlesstheconvenionsseam iene nvetsion system is known. ONLY A NAME... Transliteration makes espe- ‘ial sense with foreign per. sonal or brand names, ‘where unfortunate conse farm maciinery—both ‘examplesof translations from Chinese, Hilarity or vere unfortunate embarrasiment accompa- hon canreatifatrans. nieenameswnenate ave Iwouldberiky tounge beer tigud soce Fan Game CHibe Btn simue kore eee intosome emis Beet ‘Bimbo bread from spain. frrecrmestiialent word" Theproman onsen fenguage,incasethecon. _SQmPlEated when western rotate wetetecor- — namesareturneainte ee, Jameriskappliesin aby languages (p. 202). soon ric aneta Fistthename hos tobe Se Cee tire eae Taneforrateting or ge lenange Costa for Sind athename st jie yracbeen fumed RUSSiany Ones se alerts for Rage Over ata Py Pree ele |e =\5 tar, Each yl ne cante bere" characters exch ihe sit meaning Aarts ofthe aretherctore pestle Somectwich mi flattering, otess 8 (One! meaning oe seems, shes | tion-bare) Terese tionaries which ove! ‘safe’ equivalents 10" ‘emnames—usualy hosting te oa which prompt Pron rater than asema interpretation.‘contemporary needs is the ¥ rerminology. If people are the terms used N AN get faton 0 1 cT0ss PUFPOSES, ie eT an cokes of practice scientific 10 ctinal mar cetings, and many other situa- ernment ized. S tific a" rf opel ically at risk, where a a plcaced. Several efforts have ld eal organize data banks of termi- tising computational tech- crags and rial: Fer a vepean Union (EU) term bank, eagle a TOM (European Automatic Dictio- RODICAT lily accessible on-line via Euronet, sag) 80% PY nw in the official languages of the and comming cetm or abbreviation is listed along, Community example, equivalent items in other 920M denn, ad bibliographical reer zg et saute tan eee etna ee abut ‘abusivelincorrect excessive heer cachievefinsh (of) trertsement_ *advertisementivaraing bande “bandigang Capette “carpevrig denender_*demandirequest ‘ente! *eventualpossibie ‘urdewr *fatidiouskivesome ‘tome *iiorvlanguage (of a group) ‘rchuent.*incoherenvinconsistent ‘sarnérient_tinconvenientidrawback “rfematin‘informationinews Sit larbscon Ibeler slbebmake outa cheque Paton partion musical score int “petulanvtvely ‘ohratelsentence “sympatheticagreeable realitie oe Wa Seaton TRANSLATING AND INTERPRETING HO TER (PON ALL ues PRESS 28S FOR TeRAUNOLOGY 8 XR seaeUtnioy TYPE COME OF souRce Lancunce 56 GERMAN BK BRNISH EG neLigy TT TTALIAW oe a NO WTCH FT RTUGUEsE SP SBN TPE CODECS) OF TARGET Lancunce;s) {FOE NPE: 6 HO Oe Fo UY aes SHE SOURCE LaNGuRGE TARGET LANGUAGE(S) 206 1T Fi SURECT COME en PRESS 9 OR ANOTHER coMHAKD 0 TWPE YOUR QUESTION INFLATION a Does pate =f | Bee pra F124 Me O00128 ames MERE Cfo oh EC4 ECB £6 VE rerarion . PH atce INFLATION IS most LIKELY To OCCUR ANEW aenmaD INCREASES MMTCE TAE LAVOUR SUPPLY 15 TURE MKB THE INDUSTRIAL CAPRCETY 15 FULLY UTILIZES. «AEN SOURCES. SUPPLY aRY UP. 2WFLATION EXCES DE POWOIR DPACMHT OU EXCES ES HOTENS DE ON UA CONFORP SOUVENT...AUEC LA SINPLE MRUSSE 265. ‘Op CELLE*CI EST LA CONSEQUENCE DE LPIHLATION UIIRFLATION ELLE-HENE. PRESS CTO CONTINUE OR Gt Fo ME PH ‘The result of an on-line search for the term inflation in EURODICAUTOM. Several fepsareimel 4. Does the enquirer require a term or an abbre- viation (Lorxi? 2. Which source and target languages are involves? (in ‘the example, the source an ‘guage was nash, and: other EU languages were targets) TRANSLATION SHIFTS TheFrenddetinte ate Jellatt'iles' thought of asthe eave lent of the English definite article the. sty are many except rule, where ee orcPART X ATIONS? a irene pepe have sing versions of the BEST TRANSLATH ‘Ageadeal cin beleuned about he lating by comparing cass translated the of Bau opening verse attention paid to for ‘ment of rhyme, rhythm, wor Sonvent pour sas leshonomesdéipage Prennent des albatos,vastesiseae des mers Quisuivent, indolenn compagnons de vase, Lenavne glivantsurlesgouffesamers Ofien coamuse themselves, the men ofthe crew trap alba- twose, the reac se-birds, tat follow the ship slipping over the bitte deeps, ke idle travelling companions. erafof trans ue eae the different the varying treat- rm and content in r order an excl choice. (A. Hadley) (Often for thcie amusement, sailors catch albatross, those vastbinds of these, indolent companion of their voyages, that follow the ship gliding across the bitter depths. (F Scarfe) Inorderto amuse themselves, the members ofthe crew fen catch albatoses, those huge seabirds which, asindo- lene companions on the voyage follow the ship gliding over the bier depths (LE Fiaky) WHEN FOREIGN IS BEST Sometimes it pays norco translate, asthe business world has long known, Sales can beneficifa product is given a foreign name. In 1960 a Finnish firm distributed tinned coffee for the home marker using Finnish labels, Sales ‘were poor. The firm then had new labels made with {excin English on thesame tins, and sales rocketed. Sim. ilar, English marketing firms and other businesses make use of foreign languages to convey special effects ~ such a the use of French for the names of restaurants, night-clubs, and perfumes. In one page ofa British tele, Phone directory, under ‘Restaurants, nearly half the ‘ames were ina language other than English ~ La Bella quality and reliability (eg ‘Crown stressed, a French nameischosen ar often hasan Italian mame (ey ‘The linguistic effecs are most commercials, where appropriate Amen LANGUAGE IN THE WORLD sechalbatrosses, great birds oft hi follow indolently svhich companions the vesel gliding over biter gulf. (CE Macingyre) en of loafing crews Sometimes for sport the men of loafing Snare the great albatrosses of the deep, The indolent companions oftheir cruise ‘As through the biter vastitudes they sweep. (R. Campbell [screw will often take In sport a vessel's crew The mighty albatross, who on the breeze Doth idly sail and follow in the wake Ofships thar glide upon the biter ses. ee (A. Conder) (Often, asan amusement, crewmen. Catch albatrosses, huge birds of the ea, Who follow, indolent companions ofthe voyage, ‘The ship gliding over the salty decps. ‘cael (W. Fowlic) (From I. E Finlay, 1971, pp. 129-32.) not understand them, but the connotations of prestige associated with these languages are enough to warrant their use. The purpose ofthe language (§4) is not to com- municate ideas, but to appeal to the sensibilities of che Japanese viewer, who the manufacturers believe is much influenced by the values of modern cosmopolitan society. FOREIGN ELEMENTS IN JAPANESE COMMERCIALS English Ses everybody, new, life now, healthy, power, bi S31 open happy nic, beautiful night eee Oe French Plasir‘pleasure’, mage image’, ‘spring’. chocolat ‘chocolate German Auslese ‘selection, s ‘cream, Wagen ‘car ‘0fé ‘coffee’, printemps accessoire ‘accessory’, hick ‘chic, schon beautiful, Sahne Sweetheart, manifesto mani, “good morning’. ifesto ‘manifest’, buongiorno 7 SOME JAPANESE Sunday, bonita ‘pretty: NAMES a Ko common, suchas high Doibat Nisa, feeandmooth, quickand | Domino Biuebié sreleaance‘thestandardot | Cherage cher Sip ioanerece: Mole sentences — | Charmant Laurel iy lune ferme qui aime la vie" Mier ct \$poken by Japanese Pate Ges? oF foreign engage pls ru soR capsule auto akkuse Hasrman, 1964),day widely known from its use in is cod fe (p- 344). When senior min we language backgrounds meet, the jlferent ply shows a pair of interpreters od invariably m8 olit al pe wound, At major conferences | he ba jons General Assembly, the he deh nes is a clear indication that a * headphone is taking place. In everyday soe ings Frequently are ceded oo. interpreters sane elitan societies formed by new ily in coersrants and Gastarbeter ($9) ns nes of I courts, hospital, local mics, eassr00S; 6 loo without the prsence ean ine importance and frequency of this pe eis emarkable that so litle study h wha actually happens when interpreting, se ofhow successul an exercise itis | sree he eecency of developments in the field eatin this neglect. One procedure, known as Tee oe vimerpreting, is very old ~ and presumably ‘rin che Tower of Babel! Here, the interpreter spas afer the speaker has finished speaking Wier in shor bursts, ot at the very end of a dis- sum), Thisapproach is widely practised in informal hatin aswell 3s in committees and small confer- feat In lrgerand more formal setings, however, i to fen generally ceplaced by simultaneous incerpret- og-areeent development that arose from the av ahiy of modern audiological equipment and the ‘draofinceased international interaction following seScond World War, OF he vo procedures, itis the second chat has stated mos interest, because of the complexity of thaland the remarkable skills required. In no other ‘atest ofhuman communication is anyone routinely "ied licen and speak a the same time, preserv- ‘ana semanticcorespondence between the wo ab Mateos ther is invariably a delay ofa few Tabewen the scimulus and the response, because fests to assimilace wha is being said in ale ing and to translate it into an accepr- inter in the target language. This ‘ear-voice span’ Pisusully about? ah Stell abour2 03 seconds, bute may be as bra 2A oF, ifthe text is complex. The | Sree what has just been said, atend Snaion tty being said, and anticipate the con- atisabout to be said. As one writer has ot industrial cribunals can- Pui saa sen | a "eticeyouare taking a leap in the dark, you sudden Sm matical Future; the original sen- Suddenly be turned i ‘urned in such a way that your trans ‘easily be reconciled with your reat nimbleness is called for to onto itsstar, Ss Bue the mind through chs ime itis engaged upon R.Glemer Alisasource cfs ntl lear. Tha ing bood nce of back ateness that Pon the accuracy of the werk, Rens ave now begun to look a dletermine, for example, how fy ing is affected by speed at which the sourelangus that an input speed of betwee | minute is a comforta may depend pect Reena at these factors — eo succes interpre & conditions, of the phen Isms and 120 words per ble rat for intepraing ate upper limic of around 200 w.pim. But even smal increases in speed can dramatically affct the secur of output. In one controled study, when speeds wer sradually increased in a series of stages rom 95 to 164 wep.m,, the ear-voice span also increased with cach stage, and the amoune correctly interpreted showed a clear decline (D. Gerver, 1969), Als, asthe eandating load increaus, no ony are there mow eros con mission (mistranslations, cases of vagueness replcin precision) there are also more etre of omiston, 2 words and segments of meaning ae filtered out. These are important findings, given the neec for accuracy in international communication, What is needed is a more detailed identification ofthe problem ares, and of the strategies speakers, lsteners, and interpreters use tosolve them. There is urgent need to expand what has so far been one of the most neglected fields ofcommu- nication research, ee Incpmee Hae Hag Hap Hand ee Soh ala sit | Sooty ‘THE EAR-VOICE SPAN When the inputlanguageis Sraihtionard the met eter willbe aeguar20°3 Seconsbehind the speakes But when an unexpected: tueletfity emerges ofthe Speaker sudden pee, re clay mayinceae wth Seonsaquental ctor" fect at ay ake some Sime toresohe,Duting this paid, hen theearoke Sean serra econdslona Reterpreterscoontie Trane procesig abies sounder great pres ‘teommon inguisicre sontoritepretng dl Sconwher range SSnguage where te ae cen nese {hee Gov languages $14 intoalanguage wher rate ening materia 7 frobiem oxen ae Protating panes tO Engi and 0oroneu transiting ean ateclausesMACHINE TR NSLATION 10 provide eanslacons Me ied since he idea of using machines t os peoween ature guages pee ee ad ory, the success of advanced gi baking pas and the invention of the electronic computer § cee tlt mancenon (4 T) could bea reality. Warren Weaver, a founder of the field, caught the optimism in a 1947 memorandum: ‘One naturally wonders if the problem of translation could conceiv- ably be reated asa problem of eryprography. When I ook at an article in Russian, I say: “This is really writ- ten in English, but it has been coded in some strange symbols. I will now proceed t0 decode.” "As a resul several groups began rescarch programmes into MT. during the 1950s, and great claims were made for the furure of the subject (W. Weaver, 1955). However, initial results were not encouraging. The systems proved to be very limited in the kind of daca they could handle. lations were crude, full of errors, and required so much human post-editing that they proved to be more expensive than having a human translator carry out the whole task in the fitst place. The main reason was the lack of a sufficiently sophisticated linguistic theory to provide a frame of reference for the tasks that MT needed to undertake, The earliest MT. system did litle more than look for equivalences benween the words in cach language ~ in effect, they acted as an automatic bilingual dictionary. After several decades of linguistic rescarch, iti easy to se why these approaches could not have worked. They ignored the problem posed by the grammatical dimension of lan guage analysis ~ the different levels of syntactic organi- zation (916), and the absence of straightforward formal correspondences between units of grammar (such as is LANGUAGE IN THE WORLD illustrated by the use of the definite article, p. 349) They also ignored the different ways in which languages structure meaning; word-for-word translation is offen not possibleand usually nor desirable (p. 346). Nor was there any way of distinguishing benveen the different senses of words or deciding whether a group of words was an idiom. Many ambiguities can be resolved only by using an analysis in terms of semantics ($17) or of real-world knowledge, and such analyses were notavail- able at chat time, There was evidently a great deal more o MT than ‘code breaking’. w Thediseatsfuaion was summarized in a US report of 1966 by the Automatic Language Processing Advisory Committee (atrac), which concluded that human translating was faster, more accurate, and less expensive than MT, and that no further support for the latter should be provided. Asa consequence, only a minimal amount of MT research was carried on in subsequent years, either in the USA or in Europe (though contin- ued support was provided in the Soviet Union), ANEW MOOD ‘The pendulum has begun to swing back again in rccent -yeats, following the major intellectual and technological developments of the 1970s in linguistics ($65) and com. puting, A new mood isabroad, promoted by the promis- jing. practical achievements of new commercial MT projects, by the great potential of the new research pro- grammes in artificial intelligence, and by an increased theoretical awareness of the translation task which has come from progress in linguistics. There is also a greater realism concerning what MT can and cannot do, and a recognition of the need to devise techniques of humanfmachine collaboration, in order to get the best results from both. The main developments have been to Provide systems of analysis that allow for grammatical and semantic complexity. The first steps were in de auromatic procedures (algorithms) for parsing the syn= tactic strucrure of a sentence, and for carrying. out an SYSTRAN (ne of the best-known fst : ret inn temiceskixstedsty, Moscow firpor Stan waster Sertannglh | Slvamesenatotsoge — OSSAIDON menceronaer eter thekindefoutputpro. Lmersemadtenceny, NE YKKAIBOr Temietgeaan eral ducediySYaTRAW andthe suceana yen Revised output ianunen foramletndol go eateg {neprimer moskovskit ‘The maderitport isan Mringteairathe scons anit me ageeonmede, dabortecomplecof ra awrareeemeamohe tumbasinennan ™ perscitl aropor nieldateesocuinnase came ob ansitertion hove boen : cal vies requtring scelemaipeelieiby atge territory whi Beier eee en, SRR ree feundintngcampnt’et Acontemporay armors Semspunae Measures hou: tre inthe computer nuove co omod Feral ete, (anBesed raters, gemeaedcomplnct | Deer rie ar Lunceacinvewof he emer orgin) menrorenegyatange. Airport in New York) seer ann Mentofwhichtheterrion, (Eknowies Whee 3 pena fore po enny! a3roport measured sme esol siehesEonornA MM Preto oiy — Meuledsometies Komelehsnenerny soon, hetero me — APRE-EDie man Theor Senses. intongose ets Universi sorte tor CULT tear Pens Teavtomstcgen ea Chinese urna ating mathemati a ireemay A amount of reef inci Pe, Ponta ‘Stig innernat on the struct ‘ tala between thelengu elenage emer mmulatedtocnabere” enced andevenmor People to ranean the inputintomastt, latable form Wants ‘arrangement poses ‘an bekepttoamimny notaltogethercmars, Giventimeanadeteipe ramming technic pre-editing rules gud Bly be redcedo tht Computer weer take up his route vn Preeditingcanthuclon Solve many of heres uisticprobiems eet erwise dependentontte research in natura ago, computations ingusss transformation ater Inthe words, mode are much mare cone sive and sphistiated the present onesare designed. Thesemoseb PA take years topertetard present, preeing 335" Iitelyesertiatinoe achieve the goal np ‘age of our develope wey complxseter, be translated wth es pre-eaitng. Aserte%® ‘which has cor structurecan beans the existing posta broken upintosimpe' ences icv et ranslatedy Sentencesin hin often without verbs orn Fecsand pee theverbs ors! ie these sentencescanti analysed andres! {Sc ton, 1976.0."