<?xml version="1.0" standalone="yes"?>
<Paper uid="W05-0109">
  <Title>Natural Language Processing at the School of Information Studies for Africa</Title>
  <Section position="3" start_page="49" end_page="50" type="intro">
    <SectionTitle>
2 Languages and NLP in Ethiopia
</SectionTitle>
    <Paragraph position="0"> Ethiopia was the only African country that managed to avoid being colonised during the big European power struggles over the continent during the 19th century. While the languages of the former colonial powers dominate the higher educational system and government in many other countries, it would thus be reasonable to assume that Ethiopia would have been using a vernacular language for these purposes.</Paragraph>
    <Paragraph position="1"> However, this is not the case. After the removal of the Dergue junta, the Constitution of 1994 divided Ethiopia into nine fairly independent regions, each with its own &amp;quot;nationality language&amp;quot;, but with Amharic being the language for countrywide communication. Until 1994, Amharic was also the principal language of literature and medium of instruction in primary and secondary schools, but higher education in Ethiopia has all the time been carried out in English (Bloor and Tamrat, 1996).</Paragraph>
    <Paragraph position="2"> The reason for adopting English as the Lingua Franca of higher education is primarily the linguistic diversity of the country (and partially also an effect of the fact that British troops liberated Ethiopia from a brief Italian occupation during the Second World War). With some 70 million inhabitants, Ethiopia is the third most populous African country and harbours more than 80 different languages -exactly how many languages there are in a country is as much a political as a linguistic issue; the count of languages of Ethiopia and Eritrea together thus differs from 70 up to 420, depending on the source; with, for example, the Ethnologue (Gordon, 2005) listing 89 different ones.</Paragraph>
    <Paragraph position="3"> Half-a-dozen languages have more than 1 million speakers in Ethiopia; three of these are dominant: the language with most speakers today is probably Oromo, a Cushitic language spoken in the south and central parts of the country and written using the Latin alphabet. However, Oromo has not reached the same political status as the two large Semitic languages Tigrinya and Amharic. Tigrinya is spoken in Northern Ethiopia and is the official language of neighbouring Eritrea; Amharic is spoken in most parts of the country, but predominantly in the Eastern, Western, and Central regions. Oromo and Amharic are probably two of the five largest languages on the continent; however, with the dramatic population size changes in many African countries in recent years, this is difficult to determine: Amharic is estimated to be the mother tongue of more than 17 million people, with at least an additional 5 million second language speakers.</Paragraph>
    <Paragraph position="4"> As Semitic languages, Amharic and Tigrinya are distantly related to Arabic and Hebrew; the two languages themselves are probably about as close as are Spanish and Portuguese (Bloor, 1995). Speakers of Amharic and Tigrinya are mainly Orthodox Christians and the languages draw common roots to the ecclesiastic Ge'ez still used by the Coptic Church. Both languages use the Ge'ez (Ethiopic) script, written horizontally and left-to-right (in contrast to many other Semitic languages). Written Ge'ez can be traced back to at least the 4th century A.D. The first versions of the script included consonants only, while the characters in later versions represent consonant-vowel pairs. Modern Amharic words have consonantal roots with vowel variation expressing difference in interpretation.</Paragraph>
    <Paragraph position="5"> Several computer fonts have been developed for the Ethiopic script, but for many years the languages had no standardised computer representation. An international standard for the script was agreed on only in year 1998 and later incorporated into Unicode, but nationally there are still about 30 different &amp;quot;standards&amp;quot; for the script, making localisation of language processing systems and digital resources  difficult; and even though much digital information is now being produced in Ethiopia, no deep-rooted culture of information exchange and dissemination has been established. In addition to the digital divide, several other factors have contributed to this situation, including lack of library facilities and central resource sites, inadequate resources for digital production of journals and books, and poor documentation and archive collections. The difficulties of accessing information have led to low expectations and consequently under-utilisation of existing information resources (Furzey, 1996).</Paragraph>
    <Paragraph position="6"> UNESCO (2001) classifies Ethiopia among the countries with &amp;quot;moribund or seriously endangered tongues&amp;quot;. However, the dominating languages of the country are not under immediate threat, and serious efforts have been made in the last years to build and maintain linguistic resources in Amharic: a lot of work has been carried out mainly by Ethiopian Telecom, Ethiopian Science and Technology Commission and Addis Ababa University, as well as by Ethiopian students abroad, in particular in Germany, Sweden and the United States. Except for some initial efforts for the related Tigrinya, work on other Ethiopian languages has so far been scarce or non-existent -- see Alemu et al. (2003) or Eyassu and Gamb&amp;quot;ack (2005) for short overviews of the efforts that have been made to date to develop language processing tools for Amharic.</Paragraph>
    <Paragraph position="7"> One of the reasons for fostering research in language processing in Ethiopia was that the expertise of a pool of researchers in the country would contribute to maintaining those Ethiopian languages that are in danger of extinction today. Starting with Amharic and developing a robust linguistic resource base in the country, together with including the Amharic language in modern language processing tools could create the critical mass of experience, which is necessary in order to expand to other vernacular languages, too.</Paragraph>
    <Paragraph position="8"> Moreover, the development of those conditions that lay the foundations for language and speech processing research and development in the country would prevent potential brain drain from Ethiopia; instead of most language processing work being done by Ethiopian students abroad (at present), in the future it could be done by students, researchers and professionals inside the country itself.</Paragraph>
  </Section>
class="xml-element"></Paper>