SADiLaR Language Resource Repository
Not a member yet
536 research outputs found
Sort by
Linguistically enriched corpora for conjunctively written South African languages
No full textThis resource contains linguistically annotated data for four official South African languages with a conjunctive orthography from the Nguni family (isiNdebele, isiXhosa, isiZulu and Siswati) as well as English. The data set is parallel for all five languages and the Nguni languages have been annotated for three different types of linguistic information: morphology, part-of-speech and lemmas. We have also included the protocols and tagsets used during annotation
Code-switching among bilingual Afrikaans-Dutch children
No full textTo describe the code-switching in the speech of pre-school, bilingual Afrikaans-Dutch children, novel language data of eight preschool children between the ages of 2;3, and 6;4 who are raised bilingually in Afrikaans and Dutch in South Africa were collected and analysed. The parent-report method was employed to collect the data, and the data collection period lasted a total of approximately two months.
The NWU HSSEC ethics number for the research is: NWU-02043-20-S
CGE's Afrikaans Gender Terminology List
No full textCGE's Afrikaans Gender Terminology List is a list of terms, either words or phrases, related to the promotion of gender equality. All 436 words or phrases appear in English, with definitions in English, and with equivalent Afrikaans words or phrases. The list is the result of a collaborative effort between the Commission for Gender Equality (CGE) in partnership with the PanSALB Afrikaans National Language Body (ANLB), one of PanSALB's language boards
Autshumato English-isiNdebele Parallel Corpora
No full textAligned parallel corpora for the following language pair: English-isiNdebele. Data was crawled from various multilingual government websites, sourced from translated material and created by translating English sentences into isiNdebele
Autshumato Monolingual isiNdebele Corpus
No full textMonolingual corpus for isiNdebele. The data is given as a single UTF-8 text file, with each segment on a newline
Core technologies for conjunctively written South African languages
No full textDuring this SADiLaR funded project, enriched corpora for the four official South African languages with a conjunctive orthography,
i.e. isiNdebele (NR), isiXhosa (XH), isiZulu (ZU), and Siswati (SS) was developed. The corpora consist of approximately 50,000 tokens,
parallel on sentence level, with English as source language, for each language. Each language’s corpus was annotated on three levels, namely morphological analysis, part of speech and lemmatisation (see: https://repo.sadilar.org/handle/20.500.12185/546).
Using the annotated data, 12 core technologies, i.e. morphological analysers, POS taggers and lemmatisers for each of the four languages were developed and packaged in a single graphical user interface (UI)
Mburisano Covid-19 multilingual corpus
No full textThis corpus was created to aid development of the AwezaMed Covid-19 speech-to-speech mobile application. The project within which it was created, Mburisano, was funded by the Department of Sport, Arts and Culture (DSAC).
A selection of English sentences was generated in consultation with medical domain experts, and these sentences were manually translated into all official South African languages. The sentences formed the basis of the rapid development of Grammatical Framework (GF) application grammars for all the languages, to aid spoken communication about Covid-19 with a particular focus on screening and triage. The corpus is presented as a limited domain, manually translated parallel corpus in all 11 official South African languages.
The AwezaMed Covid-19 application can be found [here](https://play.google.com/store/apps/details?id=za.co.aweza.covid19&gl=ZA)
Autshumato Monolingual Tshivenḓa Corpus
No full textMonolingual corpus for Tshivenḓa. The data is given as a single UTF-8 text file, with each segment on a newline
Denominal adjectives in Afrikaans dataset
No full textThis dataset contain a collection of Afrikaans denominal adjectives that were extracted from the Virtual Institute for Afrikaans' corpus portal. The denominal adjectives are only those that end with either ·agtig or ·erig - two common adjective-forming affixes in Afrikaans. This dataset has been enriched by mainly morphological annotation, but also some semantic annotation.
An updated version of this dataset (version 2020-10-28) is used in Denominal adjectives in Afrikaans: The cases of ·agtig and ·e·rig (2020)
Corpus of multilingual code-switched soap opera speech
No full textThe corpus comprises 26.9 hours of annotated multilingual speech that contains examples of code-switching in isiZulu, isiXhosa, Setswana, Sesotho and English. The speech was obtained from South African soap operas. Code-switching between English and one of the Bantu languages is by far most prevalent in the data. Although not very common, switches between the Bantu languages themselves also occur. An initial attempt to align the audio extracted from soap opera episodes with the corresponding scripts revealed that actors very often perform ad lib. The speech and the examples of code-switching it contains can therefore be considered to be spontaneous