Text in English / Abstracts in English, isiXhosa and Afrikaans / The Constitution of the Republic of South Africa, Act 108 of 1996, mandates the state to take practical and positive measures to elevate the status and the use of indigenous languages. The implementation of this pronouncement resulted in a growing demand for specialised translations in fields like technology, science, commerce, law and finance. The lack of terminology and resources such as specialised bilingual dictionaries in indigenous languages, particularly isiXhosa remains a growing concern that hinders the translation and the intellectualisation of isiXhosa. A growing number of African scholars affirm the importance of specialised dictionaries in the African languages as tools for language and terminology development so that African languages can be used in the areas of science and technology. In the light of the background above, this study explored how parallel corpora can be interrogated using a bilingual concordancer, ParaConc to extract bilingual terminology that can be used to create specialised bilingual dictionaries. A corpus-based approach was selected due to its speed, efficiency and accuracy in extracting bilingual terms in their immediate contexts. In enhancing the research outcomes, Descriptive Translations Studies (DTS) and Corpus-based translation studies (CTS) were used in a complementary manner. Because the study is interdisciplinary, the function theories of lexicography that emphasise the function and needs of users were also applied. The analysis and extraction of bilingual terminology for dictionary making was successful through the use of the following ParaConc features, namely frequencies, hot word lists, hot words, search facility and concordances (Key Word in Context), among others. The findings revealed that English-isiXhosa Parallel Corpus is a repository of translation equivalents and other information categories that can make specialised dictionaries more user-friendly and multifunctional. The frequency lists were revealed as an effective method of selecting headwords for inclusion in a dictionary. The results also unraveled the complex functions of bilingual concordances where information on collocations and multiword units, sense distinction and usage examples could be easily identifiable proving that this approach is more efficient than the traditional method. The study contributes to the knowledge on corpus-based lexicography, standardisation of finance terminology resource development and making of user-friendly dictionaries that are tailor-made for different needs of users. / Umgaqo-siseko weli loMzantsi Afrika ukhululele uRhulumente ukuba athabathe amanyathelo abonakalayo ekuphuhliseni nasekuphuculeni iilwimi zesiNtu. Esi sindululo sibangele ukwanda kokuguqulelwa kwamaxwebhu angezobuchwepheshe, inzululwazi, umthetho, ezemali noqoqosho angesiNgesi eguqulelwa kwiilwimi ebezifudula zingasiwe-so ezinjengesiXhosa. Ukunqongophala kwesigama kunye nezichazi-magama kube yingxaki enkulu ekuguquleleni ngakumbi izichazi-magama ezilwimi-mbini eziqulethe isigama esikhethekileyo. Iingcali ezininzi ziyangqinelana ukuba olu hlobo lwezi zichazi-magama luyimfuneko kuba ludlala iindima enkulu ekuphuhlisweni kweelwimi zesiNtu, ekuyileni isigama, nasekusetyenzisweni kwazo kumabakala obunzululwazi nobuchwepheshe. Olu phando ke luvavanya ukusetyenziswa kwekhophasi equlethe amaxwebhu esiNgesi neenguqulelo zawo zesiXhosa njengovimba wokudimbaza isigama sezemali esinokunceda ekuqulunqweni kwesichazi-magama esilwimi-mbini. Isizathu esibangele ukukhetha le ndlela yophando esebenzisa ikhompyutha kukuba iyakhawuleza, ulwazi oluthathwe kwikhophasi luchanekile, yaye isigama kwikhophasi singqamana ngqo nomxholo wamaxwebhu nto leyo eyenza kube lula ukufumana iintsingiselo nemizekelo ephilayo. Ukutyebisa olu phando indlela yekhophasi iye yaxhaswa zezinye iindlela zophando ezityunjiweyo: ufundo lwenguguqulelo oluchazayo (DTS) kunye neendlela zokuguqulela ezijoliswe kumsebenzi nakuhlobo lwabasebenzisi zinguqulelo ezo. Kanti ke ziqwalaselwe neenkqubo zophando lobhalo-zichazi-magama eziinjongo zokuqulunqa izichazi-magama ezesebenzisekayo neziluncedo kuninzi lwabasebenzisi zichazi-magama ngakumbi kwisizwe esisebenzisa iilwimi ezininzi. Ukuhlalutya nokudimbaza isigama kwikhophasi kolu phando kusetyenziswe isixhobo sekhompyutha esilungiselelwe ikhophasi enelwiimi ezimbini nangaphezulu ebizwa ngokuba yiParaConc. Iziphumo zolu phando zibonise mhlophe ukuba ikhophasi eneenguqulelo nguvimba weendidi ngendidi zamagama nolwazi olunokuphucula izichazi-magama zeli xesha. Kaloku abaguquleli basebenzise amaqhinga ngamaqhinga ukunika iinguqulelo bekhokelwa yimigomo nemithetho yoguqulelo enxuse abasebenzisi bamaxwebhu aguqulelweyo. Ubuchule beParaConc bokukwazi ukuhlela amagama ngokwendlela afumaneka ngayo kunye neenkcukacha zamanani budandalazise indlela eyiyo yokukhetha imichazwa enokungena kwisichazi-magama. Iziphumo zikwabonakalise iintlaninge yolwazi olufumaneka kwiKWIC, lwazi olo olungelula ukulufumana xa usebenzisa undlela-ndala wokwakha isichazi-magama. Esi sifundo esihlanganyele uGuqulelo olusekelwe kwiKhophasi noQulunqo-zichazi-magama zobuchwepheshe luya kuba negalelo elingathethekiyo kwindlela yokwakha izichazi-magama kwilwiimi zeSintu ngokubanzi nancakasana kwisiXhosa, nto leyo eya kothula umthwalo kubaqulunqi-zichazi-magama. Ukwakha nokuqulunqa izichazi-magama ezilwimi-mbini zezemali kuya kwandisa imithombo yesigama esinqongopheleyo kananjalo sivelise izichazi-magama eziluncedo kwisininzi sabantu. / Die Grondwet van die Republiek van Suid-Afrika, Wet 108 van 1996, gee aan die staat die mandaat om praktiese en positiewe maatreëls te tref om die status en gebruik van inheemse tale te verhoog. Die implementering van hierdie uitspraak het gelei tot ’n toenemende vraag na gespesialiseerde vertalings in domeine soos tegnologie, wetenskap, handel, regte en finansies. Die gebrek aan terminologie en hulpbronne soos gespesialiseerde woordeboeke in inheemse tale, veral Xhosa, wek toenemende kommer wat die vertaling en die intellektualisering van Xhosa belemmer. ’n Toenemende aantal vakkundiges in Afrika beklemtoon die belangrikheid van gespesialiseerde woordeboeke in die Afrikatale as instrumente vir taal- en terminologie-ontwikkeling sodat Afrikatale gebruik kan word in die areas van wetenskap en tegnologie. In die lig van die voorafgaande agtergrond het hierdie studie ondersoek ingestel na hoe parallelle korpora deursoek kan word deur ’n tweetalige konkordanser (ParaConc) te gebruik om tweetalige terminologie te ontgin wat gebruik kan word in die onwikkeling van tweetalige gespesialiseerde woordeboeke. ’n Korpusgebaseerde benadering is gekies vir die spoed, doeltreffendheid en akkuraatheid waarmee dit tweetalige terme uit hulle onmiddellike kontekste kan onttrek. Beskrywende Vertaalstudies (DTS) en Korpusgebaseerde Vertaalstudies (CTS) is op ’n aanvullende wyse gebruik om die navorsingsuitkomste te verbeter. Aangesien die studie interdissiplinêr is, is die funksieteorieë van leksikografie wat die funksie en behoeftes van gebruikers beklemtoon, ook toegepas. Die analise en ontginning van tweetalige terminologie om woordeboeke te ontwikkel was suksesvol deur, onder andere, gebruik te maak van die volgende ParaConc-eienskappe, naamlik, frekwensies, hotword-lyste, hot words, die soekfunksie en konkordansies (Sleutelwoord-in-Konteks). Die bevindings toon dat ’n Engels-Xhosa Parallelle Korpus ’n bron van vertaalekwivalente en ander inligtingskategorieë is wat gespesialiseerde woordeboeke meer gebruikersvriendelik en multifunksioneel kan maak. Die frekwensielyste is geïdentifiseer as ’n doeltreffende metode om hoofwoorde te selekteer wat opgeneem kan word in ’n woordeboek. Die bevindings het ook die komplekse funksies van tweetalige konkordansers ontknoop waar inligting oor kollokasies en veelvuldigewoord-eenhede, betekenisonderskeiding en gebruiksvoorbeelde maklik identifiseer kon word wat aandui dat hierdie metode viii doeltreffender is as die tradisionele metode. Die studie dra by tot die kennisveld van korpusgebaseerde leksikografie, standaardisering van finansiële terminologie, hulpbronontwikkeling en die ontwikkeling van gebruikersvriendelike woordeboeke wat doelgemaak is vir verskillende behoeftes van gebruikers. / Linguistics and Modern Languages / D. Litt. et Phil. (Linguistics (Translation Studies))

Rekenaars kan ’n nuttige rol speel in vertaling. Twee benaderings is vertaalgeheuestelsels en masjienvertaalstelsels. By hierdie twee tegnologieë word ’n vertaalgeheue gebruik—’n tweetalige versameling vorige vertalings. Hierdie proefskrif bied metodes aan om die kwaliteit van ’n vertaalgeheue te verbeter. ’n Masjienleerbenadering word gevolg om foutiewe inskrywings in ’n vertaalgeheue te identifiseer. ’n Verskeidenheid leerkenmerke in drie kategorieë word aangebied: kenmerke wat verband hou met tekslengte, kenmerke wat deur kwaliteittoetsers soos vertaaltoetsers, ’n speltoetser en ’n grammatikatoetser bereken word, asook statistiese kenmerke wat met behulp van eksterne data bereken word. Die evaluasie van vertaalgeheuestelsels is nog nie gestandaardiseer nie. In hierdie proefskrif word ’n verskeidenheid probleme met bestaande evaluasiemetodes uitgewys, en ’n verbeterde evaluasiemetode word ontwikkel. Deur die foutiewe inskrywings uit ’n vertaalgeheue te verwyder, is ’n kleiner, skoner vertaalgeheue beskikbaar vir toepassings. Eksperimente dui aan dat so ’n vertaalgeheue beter prestasie behaal in ’n vertaalgeheuestelsel. As ondersteunende bewys vir die waarde van ’n skoner vertaalgeheue word ’n verbetering ook aangedui by die opleiding van ’n masjienvertaalstelsel. / Computers can play a useful role in translation. Two approaches are translation memory systems and machine translation systems. With these two technologies a translation memory is used— a bilingual collection of previous translations. This thesis presents methods to improve the quality of a translation memory. A machine learning approach is followed to identify incorrect entries in a translation memory. A variety of learning features in three categories are presented: features associated with text length, features calculated by quality checkers such as translation checkers, a spell checker and a grammar checker, as well as statistical features computed with the help of external data. The evaluation of translation memory systems is not yet standardised. This thesis points out a number of problems with existing evaluation methods, and an improved evaluation method is developed. By removing the incorrect entries in a translation memory, a smaller, cleaner translation memory is available to applications. Experiments demonstrate that such a translation memory results in better performance in a translation memory system. As supporting evidence for the value of a cleaner translation memory, an improvement is also achieved in training a machine translation system. / School of Computing / Ph. D. (Rekenaarwetenskap)

South Africa has eleven official languages, ten of which are considered “resource-scarce”. For these languages, even basic linguistic resources required for the development of speech technology systems can be difficult or impossible to obtain. In this thesis, the process of developing Spoken Language Identification (S-LID) systems in resource-scarce environments is investigated. A Parallel Phoneme Recognition followed by Language Modeling (PPR-LM) architecture is utilized and three specific scenarios are investigated: (1) incomplete resources, including the lack of audio transcriptions and/or pronunciation dictionaries; (2) inconsistent resources, including the use of speech corpora that are unmatched with regard to domain or channel characteristics; and (3) poor quality resources, such as wrongly labeled or poorly transcribed data. Each situation is analysed, techniques defined to mitigate the effect of limited or poor quality resources, and the effectiveness of these techniques evaluated experimentally. Techniques evaluated include the development of orthographic tokenizers, bootstrapping of transcriptions, filtering of low quality audio, diarization and channel normalization techniques, and the human verification of miss-classified utterances. The knowledge gained from this research is used to develop the first S-LID system able to distinguish between all South African languages. The system performs well, able to differentiate among the eleven languages with an accuracy of above 67%, and among the six primary South African language families with an accuracy of higher than 80%, on segments of speech of between 2s and 10s in length. AFRIKAANS : Suid-Afrika het elf amptelike tale waarvan tien as hulpbron-skaars beskou word. Vir die tien tale kan selfs die basiese hulpbronne wat benodig word om spraak tegnologie stelsels te ontwikkel moeilik wees om te bekom. Die proses om ‘n Gesproke Taal Identifisering stelsel vir hulpbron-skaars omgewings te ontwikkel, word in hierdie tesis ondersoek. ‘n Parallelle Foneem Herkenning gevolg deur Taal Modellering argitektuur word ingespan om drie spesifieke moontlikhede word ondersoek: (1) Onvolledige Hulpbronne, byvoorbeeld vermiste transkripsies en uitspraak woordeboeke; (2) Teenstrydige Hulpbronne, byvoorbeeld die gebruik van spraak data-versamelings wat teenstrydig is in terme van kanaal kenmerke; en (3) Hulpbronne van swak kwaliteit, byvoorbeeld foutief geklasifiseerde data en klank opnames wat swak getranskribeer is. Elke situasie word geanaliseer, tegnieke om die negatiewe effekte van min of swak hulpbronne te verminder word ontwikkel, en die bruikbaarheid van hierdie tegnieke word deur middel van eksperimente bepaal. Tegnieke wat ontwikkel word sluit die ontwikkeling van ortografiese ontleders, die outomatiese ontwikkeling van nuwe transkripsies, die filtrering van swak kwaliteit klank-data, klank-verdeling en kanaal normalisering tegnieke, en menslike verifikasie van verkeerd geklassifiseerde uitsprake in. Die kennis wat deur hierdie navorsing bekom word, word gebruik om die eerste Gesproke Taal Identifisering stelsel wat tussen al die tale van Suid-Afrika kan onderskei, te ontwikkel. Hierdie stelsel vaar relatief goed, en kan die elf tale met ‘n akkuraatheid van meer as 67% identifiseer. Indien daar op die ses taal families gefokus word, verbeter die persentasie tot meer as 80% vir segmente wat tussen 2 en 10 sekondes lank. Copyright / Dissertation (MEng)--University of Pretoria, 2010. / Electrical, Electronic and Computer Engineering / unrestricted