IsiXhosa Lexicography : Past , Present and Future

This article presents a panoramic and critical overview of isiXhosa lexicography and its impact on the intellectualisation of this indigenous South African language. The history of isiXhosa lexicography dates back more than two centuries. However, there still exists a need for dictionaries that serve the language-speaking community as practical tools for addressing diverse communication and learning-oriented needs in the current language policy dispensation. The IsiXhosa National Lexicography Unit (XNLU) is currently working on dictionary projects that attempt to address this situation while at the same time not losing sight of the mandate that the Pan South African National Language Board (PanSALB) placed on all the National Lexicography Units (NLUs). For this to happen, the article argues that the NLU needs to put lexicographic practice into its historical perspective, i.e. conceiving dictionary projects in the light of existing dictionaries and lexicographic traditions in the language. Over and above that, there is a need to take into account the recent developments in lexicographic research, adopt co-operative lexicographic practice and develop a dictionary culture among the isiXhosa-speaking community.


Introduction
This article presents a panoramic overview of isiXhosa lexicography up to the present.It offers a critical commentary of the major isiXhosa dictionaries and current dictionary products in the view of the line function of the National Lexicography Units (NLUs), of which the IsiXhosa National Lexicography Unit (XNLU) is part, in respect to isiXhosa as one of the official languages of South Africa.The main focus is on the impact of lexicography on this indigenous South African language and the lexicographic needs of the language-speaking community.
The history of isiXhosa lexicography, as briefly outlined in Section 2, dates back more than two centuries.The language currently boasts of a monumental three-volume dictionary, The Greater Dictionary of isiXhosa (GDX).This is in addition to the Kafir-English Dictionary (KED), the Oxford English-Xhosa Dictionary (OEXD), and a medium-sized monolingual dictionary, Isichazi-magama SesiXhosa (ISX), and many other dictionaries that are enumerated without detailed discussion in this article (cf.Table 1).
This article considers the strengths and limitations of the existing isiXhosa dictionaries.In the studied dictionaries, particular attention is paid to, among other aspects: -Purpose of each dictionary (as enumerated in the blurb, preface and introductory text); -Scope and size of the dictionary; -Data categories included in relation to the purpose of the dictionary; -User-friendliness of the dictionary.
Extending a similar approach to the current dictionary projects, this undertaking offers an up-to-date and futuristic view of isiXhosa lexicography, more than two decades after Peter Mtuze conducted his own survey (cf.Mtuze 1992).
Since then, a number of significant developments have taken place.Firstly, the Xhosa Dictionary Project (XDP), at the University of Fort Hare, has been reconstituted into the XNLU, thereby reconfiguring both the primary and secondary lexicographic processes which culminate in the production of isiXhosa dictionaries.Secondly, Volume 2 and Volume 1 of the GDX and the ISX have since been published, thereby rendering the inventory of isiXhosa lexicography richer than what it was in 1992.Finally, the currently on-going projects within the XNLU are being undertaken in a different context of language development, when isiXhosa is now one of South Africa's eleven official languages.This places big responsibilities on the lexicographers.The survey of isiXhosa lexicography conducted in this article, therefore, is not undertaken for its own sake.It is a stock-taking exercise that is meant to inform the future activities of the XNLU.
The number and size of some of the available dictionaries in isiXhosa project a good image of the language, particularly when compared to the other African languages in general.However, a closer and critical analysis of the dictionaries themselves necessitates a distinction between dictionaries for the language on the one hand and dictionaries for the language users and language learners on the other.The line function of the NLUs which requires them to focus on a particular type of dictionary, if not adopted strategically, is likely to add to a long list of dictionaries that cannot be used by the majority of the language speakers who need them.A strategic approach is required that does not consider the line function of the NLUs as diametrically opposed to periodic and regular publication of other types of dictionaries which address some immediate lexicographic needs in the linguistic community, while ultimately contributing to the intellectualisation of a particular language in the long term.For this to happen, practical lexicography needs to catch up with recent theoretical developments in lexicography in order to address some challenges that are faced in the actual compilation of the dictionaries, thereby ultimately addressing the needs of isiXhosa speakers and learners in a user-friendly way.

A Brief History of isiXhosa Lexicography
The history of isiXhosa lexicography can be traced back to the 18th century (cf.Pahl 1989 andMtuze 1992) and since then, the inventory of isiXhosa has grown as partly shown in Table 1 below.According to Pahl (1989) and Mtuze (1992), the pioneer of isiXhosa lexicography is Sparrman, a natural scientist who visited the Cape coast between 1772 and 1777 and is credited with compiling a short "list of Xhosa numerals, nouns, adjectives and verbs", differing little in form from the words as used today (Pahl 1989: xxxvii).With the exception of the two volumes of the GDX, Volume 1 and Volume 2, as well as the ISX, all the other listed dictionaries were published before the establishment of the XNLU.The establishment of the Unit with a specific line function, therefore, follows a very long history of lexicographic practice in isiXhosa.However, despite a relatively long list of lexicographic works for an African language, there is a dearth of literature and research on isiXhosa lexicography.As Google search shows, only three hits for both isiXhosa lexicography and Xhosa dictionaries are academic works that specifically deal with isiXhosa dictionaries.The remainder of the hits only serve to confirm the existence of the dictionaries.Of the previous twenty-two volumes of Lexikos, the journal of AFRILEX, which has served to disseminate information about African lexicography, only five make reference to isiXhosa lexicography.Of easily accessible works on isiXhosa lexicography, Pahl (1989) and Mtuze (1992) offer the most comprehensive accounts.In Section 6 of the introduction of the third volume of the GDX, Pahl (1989: xxxvi-xxxix) offers a historical account of isiXhosa lexicography in order to provide the GDX with the necessary context.He benefits largely from Rev. Robert Godfrey who, in the first edition of the KED, devoted some space to credit his predecessors and review their works.Mtuze (1992) adopts a similar approach in his survey of isiXhosa lexicography between 1972 and 1989, when the third volume of the GDX was published.Another publication worthy of mention because of its specific focus on isiXhosa lexicography is Moropa and Kruger (2000), which deals with the mistranslation of cultural terms in the KED.
The history of lexicography in general, or the historical perspective of dictionary research, in terms of Hartmann (2001), cannot be overemphasised.According to Wiegand (1984), it is important to study the history of lexicography in order to understand and learn from previous experiences and avoid repeating earlier mistakes.This point is also expressed by Hartmann (1983), who observes that every dictionary has a forerunner, as well as Landau (1984), who argues that dictionary making is a judicious act of piracy whereby lexicographers heavily rely on their predecessors.By revisiting the history of isiXhosa lexicography, the authors hope to chart the way forward for future works in this and other South African languages.This is particularly important at a time when both the NLUs and the Pan South African Language Board (PanSALB) are in need of a turn-around strategy for South African lexicography, more than ten years since the NLUs were established.
That isiXhosa lexicography is under-researched is unfitting, given such a long history of dictionary making.Not only lexicographers in the indigenous South African languages can learn from this history, but also lexicographers from other languages across the continent.However, for this to be possible, research on isiXhosa lexicography is still a long way behind.A student or researcher of isiXhosa lexicography still has to rely on primary sources such as the dictionaries themselves and some relevant documents which are in the possession of the XNLU, PanSALB and dictionary publishers.Furthermore, the researcher of isiXhosa lexicography still has to benefit from research on the lexicography of the other languages in streamlining research topics.To really make progress, the researcher may have to start from very general surveys before venturing into specific and detailed engagements with certain lexicographic issues.What this article hopes to achieve is to open up isiXhosa lexicography for metalexicographic scrutiny by identifying, albeit in less detail, some key areas for research.

A Kafir-English Dictionary
The KED was first published under the editorship of Rev. Albert Kropf in 1899 before its second, and so far, the latest edition, edited by Rev. Robert Godfrey, in 1915.According to its original editor (Kropf 1899: iii), it was first proposed as "a memorial of the Kafir language which would soon be supplanted by English".In terms of typology, the KED is a mono-directional bilingual dictionary which describes isiXhosa using English.Beyond the fact that the language http://lexikos.journals.ac.za lacked the relevant metalanguage for lexicographic description, the use of English as descriptive language is consistent with the view harboured by some missionaries when the dictionary project was first proposed.The language was being archived for the benefit of non-speakers of isiXhosa who had to study the language in order to understand the mother-tongue speakers (cf.Gallardo 1980).Contrary to the original view among some missionaries, the KED had an unintended outcome.It has become an important element of the lexicographic inventory of the language, a real milestone in the intellectualisation of isiXhosa.Pahl (1989: xxxviii) regards Albert Kropf, as "the first lexicographer to have fathomed and thoroughly mastered the intricacies of Xhosa phonology ... so thorough and accurate that little fault can be found with it".He further suggests that "all future lexicographical work in isiXhosa must be based upon it".The usefulness of this dictionary has transcended the isiXhosa-speaking community of South Africa and is one of the South African dictionaries, the others being in Zulu, that have been used in Zimbabwe by Ndebele speakers and scholars (cf.Hadebe 2006).It has also transcended the period during which it was produced, with Pahl (1989: xxxviii) rightly observing that it "has stood the test of more than three quarters of a century".A number of isiXhosa academics and language practitioners continue to use it as an invaluable reference for different purposes, such as translation.Because of its enduring importance, the KED has been converted into a PDF document, thereby turning it into an electronic resource to increase its accessibility, portability and durability many years after it went out of print.
The article below is an example of the fairly comprehensive treatment of lemmata in the KED.As can be seen from this article, type of speech data, meaning, both literal and figurative, as well as usage examples are provided for the verb ukutwala (to carry).

Example of a Dictionary Article from the KED
http://lexikos.journals.ac.za The KED also has thirty-two pages of front-matter texts in which, among other issues, it describes the grammatical features of isiXhosa and enumerates their treatment in the dictionary.The dictionary, therefore, fits in with the earliest grammars of isiXhosa, but there would be need to improve or revise some of its grammatical representations.
Perhaps its major impediment in post-apartheid South Africa is the K word in its title, the main feature of many other isiXhosa and isiZulu dictionaries published during the colonial and apartheid periods (cf.Table 1).This word is offensive, its presence in the title remains a major put-off to some potential users of the dictionary today.Besides the title, the dictionary has also been criticised for the mistranslation of certain culture-specific terms (cf.Moropa and Kruger 2000).Associating a traditional marriage practice with abduction in the above KED article is a good example.In modern times, abduction would be associated with a criminal activity and this limits the dictionary in terms of its communicative and cognitive functions.Finally, with its second edition having been published in 1915, the old orthography becomes another impediment of the dictionary.This can be illustrated by the KED article where uku-twala (to carry) is entered instead of ukuthwala and other entries such as tabatani (take) instead of thabathani in the same article.Thus, notwithstanding its usefulness, modern lexicographers and language users will need to pay attention to such shortcomings of this otherwise timeless work of isiXhosa lexicography.

2.2
The Oxford English-Xhosa Dictionary Based on the Oxford Learner's Dictionary, the OEXD is a mono-directional bilingual dictionary with as its primary purpose, the learning of isiXhosa as an additional language.In its preface, Fischer (1985) states "the idea of this English-Xhosa Dictionary was conceived during many years of studying and teaching Xhosa, and being confronted with the lack of an appropriate English-Xhosa Dictionary".The users are expected to know and use English as the point of departure when learning isiXhosa.Thus, English is the lemmatising language for which data categories meant to support mainly text production in isiXhosa are provided.The following is a typical article from this dictionary:

Example of a Dictionary Article from the OEXD
http://lexikos.journals.ac.zaHowever, Fischer (1985) reports that the discussion of the original idea of this dictionary with teachers indicated "a long-felt need amongst Xhosa-speaking students as well", particularly when "confronted with essays and literature in the English curriculum".This resulted in the inclusion of literary English words which are explained in isiXhosa.Pahl (1989: xxxix) describes this dictionary as "a godsend to students of Xhosa" as it was produced in the then new orthography, compared for instance to the KED.Furthermore, the enriched dictionary articles also facilitate the comprehension of lemmata using isiXhosa, meaning that the dictionary also provides limited English-comprehension support to isiXhosa mother-tongue speakers.Finally, its size also makes the dictionary handy for Xhosa-English or English-Xhosa translators, interpreters and editors.
With the latest reprint of 2011 being the 21st impression of the OEXD, this dictionary is without doubt the most well-known and successful isiXhosa dictionary on the market.For example, in both 2012 and 2013, an annual prelearning questionnaire survey that is conducted at the beginning of the undergraduate lexicography course at Rhodes University indicates that isiXhosa dictionary is synonymous with the OEXD.This knowledge is based on the students' previous but limited use of the dictionary at high school and its recommendation as an important reference work for isiXhosa students at the university, especially for translation studies.Despite this success, a close study of this dictionary is recommended to determine whether or not future reprinting should be accompanied by revision of certain areas.From a cognitive perspective, the dictionary is limited, like many other dictionaries in isiXhosa, as it displays cultural prejudices where the verb abduct is explained as forcefully dragging a girl into a marriage (cf.OEXD: 1).Yet abduct generally means much more than this.

2.3
The Greater Dictionary of IsiXhosa The Greater Dictionary of isiXhosa (GDX) is a three-volume trilingual isiXhosa dictionary in which English and Afrikaans are the other featured languages.Its volumes were published separately and, curiously, in reverse order, i.e.Volume 3 (Q-Z), Volume 2 (K-P) and Volume 1 (A-J) in 1989, 2003 and 2006 respectively.Although the GDX was completed in the NLUs dispensation, it is a product of the Xhosa Dictionary Project (XDP) which started in 1968.Having co-opted the XDP initiative into the NLU initiative which had a particular line function, the XNLU was allowed to complete this monumental lexicographic endeavour.The project was conceived on the backdrop of an acute need for "a modern, scientific standard Xhosa dictionary" (Pahl 1989: xxxix), the revision of isiXhosa orthography and a period of more than a century since the KED was out of print.Following the initiative of the rector of the University of Fort Hare and the endorsement by its Council, the project was to be housed at the University of Fort Hare.According to Pahl (1989: xxxix), "the Fort Hare University Council had decided that the Xhosa entries were to be defined in English and Afrikaans, but at a meeting of the Xhosa Dictionary Committee in 1968 it was decided that the definitions be given in Xhosa as well, thus making the dictionary fully trilingual".As a result, the GDX would appear as illustrated below: Example of a Dictionary Article from the GDX IsiXhosa is the lemmatising language but all three languages are used to provide explanations and other forms of lexicographic description.Because the dictionary also uses isiXhosa to provide information, it caters for the needs of both mother-tongue and non-mother-tongue speakers of isiXhosa.This was a major break from the tradition whereby all information about the language was hitherto given through another language, implying that those who did not know such a language, i.e.English or Afrikaans, could not benefit.
Besides the main text, the dictionary provides encyclopaedic information about isiXhosa and some culturally significant aspects of the language and its speakers.For example, between pages 685 and 754 of the third volume, the dictionary contains back matter texts which provide information on isiXhosa grammar and aspects of folklore in isiXhosa, e.g.tikoloshe, mamlambo, ichanti, impundulu, etc. as well as cultural practices such as circumcision, ukuthwalwa kwentombi, etc.As a container of knowledge (McArthur 1986), this dictionary provides information that non-mother-tongue speakers of isiXhosa and the younger generation of isiXhosa speakers may find very valuable.Thus the dictionary is rich in both linguistic and extra-linguistic (encyclopaedic) information.
However, despite its stature and richness, this dictionary has probably made more impact on the language itself than on the speakers of the language.Hence the distinction made between dictionaries for the languages and dictionaries for the language speakers and learners.In other words, it has not proved to be a popular reference work, especially for students at both university and school levels.This has less to do with the quality of the dictionary, but more probably with its voluminous nature.Because of the amount of resources that were required to produce it, the dictionary is being sold at R425 per volume, which adds up to R1 275 for the three volumes.In addition to its price, the three-volume dictionary is not a portable reference work.As a result, over 4 300 of the 4 500 copies (1 500 per volume) are still with the publisher instead of dictionary users, which indicates that the dictionary has not sold very well.It represents archival material for isiXhosa, but is not a utility product as a dictionary should be.
Firstly let me acknowledge PanSALB for supporting us as the IsiXhosa National Lexicography Unit for all these years.I appreciate the view that we should start to compile monolingual dictionaries.Yet PanSALB allowed our Unit, with the advice of the Board of Directors, to first complete the project of compiling the trilingual dictionary, namely The Greater Dictionary of IsiXhosa, which began in 1968.
This dictionary, therefore, may be regarded as a major step towards the accomplishment of the line function of the NLUs, i.e. to compile national dictionaries for the official languages (cf.Section 3).As a standard dictionary, the main components of the articles in the ISX are the lemma, word-class data, noun http://lexikos.journals.ac.za classes for noun lemmata, explanations of meaning and illustrative examples, as can be seen from the example below.The dictionary also provides some data types characteristic of national dictionaries as linguistic and cultural reservoirs.
For example, limited etymological data is given in the form of the source languages of borrowed words treated as lemmata, with abbreviations such as Afr (Afrikaans), Eng (English), Jap (Japanese) and SAE (South African English).Isihlonipho (euphemistic language of women), described in the front-matter, is also a key element of the cultural data that is included in the dictionary, but it is clear that its provision and treatment is even less comprehensive and less satisfactory compared to the GDX.Thus, there is still some way to go to accomplish the line function of the NLUs, for the national dictionary will have to be more comprehensive (cf.Section 3).
(They (Xhosa speakers) are correct to stop developing other 'nations' using their language -it has been long since isiXhosa was treated as something that had to be learnt by other 'nations' in order to understand the mindset of "these Xhosas".As a result, this book that is being discussed will give dignity and confidence to the speakers of the language, that their language can also serve functions that have been served by other languages of the world.Surely, even the terminology of technology is also being defined using isiXhosa in the same way as English.) http://lexikos.journals.ac.za Clearly, there is resonance to the sentiments expressed by Béjoint (2000) that monolingual general-purpose dictionaries are national emblems and expressions just like national flags and national anthems, a view that informed Samuel Johnson's approach to the mission of compiling The Dictionary of the English Language (cf.Johnson 1747).It is clear, therefore, that the future of the language and its ability to fulfil the ideals that saw isiXhosa being accorded its status as one of the country's official languages is felt in the dictionary.The mandate that PanSALB placed on the NLUs, therefore, is conceived in such a nationalistic spirit.However, it is critical to consider this mandate in the light of language planning in South Africa, of which lexicography is part, and the lexicographic needs of the speakers and learners of the respective languages.

Towards the National Dictionary: Current Dictionary Projects in isi-Xhosa
After this brief review of isiXhosa dictionaries, the future of isiXhosa lexicography is now considered in the light of current dictionary projects.Perhaps the most crucial question is whether the XNLU is on course to accomplish its line function of producing a national dictionary for isiXhosa.Again, it would be important to discuss the reason for the current state and possible direction for isiXhosa lexicography.
As noted in the previous section, lexicographic practice in isiXhosa predates the establishment of the NLUs in South Africa by over two centuries.Through the XDP, an organised structure for lexicographic work already existed in the language.This structure was considered ideal to be co-opted into the new lexicographic dispensation.Yet it was equally vital that the line function of the NLUs had to be co-opted into an established tradition for the language to develop and for both its speakers and learners to obtain dependable reference tools.
A closer look at the dictionaries produced and some current projects by a number of NLUs suggests that there is no consensus with regard to the line function of the NLUs and how this function could be accomplished.Consequently, the optimism that accompanied the establishment of the NLUs has turned into frustration as their mandate seems to be difficult to execute.Whereas different NLUs are engaged in projects which seem to be distanced from their line function, the NLUs do not find it difficult to justify their projects in view of the lexicographic needs of the linguistic communities they seek to serve and the developmental stages of the respective languages.The XNLU is no exception.
The terms of reference for the establishment of the NLUs were first outlined in the National Lexicography Units Bill (B 103-96), which is discussed in detail by Kumalo (1999), and further encapsulated in the Articles of Association of the NLUs.According to the Bill, the objectives of the NLUs "shall be to initiate, maintain, continue, complete and from time to time improve the compila-tion of the dictionary and other products" (B 103-96: 4).The definite article the underlines the line function of the NLUs as focusing on a particular type of dictionary.The definitions section of the Bill describes this dictionary as "the definitive product, on the general vocabulary of an official South African language, as produced by the unit for the language concerned" (B 103-96: 3).This dictionary is the national monolingual general dictionary (Alberts 2011), or, simply put, the national dictionary.The national dictionary is defined in Hartmann and James (1998: 98) as "[a] cultural or historical dictionary of significance to the language of a country, often financed by central government or an academy and compiled by a team of lexicographers".To the best of our knowledge, none of the NLUs for the nine indigenous South African languages are currently working on such a dictionary.Instead they are busy with various types of dictionaries which, in terms of the Bill would be accounted for as the "other products" (B 103-96: 4).It is perhaps due to the failure of the NLUs to make the anticipated progress concerning their line function, and the uncertainty that they can realistically achieve this feat in the foreseeable future, that Alberts (2011: 51) argues: The NLUs need to be given the resources that are necessary to revert to the original idea of producing properly researched, definitive, monolingual dictionaries for each of the official languages instead of compiling all kinds of dictionary products to generate funds (dictionaries that commercial publishers are able to compile much better, and that they are already publishing successfully).
The XNLU is currently working on three projects, namely, the Mathematics and Science Pedagogic Dictionary, the Concise Trilingual Dictionary and the IsiXhosa Monolingual Dictionary.A brief discussion of these projects poses a number of questions in relation to the progress of isiXhosa lexicography and the line function of the NLUs.
The product of the first project is a specialised dictionary intended for teachers and learners in the intermediate and senior phase of primary schooling.It represents the Unit's endeavour to promote mother-tongue education through a dictionary that facilitates access to, and communication in, mathematical and scientific content by learners and teachers in isiXhosa.The project, therefore, addresses communication and cognitive problems in the respective subject fields of Mathematics and Science by describing specialised concepts, documenting, explaining and illustrating the use of the relevant terminology using isiXhosa which is one of the main languages of the Eastern and Western Cape Provinces.The first draft of this dictionary has been reviewed for publication and the Unit is currently incorporating the suggestions of the reviewers who included subject specialists (Mathematics and Science), a mother-tongue isiXhosa linguist and a lexicographer.
The second project is the Concise Trilingual Dictionary Project.Its main intention is to abridge the bulky GDX into a concise, portable and user-friendly dictionary that can be sold at an affordable price and be easily used by school http://lexikos.journals.ac.za and university students, both mother-tongue isiXhosa speakers and nonmother tongue speakers whose mother tongue is either Afrikaans or English.The GDX is, thus, the dictionary basis for the current project.The Unit contracted a Cape Town-based company to convert the hard copy of the GDX into an electronic rich text format (RTF) document.GDX entries were then copied into a Word document which can be edited, with editing focusing on abridging the GDX by eliminating certain types of entries while adding others, e.g.curriculum words collected from the Department of Education.The Unit intends to make use of TshwaneLex when the actual dictionary writing stage commences after the establishment of the dictionary basis and dictionary files.
The last of the current projects, i.e. the IsiXhosa Monolingual Dictionary Project, also emanates from the GDX.Its rationale is based on the size of the GDX, but unlike the concise trilingual dictionary project, it eliminates all the Afrikaans and English entries to result in a monolingual dictionary.Again, not only eliminating the entries in the other languages is the main focus, but the intended dictionary will also be updated by adding baseline curriculum terms from the Eastern and Western Cape Departments of Education and other new words.The target users of the dictionary are school and university mother tongue students of isiXhosa.The envisaged dictionary should be an updated and a more comprehensive version of the ISX.
Although only brief insights have been provided with regard to the current XNLU projects, it is clear that the Unit does not currently have the national monolingual general isiXhosa dictionary as one of its projects.Looking backwards and forwards, the following questions may therefore be posed concerning the current dictionary projects: -How different is the mathematical section of the Mathematics and Science Pedagogic Dictionary from Isichazi-magama Sezibalo published by Cambridge University Press in 2010?
-How different is the scientific section of the Mathematics and Science Pedagogic Dictionary from Isichazi-magama sezeNzululwazi neTeknoloji, a Science and Technology Dictionary published by New Africa Books?
-Is size the only real problem with the GDX such that the Concise Trilingual Dictionary Project will result in a user-friendly dictionary?-Will one concise version of the GDX be suitable for use by school (taking into account the different school levels) and university students?
-Will the IsiXhosa Monolingual Dictionary Project produce a dictionary that would be fundamentally different from the ISX that was published in 2008?
-Overall, how do the current dictionary projects advance isiXhosa lexicography cognisant of the line function of the XNLU?
These questions remain crucial for the XNLU to produce new dictionaries that http://lexikos.journals.ac.za will not only be snippets and combos of previous dictionaries, but tools that will contribute towards the intellectualisation of isiXhosa.No concerted attempt will be made to address each of the questions here, for equally elaborate discussions pertaining to the individual dictionary projects would be required to address these questions.Such discussions will constitute the research-based lexicographic practice that the XNLU aspires to adopt as it moves forward.
Here it will suffice to focus on the last two questions in as much as they resonate with the argument of Alberts (2011) cited above.
There is no disputing that isiXhosa, like all the other indigenous African languages, needs a comprehensive monolingual general dictionary in the mould of the Woordeboek van die Afrikaanse Taal in Afrikaans and other such dictionaries in more developed languages.Obviously, the NLUs will need more resources to conduct thorough research towards such dictionaries.The current budget for the NLUs, which is not even enough for the salaries of the staff, is a clear under-investment in this area.However, channelling all the available resources towards such a dictionary may not be ideal for the indigenous African languages in which other small dictionaries are needed to assert the official functions that the democratic constitution bestowed on these languages (cf.Mini 1996).Expecting commercial publishers to fill this big void by producing such dictionaries may also be asking for too much in the current situation where the implementation of the multilingual policy has been half-hearted.The success with which the commercial publishers have been producing other dictionaries for all the indigenous languages remains limited and questionable but understandable.Since commercial publishers are profit-driven, they will always be reluctant to produce books in African languages, unless the government commits funds to subsidise their costs.Strategies are needed that will, for instance, enable the NLUs to focus on the national monolingual general dictionary projects while producing lexicographic spin-offs in the form of other dictionaries that can make an immediate impact in the education sector, for instance.In any case, the material that is being used for current projects may also be employed for the comprehensive national monolingual dictionary and the XNLU may be considered to be on the right track, especially with the current IsiXhosa Monolingual Dictionary Project.

Other Issues for IsiXhosa Lexicography Moving into the Future
There are other crucial issues that should be considered for isiXhosa lexicography to move forward into the future.Some of these are discussed in the subsections that follow below.

Research-Based Lexicography
The establishment of the NLUs emerged as a response to the democratic language policy that sought to elevate South Africa's indigenous languages http://lexikos.journals.ac.za (Gouws 2003) and their successful establishment may be attributable to this enabling environment.However, the major role played by lexicographic research should never be underestimated.It was crucial for the lexicographically informed to make research-based submissions and presentations to the policymakers, and lexicographers such as Professors Danie Prinsloo and Rufus Gouws, as well as Dr Mariëtta Alberts, deserve special mention.It was also these lexicographers who organised and ran workshops on various lexicographic issues before and after the establishment of the NLUs, thereby ensuring that dictionary work would not be conducted by unskilled people (cf.Gouws 2001).
During the early years of the NLUs, the lexicographers working on different languages would also present papers at AFRILEX and other conferences on the challenges encountered in their work and the procedures they adopted to address such challenges.This provided an opportunity where lexicographic practice and theory would interact and enrich each other.Such enthusiasm later diminished.However, the Dictionary Roundtable session at the Eighteenth International AFRILEX Conference, held at the Nelson Mandela Metropolitan University, brought a revival of interest as the NLUs presented status and progress reports on their projects, in addition to showcasing some of their products.It became clear that despite some concerns with their achievements, they were, however, striving to contribute towards the development of their respective official languages, albeit under difficult circumstances.
Moving forward, the XNLU intends to engage in research-based lexicography for a number of reasons.Firstly, with the user-perspective having gained prominence in modern-day lexicography (Hartmann and James 1998), the Unit needs to be fully aware of the user needs and reference skills.In this way, the lexicographers will be able to conceive dictionary projects that will be of functional value.Secondly, the XNLU should open itself for metalexicographic research.This may entail publicising its products and projects in order to generate feedback and ideas from dictionary users and metalexicographers that may improve their products.In this way, they may also learn from the experience of their counterparts working on other languages locally and abroad.Finally, the concept of research-based lexicography will involve the lexicographers from the Unit in investigating language-specific issues, using linguistic, educational and lexicographic theories to arrive at well-informed decisions regarding their practices.Research-based lexicographic practice will not only improve the quality of isiXhosa dictionaries, it will also contribute to the available body of knowledge about South African lexicography, which is currently lacking when it comes to certain languages.It will contribute towards developing and entrenching a beneficial dictionary culture and an informed dictionary criticism.

Developing a Dictionary Culture
The practical work of compiling dictionaries at the XNLU needs to be accom-http://lexikos.journals.ac.za panied by the development of a dictionary culture among the potential dictionary users.Whereas lexicographic practice on its own will contribute towards the development of the language, developing a dictionary culture will go a long way in empowering the isiXhosa speech community through the products of that lexicographic practice.As Gouws (1996: 97) argues, dictionary-making is a cultural practice that should lead to another cultural practice, i.e. dictionary use.If the NLUs continue to produce dictionaries that nobody uses, then it would be difficult for them to get feedback that will improve their practice.Above all, they would have laboured in vain.The little resources that are allocated to the NLUs would turn out to be a waste.Yet in the South African context, it was mainly the need to develop the languages so that they could be used for expanded social spaces, that placed emphasis on the dictionaries (Gouws 2003) and other corpus development outputs.Language users and language learners should venture into those spaces that were previously reserved for English with, among other resources, dictionaries as their communication and learning instruments.Not only do they need dictionaries, they also need the reference skills that will enable them to gain optimally from using dictionaries.At the moment, isiXhosa dictionaries are generally not well-known.Dictionaries are still less used at schools and students proceed to university with little or no knowledge of the available dictionaries especially in the African languages.The XNLU has a part to play in eradicating this problem.They need to go to schools to promote their dictionaries and demonstrate their importance in the real situations of the potential users.A project is in the pipeline in this regard (cf.Section 4.5).

Lemmatisation in isiXhosa Dictionaries
As in Bantu languages in general (Gouws and Prinsloo 2005a;2005b), lemmatisation remains another critical issue that should be given attention by both practising and theoretical lexicographers in isiXhosa.This pertains to the choice between stem-based and word-based lemmatisation approaches in entering and ordering lemmata.The last time that the issue seems to have been given serious thought was when, in the second edition of A Kafir-English Dictionary, Godfrey (1915: ix) stated: The nature of the structure of the prefix-using Kafir language, differing absolutely from that of the suffix-using English, renders the construction of a Kafir dictionary on English lines an impracticability.... Were we to follow the alphabetical order of Kafir words, as we do in English, we should find the great bulk of the dictionary entered under the vowels i and u and we should have very little use for any other letter.For this reason Kafir scholars have practically agreed that a Kafir dictionary should follow the alphabetical order of stems rather than words.Kropf's dictionary was constructed on this plan.
Since then, all dictionaries in which isiXhosa is the lemmatising language have http://lexikos.journals.ac.za followed suit.The advantages of both approaches have been discussed before (cf.Gouws andPrinsloo 2005a, 2005b).On the stem-based approach, Godfrey (1915: ix) cautiously concluded: Although the greatest care has been exercised to discover the stems and to place under them only such words as are actually derivatives, further study of the language will demand a certain amount of re-arrangement.
It is such a further study that is advised in this article to determine whether the stem-based approach should continue to be adopted for all dictionaries in which isiXhosa is the lemmatising language, especially at a time when a shift has started in its sister languages such as isiZulu.Gouws and Prinsloo (2005b) observe that if applied wisely in the form of left-expanded article structures, the approach offers the best of both approaches.However, apart from further research that can ascertain user competences when it comes to dissecting words into prefixes and stems, this article also advises that in dictionaries targeted at young users and those dealing with specialised languages, where providing information regarding the structure of the language is not important, a word-based approach should be adopted.

Electronic Lexicography
Computer assistance was identified as a potentially important element in the formulation of the general theory of lexicography (Wiegand 1984).There are several ways in which computer technology has since transformed and improved practical lexicography (cf. De Schryver 2003, Sinclair 1987).This includes the use of electronic corpora for a variety of lexicographic tasks and dictionary writing systems which permit simultaneous participation of different lexicographers in one project even from different workstations.The XNLU and other NLUs need to exploit the opportunities of electronic lexicography in their practice.The national monolingual dictionary projects would benefit much from the use of corpora which would store large masses of linguistic data and display the various linguistic nuances that such dictionaries should capture.Electronic databases may also enable the lexicographers to enter all the data that will eventually be included in the national monolingual dictionary project but suppress it for the purposes of deriving smaller dictionaries from time to time.Even more, some dictionaries, including the national monolingual ones, may be published as electronic dictionaries to avoid publishing costs and render the dictionaries organic in the sense of being continuously updateable without devaluing previous publications.The lexicographers working for the NLUs need to be equipped with the skills that will ensure that their computers are not nice-to-have gadgets or mere versions of type-writers but tools that will enable them to store data, process it and address language-specific issues such as lemmatising and defining in Bantu languages.

Co-operative Lexicography
Given that the NLUs constitute the core of South African lexicography and serve as a testimony of the government's commitment towards lexicography, it should always be borne in mind that they are a vital part of the official language planning efforts in the country.The concept of co-operative lexicography was proposed from the outset of the establishment of the NLUs (Kumalo 1999: 214) and came up again in their recent review (Alberts 2011: 35-37).Co-operative lexicography should be part of strategic operations which can maximise the limited resources that the NLUs have to work with while at the same time maximising on positive impacts of lexicographic projects and products on the development, learning and use of all official South African languages.A few cases of co-operative lexicography are under way or in the pipeline for isiXhosa lexicography undertaken at the XNLU.Firstly, co-operative lexicography is essential in the XNLU's bi-or multilingual dictionary projects, given that its main focus and linguistic expertise is mainly in one language.The Unit would have to collaborate with its peers working on the respective languages covered in particular projects.So far, the Dictionary Unit for South African English (DSAE) has supplied the XNLU with its English wordlist that it compiled for its own projects and this will consolidate the dimension of the latter's projects covering English.The Bureau of the WAT may, for instance, also have to be consulted for assistance with the Afrikaans section in the Trilingual Concise Dictionary Project.In this way, it may be possible to optimise the assistance that different sets of language speakers and learners may obtain from one dictionary.
Secondly, the XNLU needs to strengthen collaboration with other stakeholders such as the Department of Education, schools and universities whose key activities constitute dictionary user situations and determine lexicographic needs.Such stakeholders need to be involved as advisors and participants in some lexicographic projects, and most crucially as potential users, future compilers and critics of prospective products.In this way, the Unit will be able to create public awareness of its activities, develop a dictionary culture among the potential dictionary users and generate constructive feedback on its products.That the University of Fort Hare library did not have copies of the GDX until 2011 is an illustration of how weak collaboration has been between the NLU and its host institution, a case of a poor tenant and landlord relationship.Currently the Unit, working with the African Language Studies Section of Rhodes University's School of Languages, have put together a joint-community engagement plan through which dictionary skills workshops will be conducted in Eastern Cape schools.
Lastly, the XNLU and all the NLUs need to find ways of working closely with commercial publishers.There is much that can be gained by co-operating rather than competing with commercial publishers who have much experience and resources required for successful lexicography.The DSAE and WAT collaboration with publishers such as Longman and Pharos have proved to be http://lexikos.journals.ac.za effective in facilitating the production of dictionaries that fall outside the line function of the NLUs but serve important immediate functions, especially in education.Hall et al. (2013) utilised the publishers slot in the Dictionary Roundtable at the Eighteenth International AFRILEX Conference to highlight some key considerations for the NLUs seeking publication deals with commercial publishers.In order for the NLUs to find some common grounds with publishers, the importance of a business plan would be reiterated, whereby consultations are made at the dictionary conceptualisation stage instead of a situation whereby the NLUs offer finished dictionary projects to different publishers.Finding themselves stranded with their completed projects and resorting to less well-known and less strategic publishers has not helped some NLUs.The main reason, in our view, why the GDX and ISX are not available in bookshops and less well-known among the isiXhosa-speaking community is working with a publisher outside the isiXhosa-speaking provinces, thereby making the marketing of the dictionaries quite difficult.If the GDX had been published by OUP, for instance, it is unthinkable that it would be difficult to acquire it in bookshops.

A Review of the Current Structures of South African Lexicography
While there has been a concern that the NLUs have not lived up to their expectations (Alberts 2011), the PanSALB Session at the Eighteenth International AFRILEX Conference confirmed that the current structures of South African lexicography, the core of which is PanSALB and the NLUs, have struggled to function.Resource constraints, which have implications on the staffing of the NLUs and the scope of lexicographic work that the NLUs could realistically undertake, emerged as a major issue.The PanSALB presentation also fell short in that, having provided the now well-known background to the establishment of both PanSALB and the NLUs, it had a blank slide when it came to the Board's vision for the NLUs.That the NLUs have collectively made a submission of revisions of the Articles of Association demonstrate the strain that the lexicographers have taken not only to produce dictionaries but also to survive under the current situation.The XNLU, like the other NLUs, faces an uncertain future and since it has been the main role player in isiXhosa lexicography, it is also the future of isiXhosa lexicography that is at stake.However, one has to be optimistic that, rather than judging South African lexicography to have failed, the challenges faced by NLUs compel the stakeholders to go back to the drawing board and find new solutions that will ensure that South Africa maintains its place on the global lexicographic map.

Conclusion
This article dealt with the past, present and future of isiXhosa lexicography.However, some issues discussed are reflective of the lexicography of African languages and South Africa in general.The article established that isiXhosa has a long and rich history of lexicographic practice.However, it noted that although there is much that can be learnt from that history, the existing dictionaries need to be either updated or supplemented by new dictionaries in order to meet current lexicographic needs among the isiXhosa-speaking community.Perhaps this explains the current projects being undertaken by the XNLU.This article does not proclaim to be a closed case.More definite conclusions could only be made with regard to the past and the present.We can only plan, but not always predict successfully what will happen in the future.Accordingly, the article only sought to initiate discussions about South African lexicography more than a decade since the establishment of the NLUs, which came after more than a century of lexicographic practice in African languages.Some potentially productive proposals have been made in this article.The article therefore hopes to have opened up possibilities for research in and debate on isiXhosa lexicography, with the major objective being to contribute to lexicographic practice and dictionary use.At the same time, it highlighted that South African lexicography is at a cross-roads, e.g. the choice between a national dictionary and other projects.As such, there is an urgent need to plan anew to ensure that the optimism of the past decade does not vanish into a false promise.

Table 1 :
Sparrman (1776)pended to his book entitled A Voyage to the Cape of Good Hope Towards the Antarctic Polar Circle and Round the World but Chiefly into the Country of the Hottentots and Caffres from the Year 1772 to 1776, indicated as Appendix toSparrman (1776)in Table1.A partial inventory of isiXhosa dictionaries, their editors and dates of publication