An investigation of partial meaning and non-equivalence in English translations of Ekegusii kinship terms

The main challenge that translators face is that of non-equivalence as the source language (SL) and the target language (TL) almost always have cultural and linguistic differences (organize words differently). Accordingly, one needs to be aware of the linguistic differences between the two cultures. A group uses a set of terms to name the people they are related to by blood, marriage or adoption which should be recognized by members of the speech community. This study explores kinship terms in Ekegusii, a language that has been coded as E42 in Guthrie's (1971) classification of Bantu languages and belongs to the interlacustrine zones, namely D and E. Both Ekegusii and English languages have their unique kinship term systems which pose translation problems with regard to equivalence. Unlike English nouns that denote family relations in a pretty straightforward way, Ekegusii’s way of addressing family members and relatives differs and to some degree it may be difficult to those who are not familiar with the Ekegusii culture. Consequently, this paper investigates the problem of non-equivalence at word level in translation between Ekegusii and English. The paper presents background knowledge and different approaches related to non-equivalence. Subsequently, conceptual kinship terminologies are contrasted to prove that there is a significant linguistic gap between Ekegusii and English. The paper then proposes a classification of non-equivalence at word level based on Mona Baker’s point of view. Finally, the paper suggests effective strategies to deal with non-equivalence at word level in translation of Ekegusii kinship terms.

The native speakers of Ekegusii are known as Abagusii. Ekegusii is used as a first language by approximately three million speakers in the three larger administrative districts of Nyanza Province namely: Kisii, Nyamira and Gucha, collectively known as Gusii (Ogechi, 2002).
All human languages have a system of kinship terms which is clearly highlighted in their address systems. Without exception, all kinship term systems make use of such factors as sex, age, generation, blood and marriage in their society (Wen and Min, 2007). Kinship terminologies are thus the terms of address used in different languages or communities for different relatives and the terms of reference used to identify the relationship of these relatives to ego. Kinship terms consist of a category of words by means of which an individual is taught to recognize the significant groupings in the social structure into which he is born (Leach, 1958).
Different societies classify kinship relations differently and therefore use different systems of kinship terminology. For example, some languages distinguish between affinal and consanguine uncles, whereas others have only one word to refer to both a father and his brothers.
All human beings are connected to others by blood or marriage. Connections between people that are traced by blood are known as consanguineous relationships while relationships based upon marriage or cohabitation between collaterals (people treated as the same generation) are affinal relationships (Essig, 2011).
Supposedly, these kinds are the same for all individuals in all cultures. However, people in different societies customarily regard genealogical connections differently, recognizing some kinds of relationships and ignoring others. The culturally determined genealogies turn objective relationships of blood and marriage between people into kinship so there is no culture in which all genealogical relationships are recognized as kin relations. Consequently, all the genealogical ties that a culture chooses to recognize are presumably what constitute an individual's kin (Essig, 2011).
According to Henning (1995), kinship terms have been widely analyzed across languages which often leads to making quite different distinctions. Moreover, English kinship terms can be broken down into generation, sex and lineage, but to translate terms from other languages we will need to add additional semantic components, for instance, side of family, relative birth order and person.

GENERAL REMARKS ON TRANSLATION
Africa is a multilingual society, thus audiences greatly depend on translation as a means of communication. This implies that translation comes in to fill the gaps between languages.
Translation has typically been used to transfer written or spoken source language (SL) texts to equivalent written or spoken target language (TL) texts. Generally, the purpose of translation is to reproduce various kinds of texts in another language hence making them available to wider readers. Just like any other speech community, the Abagusii rely on translation in their communication because some members of audiences may not be speakers of the language. However, the translations to the target language may be inadequate as there may not be exact equivalents of the SL in it leading to misinterpretation hence the need for the study.
It is not easy to translate from SL to TL because language is not just a classification for a set of general or universal concepts (Ordudari, 2007). In this regard, Culler (1976) believes that languages are not nomenclatures and the concepts of one language may differ radically from those of another, since each language articulates or organizes the world differently, and languages do not simply name categories; they articulate their own. Accordingly, one of the troublesome S e p t e m b e r 1 1 , 2 0 1 5 problems of translation is the disparity among languages. The bigger the gap between the SL and the TL, the more difficult the transfer of message from the former to the latter will be (Newmark, 1991).
The difference between an SL and a TL and the variation in their cultures makes the process of translating a real challenge. This paper is going to concentrate mainly on the procedures of translating culture-specific concepts (CSCs) and on the strategies of rendering their equivalents.
Translation consists of studying the lexicon, grammatical structure, and cultural context of the source language text. In the process one analyses it in order to determine its meaning and then reconstructs the same meaning using the lexicon and grammatical structure which are appropriate in the target language text and its cultural context. Newmark (1991) posits that many people hesitate to define translation and as far as many dictionaries offer synonyms for the verb, for instance, render, rephrase, transform, transfer, convert and add 'from one language to another,' they do not explain what is being translated. However, scholars have defined translation variously, for example, Newmark (1991) defines translation as transferring the meaning of a stretch or a unit of language, the whole or part of a text, from one language to another. Catford (1965) points out that translation is the replacement of textual material in one language by equivalent textual material in another language. He also defines it as "an operation performed on languages: a process of substituting a text into another" (Catford, 1988, p. 21). This substitution often involves the translator either considering the form of the SL (formal equivalence) or considering the meaning of the SL (functional equivalence).
In regard to the above definitions, translation is therefore, converting one language to another so that the TL could convey the intended message in SL. The translator's role is thus to decode SL and encode his understanding of the TL form as well as reproducing the closest natural equivalent of the SL. Accordingly, functionally relevant meaning is transferred leaving out the superfluous features of meaning that can also be found in the text.
As much as the translating activity is complex and difficult to define, Newmark (1991) observes that a good translation is not difficult to identify because it looks slightly similar to the original text to a reader or listener competent in both languages, unless the original contains errors of fact and deficiencies of style. He further points out that the translation of any original work is bound to make in roads on and to benefit the norms of the TL. Hence, the invisible nature of the translator's activity will emphasize the differences between the SL and the TL.
As acknowledged by Newmark (1981), equivalence is an imperative procedure in translation which goes hand in hand with meaning. Equivalence occurs at word, grammatical, textual and pragmatic levels. Besides, most people tend to recognize a translation if they find many corresponding features between the target and source language texts. By examining some examples drawn from Ekegusii, one would discover that there exists no total meaning and equivalence in translation. This is because meaning belongs to language and culture, and, some languages use only one word to refer to a concept where other languages use two or more words to refer to the same concept. Additionally, some words in Ekegusii are polysemous thus causing ambiguity. This paper therefore, aims to explore and explain the translation strategies employed in translating kinship terms from Ekegusii to English.
The art of translation is irrefutably complex as it involves many different kinds of processes and also there are known problems associated with the use of translation texts. The study will examine language particular encoding of kinship terms and semantically related terms in each language. Nevertheless, this paper will assess equivalence and how it will affect the interpretation of the target text vis-a-vis the source text. These corpora will provide substantial data for a contrastive study because they involve translation of languages in different language families. Ekegusii is a Bantu language whereas English is a West Germanic language which implies that meaning is less likely to be preserved because of the disparity in context and culture.

TRANSLATION AND EQUIVALENCE
Translations are assumed to be somehow "the same" as the original, but they are perceptibly different. As a result, many translation scholars have put forth their views on equivalence. For instance, Halverson (1997) defines equivalence as a relationship existing between two entities, and the relationship is described as one of likeness or sameness or similarity or equality in terms of any of a number of potential qualities.
As pointed out by Zaky (2000), a translator must look for a target-language utterance that has an equivalent communicative function, regardless of its formal semblance to original utterance in so far as the formal structure is concerned. This implies that translation is geared towards conveying the textual effect of the original.
The first debate of the notion of equivalence in translation initiated further elaboration of the term by contemporary theorists. Therefore, the difficulty in defining equivalence, however, seems to result in the impossibility of having a universal approach to this notion. Thus many translation scholars have given their varied views on the concept of translation equivalence.
According to Leonardi (2000), the notion of equivalence is said to be undeniably one of the most problematic and controversial areas in the field of translation theory that has caused, and it seems quite plausible that it will continue to cause, heated debates within the field of translation studies. Moreover, the term has been analyzed, evaluated and extensively discussed from different points of view and has been approached from many different perspectives all in vain. S e p t e m b e r 1 1 , 2 0 1 5 Proponents of equivalence based theories of translation usually define equivalence as the relationship between a source text (ST) and a target text (TT) that allows the TL to be considered as a translation of the ST in the first place. As a result, equivalence relationships are also thought to hold between parts of ST and parts of TT (Halverson, 1997). Moreover, Pym (1992) has pointed to the circularity of translation equivalence by asserting that equivalence is supposed to define translation, and translation, in turn, defines equivalence.
Translating consists of reproducing the closest natural equivalent of the source language message in the target language both in terms of meaning and style (Nida and Taber, 1982). Therefore, it is concerned with moral and factual truth which can be effectively rendered only if it is grasped by the reader or listener as this is what is taken to be the end product of translation (Newmark, 1991). Thus the translators, by finding equivalence in translation can show the tentative nature of their assertions, invite the readers or listeners to join and decide which translation accurately renders the ideas, concepts and words of the original text.
Various scholars have presented varied views on equivalence, for example, Jakobson (1992) bluntly says, "Equivalence in translation is almost always only partial as ordinarily there is no full equivalence between code units" (p.45). He also explains the differences between structures, terminology, grammar and lexical forms of languages are the main reasons of non-equivalence. He further observes that equivalence the pivotal concern of Linguistics as it is the basic problem of language.
Moreover, Jose Ortega (1992) also supports the view by saying that each language community gives a mental representation to any terms it uses. Thus, it is utopian to believe that words of different languages would give us translations that refer to exactly the same objects or ideas (cited in Miyanda, 2007).
For Newmark (1988), the equivalent effect is a result which all translators long to achieve. Newmark further argues that the text may reach a direct equivalent only if it is 'universal' that means the cross cultures share common ideas which of course may not be realistic.
Wen and Min (2007) exemplify the differences between English and Chinese kinship terms and reveal the causes it results from. They however do not suggest the ways in which to deal with the non-equivalence. Ghadi (2009) has presented an interesting study analyzing equivalence at word level in the English technical text and the translations in Persian. He introduces Baker's non-equivalence at word level and it is evident that translating by a general term, the use of loan word and loan word plus explanation are the chief strategies applied by both the expert and nonexpert translators. Giménez (2005) evaluates student's translation from English to Spanish and notes that translation using a related word; translation by paraphrase using unrelated words and translation by omission seem to be favoured by Spanish native speakers.
In a nutshell, Mona Baker's categories of non-equivalence at word level and the strategies that address the problem have been widely used and recognized by many linguistic theorists and researchers. Its application is not restricted only in written text translation and that in a classroom set up but also translations in social gatherings because of its soundness and comprehensibility. As observed in fundraisings and funerals translators also share the disadvantage in dealing with non-equivalence, especially in translating Ekegusii kinship terms to English.
In most cases we could contemplate on whether one will get total meaning and equivalence or not in translating words from one language to another, for instance, Ekegusii to English. What will always linger in one's mind is whether an equivalent in the TL covers all the aspects of the corresponding term in the SL especially in the event of back translation since the languages are from different families.
A majority of translation scholars emphasize the role of equivalence in the process or product of translation directly or indirectly and as a result, it is in the center of the translation studies. Consequently, this paper explores the nature of nonequivalence, its nomenclature and strategies of tackling non-equivalence at word level.

THEORETICAL FRAMEWORK
The comparison of texts in different languages inevitably involves a theory of equivalence (Leonardi, 2000). As a result, equivalence can be said to be the central issue in translation although its definition, relevance, and applicability within the field of translation theory have caused heated controversy. In addition, many different theories of the concept of equivalence have been elaborated within this field in the past fifty years.
The theory of equivalence has been interpreted in various ways by some of the most renowned theorists in the field of translation who include: Vinay and Darbelnet, Jakobson, Nida and Taber, Catford, House, and finally Baker. These theorists have studied equivalence in relation to the translation process, using different approaches, and have provided abundant ideas. However, according to Leonardi (2000) an outstandingly remarkable discussion of the notion of equivalence is found in Baker (1992) who offers a more detailed list of conditions upon which the concept of equivalence can be defined. Accordingly, Baker (1992) explores the notion of equivalence in translation at different levels, as well as the different aspects of translation. She thus puts together the linguistic and the communicative approach. Consequently, she distinguishes between: equivalence at word level, equivalence above word level, grammatical equivalence, textual equivalence, and pragmatic equivalence. S e p t e m b e r 1 1 , 2 0 1 5 Baker (1992) acknowledges that equivalence at word level is the first component to be considered by the translator since we are concerned with communicating "the overall meaning of a stretch language" (Baker, 1992, p. 10). In actual fact, when a translator starts analyzing the ST he looks at the words as single units in order to find a direct equivalent term in the TL. At word level therefore, Baker (1992) defines the "word" and notes that a word can sometimes have different meanings in different languages. She goes forth to relate the meaning of words with the morpheme, an implication that the translator should pay attention to a number of factors such as number, gender and tense when considering a single word. Moreover, she introduces problems at word level and above word level, and later on suggests some strategies in dealing with them.
Of all the approaches to equivalence, Mona Baker's stands out in this paper as she focuses on equivalence at word level. Moreover, she claims that the word is the basic unit to be considered in the meaning of translation of a text. Baker (1992) gives a definition of the term "word" since a single word can sometimes be assigned different meanings in different languages and be regarded as being a more complex unit or morpheme. Consequently, the translator should pay attention to a number of factors when considering a single word, such as number, gender and tense. Above and beyond, such an analysis is lucid and easy to comprehend.
There are many factors that can cause the problems of non-equivalence. Baker (1992, p. 21-25) categorizes some of the problems of non equivalence at word level which are presented in the following list:

METHODOLOGY
This qualitative research study was carried out amongst the Ekegusii speakers of Kegati, Kisii County, Kenya. Descriptive design was used as it involved the usage of primary data collection methods as the most suitable route for the attainment of the study objectives. A population sample of 30 respondents were studied in the larger project. Because this paper is a preliminary report, only a brief part of the data collected and analyzed will be reflected in the findings section. The samples were arrived at through both purposive and snowballing techniques. Data was collected through the use of interview. Comprehensive data analysis was done through Baker's nomenclature of non-equivalence at word level and also strategies to solve it. During the period of data collection, the researcher presented a translation task to them. The sampled population exhibited almost homogenous linguistic processes among the members. The adult speakers were considered the main engineers of communication and this was done mainly in Ekegusii. The various responses from most of the respondents showed that they favoured particular translations.

RESULTS AND DISCUSSION
The purpose of this study was to propose a classification of non-equivalence at word level based on Mona Baker's identify and discuss strategies used to deal with non-equivalence at word level in translation of Ekegusii kinship terms. Translation is a kind of activity which inevitably involves at least two languages and two cultural traditions (James, 2002). Translators are thus permanently faced with the problem of how to treat the cultural aspects implicit in a source text (ST) and of finding the most appropriate technique of successfully conveying these aspects in the target language (TL). These problems may vary in scope depending on the cultural and linguistic gap between the two (or more) languages concerned (Nida, 1964).
Typical Ekegusii-English kinship concepts will be examined so as to prove the existence of differences between Ekegusii and English. It is believed that there always are differences between two languages, as a result, the occurrence of nonequivalence in translation is inevitable. The discussion takes the word level as the main focus. Thereafter, a classification of the non-equivalence at word level between Ekegusii and English is proposed based on Mona Baker's (1992) formbased approach. We all know that there are a couple of words which are untranslatable. Furthermore, there are other words which can be translated but they demand much more thought from us. Noteworthy, is that the Abagusii are strongly oriented around family and ancestry, and accordingly their language is richer in kinship terms than English. Karimi (2006) believes that a translator should not always find one -toone categorically or structurally equivalent units in two languages, that is, sometimes two different languages may carry the same function. This is evident even in the event of translating Ekegusii kinship terminologies to English.

Ekegusii-English Word Level Conceptual Contrastive Analysis of Kinship Terms
Ekegusii does not make a distinction between 'a step-mother' and 'an aunt' since both are referred to as makomoke, for instance in Makomoke ochire, which means 'My small mother (aunt) has come'. English, conversely, uses the prefix stepto refer to relatives related only by re-marriage, not blood, such as 'the arrogant stepmother' and 'the proud stepdaughter'. Moreover, when there are co-wives in the home, the children would consider their order in marriage when referring to them, for example, baba omonene and baba omoke translated as 'big mother' and 'small mother' respectively as both are in the Gusii context regarded as mothers because they are the father's wives.
In Ekegusii kinship, it is assumed that the mother's husband has also served as a biological father, therefore there exists no distinction between children who share one parent or are related by re-marriage. English, however, uses the prefix step-or the prefix half-to refer to children who share only one parent (half-brother and step-sister). It is easy to see the difference between 'step sister' and 'sister' in English. A 'half sister' shares one parent biologically whereas a 'step sister' is by marriage, with no blood relation. This distinction does not exist in Ekegusii since there is no concrete name for that kinship and the fact that both a step sister and half sister are considered as sisters.
For the term 'uncle', Ekegusii makes a distinction depending on whether it involves a paternal uncle or a maternal one. For that reason, the words tatamoke-'small father' that is father's brother and mame-'mother's brother' are used in different situations. Nonetheless, English makes no distinction between affinal and consanguineous uncles since it uses a general term for both thus 'uncle' cannot be considered as a total equivalent of the Ekegusii terms.
The terms tata and baba in Ekegusii are quite unclear since they may refer to one's 'father or mother' on the one hand or one's 'father in-law or mother in-law' on the other hand. These two words have more than one meaning depending on the context in which they have been used as single words which may result in misunderstanding for a foreigner. Additionally, the term omwana is ambiguous as it is used to refer to any child, one's own child, a daughter or son in-law and a niece or nephew whereas English makes a clear distinction amongst all these terms.
Ekegusii kinship system makes distinctions on relatives related by marriage (in-law status), for example, by distinctively using the terms mokoyone to refer to a 'husband/wife's sister or brother'; semo to refer to 'a man that has married one's wife's sister', and kamati used by a wife and a husband's sister to refer to each other. In comparison, English fails to make such distinctions as it uses a general term to in-laws and does not regard kinship between two men that are married to sisters.
Ekegusii makes use of the terms omokungu and omorugi interchangeably to refer to one's wife, yet the direct translation of the terms are 'a woman' and 'a cook' respectively. Accordingly, in the Gusii context the wife is regarded as a cook because that is one of the major duties that she should perform in the home. Another fact is that some times omokungu is used to refer to any woman. Comparatively, English clearly distinguishes between the terms woman as 'any adult female' and wife as 'one's female spouse'. On the same note, omogaka and omosacha can be used interchangeably to refer to one's husband, yet the terms may not be directly synonymous. Other meanings of omogaka are 'the man of the homestead' or 'an old man'. The term omosacha can also mean 'any man'. A lot of difficulties will thus be faced in the process of translating these terms.
Ekegusii makes a clear distinction between a sister's child and a brother's child by referring to them using the terms omoigwa and omwana respectively. In contrast, English refers to a sister's child, wife's brother's child, husband's brother or sister's child as either nephew or niece depending on the gender in spite of the relation.
In English the terminology 'cousin' is the most classificatory term for the children of aunts or uncles. One can further distinguish cousins by degrees of collaterality and by generation. In regard to collaterality, two persons who share a grandparent count as 'first cousins', and if they share a great-grandparent they count as 'second cousins'.
In regard to generation, if two persons share an ancestor, one as a grandchild and the other as the great-grandchild of that individual, then the two descendants class as 'first cousins once removed'; if they share ancestors figures are as the grandparent of one individual and the great-great-grandparent of the other they class as 'first cousins twice removed'. If they share ancestor figures as the great-grandparent of one person and the great-great-grandparent of the other the individuals class as 'second cousins once removed'. Ekegusii however, fails to distinguish such categories as it makes use of an umbrella term bamwabo whose equivalent is 'cousin'. This would imply that there is only one equivalent to all these terms in this language.
Ekegusii distinguishes between mother and father in-law as makobiara and tatabiara respectively, but no clear distinction exists between daughter/son-in-law and son/daughter as both are referred to as omwana one translated as 'my child'. English speakers however, mark relationships by marriage except for husband/wife with the tag 'in-law'. The mother and father of one's spouse become one's mother in-law and father in-law; the female/male spouse of one's child becomes one's daughter/son in-law respectively. S e p t e m b e r 1 1 , 2 0 1 5

Classification of Non-Equivalence At Word Level Drawn from Ekegusii Kinship Terms
As far as Baker is concerned there may not be equivalent words between two languages, especially when it comes to culture specific concepts (Baker, 1992). This means that an SL expresses a concept totally unknown in the TL. As a result, there arises difficulty to translate a word in Ekegusii to English and vice versa once it does not exist in the TL because it is a concept that the target audience has never heard. For instance, we understand that African kinship is extensive in comparison to the English one. Therefore, the translator may be forced to introduce an exotic concept to the TL audience. For instance, the term semo in Ekegusii is used by men that have married sisters to refer to each other and korera used by a mother or father to a son/daughter in-law referring to one another, yet this does not occur in English because such kinship is not recognized in their culture. Palmer (1976) is in support of this because he observes that: "Words of a language often reflect not so much the reality of the world, but the interests of the people who speak it" (p. 21).
A second common non-equivalence at word level is when a TL lacks a superordinate that is it may have specific words but no general word to head a given semantic field (Baker, 1992). For example, in Ekegusii omwana is an umbrella term which refers to child, niece, and nephew, daughter, son as well as son and daughter in-law. Accordingly, all members of the descending generation are omwana one 'my child'. This results because the Gusii kinship terminology merges lineals with collaterals. Specific lineal terms are used to denote the immediate family: tata 'own father', baba 'own mother', momura one 'own son', and mosubati ominto 'young woman of our house'.
All other men of ego's generation, however, including "real" brothers, are called momura ominto 'my brother'. In the mother's family, the reciprocal term mame is applied to mother's brothers and their wives. In any clan in which ego has kinship connections, individuals of ego's parents' generation are called tatamoke 'small father' or makomoke 'small mother'. All members of the grandchildren's generation are omochokoro, and those of the grandparents' generation are sokoro 'grandfather' and magokoro 'grandmother'. Contrastively, English has no general term for such lexical sets.
The TL may lack a hyponym that is it has not enough specific terms to illustrate words in the SL (Baker, 1992). For example, Ekegusii has two different words for 'uncle' depending on the exact relationship, mame for 'mother's brother' and tatamoke 'small father' to refer to a father's brother. In contrast, English uses 'uncle' for both.
The SL concept may not have an equivalent in the TL because the concept is not lexicalized in the TL (Baker, 1992). This implies that concepts are known but no equivalents exist in the TL that is they are not allocated a word to express it. For example, the Ekegusii term omorugi which can be translated as 'cook' is a word representing the concept that is held by almost all cultures that a wife's chief role is that of ensuring that the family has been cooked for and its well fed. This term is however, absent in the English kinship system.
The TL may at times make more or fewer distinctions in meaning than the SL. Accordingly, what one language may not perceive as relevant another may regard as important. For example, English makes clear distinctions between 'father' and 'father in-law' whereas Ekegusii may not make the distinction because sometimes both can be referred to as tata meaning 'father'. The English translator may find it difficult to choose the right equivalent, unless the context makes it clear whether the person being talked of is a biological father or one whose relation is by marriage.
A word in the TL which has the same propositional meaning may have a different expressive meaning in SL and TL (Baker, 1992). In consequence, the meaning in SL is different from that in TL; for example, when a TL equivalent is neutral compared to the SL item, the translator is forced to add a modifier or an adverb, or builds it somewhere else in the text. For example, in English one would modify 'uncle' with the expression maternal or paternal in order to handle the more loaded Ekegusii terms mame and tatamoke.
Differences in physical and interpersonal perspective may exist between the TL and SL resulting to non-equivalence. Baker (1992) notes that it is possible for the TL to make more or fewer distinctions in meaning than the SL. A word in Ekegusii might convey additional meanings relative to the English ones; hence it makes translators confused of which words to be used properly provided that the context itself offers them enough detailed information. For example, Ekegusii has several equivalents for the term omwana that is child, niece, and nephew, daughter, son as well as son and daughter in-law. In addition, omogaka might refer to husband, the man of the homestead or an old man depending on who the speaker is referring to.

Translation Strategies Employed in Word Level Translations of Ekegusii Kinship Terms
Most researchers agree that strategies are used by translators when a direct, literal translation is not sufficient for the task they are working on (Berger, n. d). Consequently, good professional translators should continuously make such strategies to the text they are working with as circumstances demand.
According to Zaky (2000), Loescher (1991, p. 8) defines a translation strategy as "a potentially conscious procedure for solving a problem faced in translating a text, or any segment of it." Translation strategies are thus a set of mental and behavioural steps that a translator uses to communicate the message of the SL text. Baker (1992) observes that strategies used in cases of non-equivalence are determined by the context and purpose of translation. Moreover, she proposes nine strategies to solve non-equivalence at word level after dealing with the difficulties S e p t e m b e r 1 1 , 2 0 1 5 implied in the lack of equivalence at word level (Baker, 1992, p.26-42). However, in the context of Ekegusii -English kinship terms only six of them are relevant. Consequently, the following section seeks to discuss them.
One of the strategies that can be employed in solving non-equivalence at word level is translating by a more specific word (hyponym). This strategy might lead to over interpretation of the source language meaning, which seems to be more dangerous than over generalization. However, in some cases, it may be appropriate or necessary to use a more specific word to translate an Ekegusii word into English. This usually involves choosing among several different words, as there may be many English words that correspond to the general category or meaning expressed by the Ekegusii word. For example, as in prior analysis, there is no English general word corresponding with the term omwana or tata or baba in all cases. But depending on each, a translator needs to use a different English word which is suitable and applicable in that context. In this example it is necessary to use the word 'nephew/niece' instead of 'child' in English since it sounds more natural.
2SG-child 3SG POSS-father fine SG-be-PRES 'My child, is your father fine?' The word 'child' is very general, so to clearly communicate the translator would use either 'nephew' or 'niece' depending on the addressee's gender, otherwise the impact will not be the same on the target audience. Additionally, English would use 'first and second cousin' instead of cousin when referring two persons who share a grandparent and two persons who share a great-grandparent on one hand. Ekegusii on the other hand will use the term bamwabo 'cousin' in both contexts. As such the translator is expected to use the exact terms to address the relations.
Non-equivalence at word level can also be taken care of by translation by a more general word; superordinate. This is one of the most commonly applied strategies in dealing with various kinds of problems in translation. The translator tends to use a more general word (superordinate) or a more commonly known word to replace the more specific one. Yet the possibility of relative ease of rendering a problematic specific concept with a more general one may result in excessive generalization and eventually in loss in meaning in the translated text. Principally, using a superordinate is one of the popular strategies for dealing with many types of non-equivalence. It works equally well in most, if not all, languages, since the hierarchical structure of semantic fields is not language-specific. Under certain conditions, it may be correct to use a more general word to translate an Ekegusii word with no specific English equivalent. For instance, English's 'uncle' which takes care of both the maternal and paternal uncles although are clearly distinguished in Ekegusii. For example, a back translation of the following sentence will not evoke the same meaning;
3SG POSS-maternal uncle and 3SG POSS-paternal uncle PL-come-PERF-FV 'My uncles have come.' Additionally, translation by a neutral or less expressive word may be helpful when faced with non-equivalence at word level. This strategy is mainly useful when a translator encounters an expressive word. However, if carelessly used, he might fail to convey the true meaning or even cause misunderstanding. There are cases when the translator picks up a word which seems to be an equivalent but perceived differently in the target language. For instance, when translating makomoke from Ekegusii to English, the translator may pick on 'aunt' instead of 'step-mother' when referring to an ego's relation by remarriage which is less expressive in the TL. For that reason, he decides to use a less expressive correspondent in the target language to avoid the risk and to sound natural.
Another strategy that translators are prone to use when faced with non-equivalence at word level is translation by cultural substitution. This strategy involves replacing a culture-specific item or expression with one that has a different propositional meaning but is likely to have a similar impact in the translated text. For example, using omorugi to refer to one's wife or kamati to refer to one's brother's wife who is generally recognized as sister in-law in English.
3SG-wife to my brother SG-go-PERF market.
'My sister in-law has gone to the market.' A back translation of the sentence in example 3 will not create the same impact on the target audience because Ekegusii has specific terms for particular sisters-in-law. Furthermore, translation by paraphrasing is another possible way employed in coping with word level non-equivalence in the event of translation. When using it the translator has two possible solutions at his disposal; translation by paraphrase using a related word or translation by paraphrase using an unrelated word. In translating semo, the translator is obliged to use a related expression, 'men that are married to sisters' which is presumed to have the same meaning as the SL; S e p t e m b e r 1 1 , 2 0 1 5 This strategy is used when the concept expressed by the SL item is not lexicalized at all in the TL. However, when the concept expressed by the SL is lexicalized in the TL but in a different form, and when the frequency with which a certain form is used in the ST is significantly higher than would be natural in the TL we may paraphrase using an unrelated word. For example, the Ekegusii makomoke which means small mother/ step-mother in the language is commonly associated with 'aunt' only in English.
The main advantage of translation by paraphrase is that it is possible to achieve a high level of precision in specifying the meaning of a word or concept that poses difficulties in translation (Baker, 1992). Nevertheless, its main disadvantage is that it usually involves replacing one item with an explanation consisting of several items. Thus an outstanding difference in length of the source text and target text may occur, which is barely ever a desirable effect. However, this strategy is applicable for the term that is known but not lexicalized in the target language and the case of loan word in the source language. Paraphrasing is also helpful in addressing the problem of semantically complex words.

Conclusion
Translation equivalence cannot be easily achieved since it depends on the text, the translator, and the target audience. It is a reality that no matter how competent the translator may be, the translation might lose meaning relative to the source text. Both the linguistic and the cultural gaps among languages create the possibility of non-equivalence in translation.