# Introduction en and women in the 21 st century continue to experience a juxtaposed ideological representation on matters masculinity and femininity. Language, particularly Discourse Analysis, provides a means of analyzing these discourses and how they are perpetuated. One field where sexist discourses are perpetuated in Kenya is the electronic media, more so the Radio. Being a vibrant forum for exercising gender and power struggles, Radio offers a ground where gender inequalities are portrayed and broadcast. Daily conversations, call in sessions and topical discussions provide a rich source of sexist discourses. Since 1980s, when there was only one radio station, Voice of Kenya, to date, when there are 158 radio stations in Kenya, Radio industry has generated a fair amount of scholarly attention, with research focusing on the harmful depictions of gender that portray men and women as different and men being better than women. The potential effects of such portrayal on the audience is that when repeatedly presented, such discourses tend to be accepted as the norm and transferred from one context to the other leading to repeated proliferation of stereotypical gender differences, with regard to gender roles and expectations. Given the rise of FM stations in the last 15 years, sexist discourses have been advanced with the focus being on the expected masculine/ feminine body, roles and behaviours. Previously, women were portrayed as thin/slim, home makers and nurturers, while men were portrayed as strong full of muscles, providers and protectors. The image of strong man verses beautiful women has for long existed in discourse analysis (Wen, 2003). There is clear evidence that most radio stations air gendered topics and expect callers to participate in these discussions. And listeners, in their safety of anonymity express the opinion and feeling without fear of judgement. Using one popular radio station (Classic 105 FM), an analysis of lexical items and metaphors that challenge sexism and binary hierarchical representations of men and women is carried out. Radio Call in Sessions conventionally construct and reproduce asymmetrical gendered relations in both domestic and social spheres. They therefore provide a forum through which gendered ideals can be scrutinised. I analyse them for linguistic features and discourses (viewed as language reflecting and shaping sociocultural norms and practices). Another reason is that radio conversations would be easily available during morning call in sessions, appropriate in answering my questions and analysable using my CDA methodologies. They constitute sexism related vocabulary, are rich in sociocultural aspects and are empirically original since they have not been researched before. # II. # Sexist Language Sexist language is one that reinforces and perpetuates gender stereotypes and status differences between men and women (Swim et al 2004). Sexist language condones unequal treatment of women and men (but mostly women). This can occur overtly, covertly or in subtle forms. Subtle sexism is a form of sexism that is hidden but exhibits unequal and unfair treatment of women (and men). It is not recognized by many people because it is perceived to be customary, normal and harmless, it does not appear unusual. Subtle sexist language is difficult to change because people don't believe that such a language exists or that it is problematic because people are used to it. In most instances, subtle sexism goes unnoticed. This form of sexism is different from overt sexism which is readily apparent, visible and observable. It is also slightly different from covert sexism which is defined as the hidden and clandestine unequal and harmful treatment of women (Swim and Cohen 1997). This paper focuses on sexism generally on the basis that it is difficult to recognize and consequently difficult to critique in social contexts because it is accepted. This makes its effects to be lethal. Individuals may not notice when they are being sexist until their perpetuation becomes common knowledge and unquestionable. Perpetrators may not realize that their behaviors contribute to unequal and harmful portrayal and treatment of women. # III. # Discourse and Gender Representation While earlier research (see Lakoff 1975) revolved mainly around how language is used by women and men differently, more recent approaches are concerned with how women and men create identity in and through discourse (Wodak 1997;Sunderland & Litosseliti 2002) and how gender identities are constructed in texts. Research questions have changed in recent years from how men and women speak to the kinds of linguistic resources men and women employ to present themselves and others as certain kinds of people, and the kinds of linguistic practices that support particular gender ideologies and norms, or challenge them (Cameron 1998;Litosseliti & Sunderland 2002;Sunderland 2004). Recent studies on language and gender (see Atanga 2013; Ellece 2011) have shifted focus from differences between men's and women's use of language, that is, from who and how, to what is talked about (which includes gender) and how, rather than by whom thus de-emphasising gendered speakers and writers as agents and focusing on what is said/written to and especially about women, men, boys and girls (Sunderland & Litosseliti 2008). Weatherall (2002) explains that research has moved from language to discourse by considering how language in use reflects and perpetuates or challenges gender stereotypes. While earlier gender and language work documented how individual words could be considered sexist, current works examine how texts are used to challenge sexist positions. Research (see for example Kosetzi 2012) that focuses on constructions of gender and which inform this paper include those that have questioned constructions and representations of men as logical, intelligent, responsible and challenge portrayals of women as inferior, stupid, unreliable, irresponsible, emotional, less rational and vulnerable. The terms, construction and representation, cannot be exhaustively discussed without mentioning Butler's (1990) gender performativity concept. Butler argues that gender is a performance and people normally perform their gender thus constructing their identities. In Kenya for instance, women perform duties (such as cooking, washing pots) which are perceived to be feminine, in this way they are performing their gender and with time these chores tend to be associated with femininity and any male performing them is labelled as weak. Individuals also perform their gender in speech, for instance, through supporting arguments that contain messages of gender roles and responsibilities. However, this paper will not focus on gender performativity but on representation and construction, that is, how gender is represented and constructed in the media. I view representation as showing the attributes without necessarily stating what one is or what they are perceived to be and construction when direct description is given or a speaker presents what is perceived to be the norm, the expected, the anticipated. Both cases of representation and construction are evident in the data. Gendered discourses are 'discourses that say something about women and men, girls and boys and about their gendered actions, behaviours, positions, choices, relations and identities. More specifically, gendered discourses are discourses that represent and (re)constitute, maintain and contest gendered social practice' (Litosseliti 2006:58). For instance, the gender difference discourse that represents men as aggressive and active and women as inactive in certain aspects is said to be gendered. A discourse is gendered if it suggests or states that men or women do things or behave in certain ways because they are men/women or because they are expected to behave so since they are men/women. Actual behaviour may or may not correspond to those representations and expectations (Wetherell et al 2002). Gendered discourses are articulated by both men and women in different ways and different situations. They position men and women in certain ways and people take up (or challenge) such positions that constitute gender more widely. Discourses can therefore be gendered as well as gendering (Sunderland 2004). Common gendered discourses include: discourses of parenthood, femininity, heterosexuality, feminism, female emotionality and crises of masculinity. choices, relations and identities. More specifically, gendered discourses are discourses that represent and (re)constitute, maintain and contest gendered social practice' (Litosseliti 2006:58). For instance, the gender difference discourse that represents men as aggressive and active and women as inactive in certain aspects is said to be gendered. A discourse is gendered if it suggests or states that men or women do things or behave in certain ways because they are men/women or because they are expected to behave so since they are men/women. Actual behaviour may or may not correspond to those representations and expectations (Wetherell et al 2002). Gendered discourses are articulated by both men and women in different ways and different situations. They position men and women in certain ways and people take up (or challenge) such positions that constitute gender more widely. Discourses can therefore be gendered as well as gendering (Sunderland 2004). Common gendered discourses include: discourses of parenthood, femininity, heterosexuality, feminism, female emotionality and crises of masculinity. Gendered discourses are categorised under interpretive discourses. These are those whose interpretation depends on the analyst's standpoint, for instance 'sexist discourse' or 'racist discourse'. These may seem different from one person to another because what may be viewed as racist discourse by one person may not be viewed so by another. As such, gendered discourses are viewed as so only by an analyst who aligns themselves with a feminist perspective. In order for an analyst to argue that a discourse is gendered, they must provide evidence from the language used by identifying the linguistic items used or linguistic cues/traces that are in a text. Linguistic realisations of discourses are called traces of discourse and are listed by Fairclough (2015) as lexical items, modality, agency, process types and how social agents are represented. I will identify and analyse some of these linguistic traces of gendered discourses in my data. It is worth noting that discourses exist in relation to other discourses (Litosseliti 2006), they co-exist in contradicting and conflicting relationships (Jorgensen & Phillips 2002) and some discourses are dominant while others are marginal in different texts and contexts. V. # Theoretical and Analytical Framework a) Fairclough's Textually Oriented Discourse Analysis Textually-oriented discourse analysis is a linguistically based approach that focuses on any sort of discourse, both written and spoken -conversation, classroom discourse, media discourse and so on (Fairclough 1992). Fairclough (2003) proposes a critical discourse analysis that focuses on a problem, identifies the obstacles in the efforts to solve the problem; considers whether the social order needs the problem; identifies ways past the obstacle and reflects critically in the analysis. CDA has been selected because of its explicit concern with social issues and problems (Fairclough 2001). The social problem in this paper is sexism and gender inequality. Fairclough advocates for an analysis of actual conversations, interviews, written texts, media programs and other forms of semiotic activity. His three-tier analysis model states that the first step involves close examination of linguistic features (vocabulary, grammar and textual features). He lists a comprehensive list of textual features to focus on and suggests that analysts do not have to examine all of them but can analyse a limited number relevant to their text, context and purpose of the research. The following is a list of guiding questions during textually oriented discourse analysis. A. Vocabulary 1. What experiential values do words have? ? Are there words that are ideologically contested? ? Is there rewording and over wording? ? What ideologically significant meaning relations are there between words? # What relational values do words have? ? Are there euphemistic expressions? ? Are there markedly formal or informal words? 3. What metaphors are used? From Fairclough's framework and list of guiding questions, an analysis guideline, is developed as summarised in the table below. # Discourse analysis How does the data borrow from other texts? Which discourses are dominant? Which minor discourses shore up dominant discourses? # Social analysis What gender relations common in Kenya are rearticulated in the data? What power relations are evident in the data? What is the effect of the textual features on the power and gender relations? What is the effect on the perpetuation of subtle sexism? How is discourse determined by social structures (and practices) and what effects does it have on those structures (does it contribute to their sustenance or change/challenge them?) These guideline questions were employed in the analysis of the data as stated earlier, my focus is at three levels of analysis: description, interpretation and explanation. In the description stage, I identify lexical items and metaphors in the data. In the interpretation stage, I explore intertextuality how data from the conversation draws from other outside texts in the Kenyan context, and interdiscursivity I look at how some discourses shore up others. In the explanation stage, I examine the effect of the linguistic representations on social and power relations in Kenya and on the perpetuation (or not) of sexism and gender inequalities. # b) Sunderland's Gendered Discourses: Identifying and Naming Discourses The process of discourse identification is always interpretive because there is no finite set of discourses. Sunderland (2004:28) argues that discourses are not always recognised easily; they 'are not simply out there waiting to be spotted' but are 'in flux' (see also Litosseliti 2006). There is, therefore, no discourse that self-evidences itself as a discrete chunk of a given text in its entirety. Instead, as (Baker 2008:95) argues, 'what is there are linguistic features: 'marks on the page', words spoken or even people's memories of previous conversations [?] which -if sufficient and coherent may suggest that they are 'traces' of a particular discourse'. There are no specific traces for certain discourses; the same linguistic features can be identified and interpreted as cues and traces of different discourses by different analysts. There are also no definite criteria for deciding that a linguistic feature is a trace of a certain discourse either. We can however, 'detect a discursive structure because of the systematicity of ideas, opinions, concepts, ways of thinking and behaving which are formed within a particular context and because of the effects of those ways of behaving' (Mills 2004:15). This systematicity of ideas can be manifested through repetition of words or phrases (Atanga 2010) which show that meanings are not 'merely personal and idiosyncratic but widely shared in a discourse community' (Stubbs 2001:215). Repeated patterns tend to acquire a common sense, taken-forgranted status and with time they are no longer questioned but become part of a people's vocabulary and hegemonic discourse. To be powerful, therefore, discourses must have the characteristic of repetition (Baker 2008) although single instances (and absences) are also considered. An analyst does not only consider repetition but also the articulation of a discourse by a powerful speaker and how accessible the discourse is to a large number of people. Identification requires co-construction by the language user of the text and elements of its production. Sunderland (2004) posits that some discourses can be pervasive, widely recognisable and therefore dominant while others will be marginal, supporting or shoring up the dominant discourses. To verify the presence of a particular gendered discourse in a text, the analyst should provide evidence that suggests the presence and workings of that particular discourse. Linguistic traces can be identified by applying the categories given by Fairclough (2015) such as vocabulary as well as intertextual features. Fairclough and Sunderland's approaches are therefore interdependent in the sense that while Sunderland's is used to identify discourses, Fairclough's provides the categories of the linguistic traces of the discourses. Text analysis is the main way of justifying an interpretation. To be sure that discourses have been correctly identified, an analyst should check that the identified discourses are recognised by others within and outside the same field, as suggested by both Fairclough (1992) and Sunderland (2004), because discourses can only exist if they are socially acceptable to some people, and provisionally recognisable (Jaworski & Coupland 1999). Linked to discourse identification is the naming of discourses. The analyst may often name discourses according to which particular stand point they come from. Although Chouliaraki & Fairclough (1999) argue that the label used is not that important as long as a discourse is recognisable, Fairclough (2003) warns that the lack of a 'closed-list of discourses' does not mean that every single representation of the world should be named as a separate discourse. The analyst can use existing names that are familiar and instantly recognisable to others, or can come up with their own, however they must then give evidence to justify why a discourse has been identified and named as discourse of X or X discourse. Naming can be based on the discourses' functions (e.g. damaging, liberating, resistant, subversive, conservative discourses) or their relation to other discourses (e.g. challenging, contradicting, competing, dominant, alternative et cetera (Baker 2008)). Both identification and naming of discourses are interpretive acts. Although there are many discourses, this work focuses on discourses that that challenge gendered ones because any human experience can be gendered (meaning that there is something in them that has to do with gender) and consequently can be challenged. Gender discourses exist in relation to other discourses (Litosseliti 2006), for instance to patriarchal discourses, conservative discourses and egalitarian discourses, in different ways. The term 'gender discourse' is different from (yet related to) 'gendered discourse' used in this paper. The former referring to general discourses about gender while the latter being used to specifically mean those discourses about gender which are destructive and discriminative against women (and men sometimes). Once a discourse has been identified and named, it is important to look at its social significance, to find out how it positions social actors, in this case, women and men. This is because any discourse is not just a concept but also a social and constitutive process. From a gender perspective, particular discourses construct women and men in gendered ways, and such gendered subject positions are taken up as the norm and go unchallenged. Gendered discourses have been found to put both women and men in unfavourable subject positions, though Sunderland (2004) argues that it is women who tend to be constructed in more conservative ways; such discourses are referred to as damaging discourses. Examples of damaging discourses are traditional discourses which have been circulating in more or less the same form and have become naturalised in a particular community, for instance, those that present women's subordinate position to men as natural and normal. The 'woman as object discourse' and 'gender difference discourse' are other damaging discourses, because, through them, the society legitimates genderbased hierarchies by normalising unequal gender relations and positions. Of focus in this paper are those discourses that are challenged. A challenge to a discourse occurs when the specific discourse has been criticised without being explicitly substituted by another one (Kosetzi 2012). # c) Data Data for this paper constituted 20 words selected purposively from a corpus of 100 words collected from many sessions of tuning into Classic 105 FM morning drive hosted by Maina and King'ang'i (Maina Kageni and Daniel Ndambuki). The researcher tuned in from 8am to 9am on weekdays when the hosts engaged listeners in discussing various topics ranging from: relationships, crime, corruption, family, marriage, education among others. The researcher focused on the lexical items and metaphors relating to sexism, gender and unequal representation of men and women. Data collection took place between June 2017 and June 2020 with breaks from time to time. Maina and King'ang'i host the call-in session where the audience is given an opportunity to call in and opine on the topic of the day. Classic 105 FM morning drive is a major site for exploration of sexism and gender discourses. # d) Procedures On every session, the researcher had a pen and paper where she took statistics of gender related lexical items; words that described men and women in binary hierarchical ways and paid attention to those instances where they were from men or women. She also focused on metaphors and proverbs that were deemed to be gendered. Once satisfied that the data was enough, a list was made with distinctive columns based on gender and frequencies. The researcher then analysed the 20 words which were perceived to be appropriate to answer the questions posed using CDA procedures. See Table I for a list of the analysed words. # Discussion a) Proper Woman The proper woman discourse is a conservative discourse which supports the subservient position of women. It sets the ground for asymmetrical arrangements whereby women are assigned specific gender roles, responsibilities and expectations to fulfil for them to fit in the proper Kenyan woman category. From the discussions in the Call-in-Sessions, a proper woman is one who performs her ascribed feminine duties in her homestead once married. Such duties include washing pots, cooking, serving and minding about others affairs. However, the discussers (mostly men callers) argue that Kenyan women are no longer proper women, they are not wife material, they are 'slayers' and 'slay queens' whose only concern is on whether they have appeared beautiful enough to attract the next man, whom they slay and dump. Callers argue that Kenyan women have transgressed from their conventional feminine roles and are now 'wanting to be men'. Shoring this discourse are others such as 'woman as independent' 'woman as breaking the ceiling' among others which describe women duties as no longer confined to the kitchen. Ability to go out and seek for employment has ensured that women no longer stay at home to wait for their husbands to provide. Traces of this discourses include the lexical items and metaphors identified and discussed in Table I such as slay queen, independent, madam boss among others. Through the 'boss' metaphor, a Kenyan woman is constructed as strong and in control. She is fearless and resistant to intimidation; she calls the shots. She is not shaken or pained even when she has been left by her lover, like an employer she 'sacks' her lovers and 'employs' others in a twinkle of an eye. She does not cry when her relationship is broken because doing this would be interpreted as weakness and disqualify her from being the boss, a position that she has created for herself. She quickly accepts that the relationship is over and 'moves on'. The discussers argue that crying or showing signs of fear are considered as weakness and feminine and the Kenyan woman no longer wants to be perceived as weak. Crying and other emotions are associated with cowardice and are described as women attributes which symbolise weakness. By not crying, when abandoned the Kenyan woman is elevating herself to the status of 'maleness' because men are not allowed to cry or show fear. Therefore, by not showing weak emotions such as crying, the status of a woman is raised towards that of a man. A proper Kenyan woman is also constructed as non-domestic. This is in contrast with what Sunderland (2004) and Atanga (2012) have identified and named as woman as domestic discourse. This woman as nondomestic discourse is manifested in linguistic features used in the data. Women are constructed as staying away from conventional titles such as mothers, wives, carers, family cooks, and family servers. Women are no longer the primary domestic servers whose key role is housework. They are active outside their homes, and employ home managers to assume their domestic duties. Main traces of this discourse are lexical items such as poor cook, not a slave, madam boss among others. This discourse is prominent in the discussions. The Kenyan woman is described as being unable to peel onions or chop tomatoes due to her long nails. Instead, she asks 'mama mboga' to chop all the ingredients and mix them such that her only duty is to place the mixture in a cooking pot and boil. The end result as described by one man is 'a horrendous mkorogo' that no man in his senses can tolerate. Apart from being unable to cook, a Kenya woman is described as being unable to wash. 'They would rather buy a new bed sheet than wash a dirty one,' says one man in one of the discussions on who should do domestic chores. # b) 'Woman as poor cook In the data, Kenyan women are described as improper. Through the word murderer which is used both directly and metaphorically (to mean women have ability to kill the wallet) and monster which has been used to describe women as aggressive and dangerously destructive. One caller, a male had this to say, 'Kenyan women have become worse than the devil, they eat even their own children. Maina, let me tell you, I know of a lady who has aborted five times, this one if you marry she will kill you live live'. In the African and Kenyan context giving birth has conventionally been taken to be a form of 'taking root' in a husband's home. When a woman has given birth, she has established herself in her matrimonial home. Her husband will have no reason to send her away, since she is reproductive. The woman is also expected to give birth quickly to many children, one after another. In the Kenyan society, as is the case in many African societies, a woman who gives birth as quickly as possible earns respect among her peers and honour in the society at large. However, this conventional position is challenged in the discussion where having many children is viewed as a form of slavery for the modern woman. Women are now avoiding giving birth to more than one or two children because, '?we don't have time to care for many and you cannot get a house manager easily if you have more than two children Maina. Many children mean slavery; you will never go out to do anything.' Giving birth comes with caring responsibilities that are shouldered mostly by the woman. Child care is discursively constructed as a natural responsibility of women only (Atanga 2010). But this duty has largely been delegated to baby sitters, meaning that not even child birth will keep women in domestic spheres. A proper Kenyan woman is also constructed as brazen, confident and unafraid. Traits which make her appear resilient and independent. This challenges the 'woman as timid discourse', a discourse that positions women as lesser objects that need protection and assurance (from men). Woman as brazen, confident and unafraid discourse is mainly advanced through the lexical item 'boss' which is found in the data where the woman is described as 'boss lady' or 'madam boss'. This description positions women as secure and in no need of protection and validation from men. It depicts the Kenyan woman as an increasingly confident and power-wielding being who is not subject to control by others (especially not by men). It is an example of reaffirmation of the modern woman's position in the society as having defeated asymmetrical gendered conventions, especially where men have always been constructed as spending time outside the house and the compound, women now spend more time outside the kitchen: in the office, in businesses in salons and women group meetings. The phrase, not wife material, is a trace of the discourse that challenges the 'woman as wife discourse' that has been identified by most gender scholars. As a wife, a Kenyan woman has her roles and expectations predetermined. She is expected to bear her husband children, look after them, fetch water and firewood, cook, sweep the compound and cultivate fields. Women who do not fulfil these roles are ridiculed and they suffer socially and psychologically. However, in the current study, the 'woman as wife' discourse is challenged through the word wanderer where a Kenyan woman is constructed as unable, or unwilling, to do house chores because she is always away/ out searching and looking for the next man to maul. Through linguistic items women are depicted as failing to perform the expected duties that a married woman is to do for her matrimonial family by virtue of the fact that she is married there. She is not expected to wake up before everyone else in the morning, to greet her husband and his relatives and prepare breakfast for the household, she can have a liein and have her breakfast in bed. Proper womanhood is no longer attached to being someone's wife, neither is marriage a necessary step to a woman's happiness. In this way, marriage and becoming someone's wife have been described as unnecessary by discussers, it is not something a Kenyan woman perceives to be natural; 'I cannot die for marriage, if it comes, fine, if it doesn't, life must go on', says one of the female callers. # c) Men don't commit From the discussions, this discourse was prevalent. It was discovered that men do not want to commit to marital relationships. Through the lexical items hit and run, eater and woman eater. The discourse was found to be prevalent. The argument put forth is that the modern Kenyan man does not want to be tied down with responsibility, he wasn't to be independent, and marriage is perceived to be a way of taming men. They are therefore unwilling to commit to anything more than sex and would put any effort to bar commitment to relationships and marriage (see a similar finding in Kosetzi 2012). The Kenyan man therefore meets a girl/lady has sex with her and 'disappears into this air, you think he enjoyed it and the two of you had a good time, but when you call him in the morning, Maina, you find that you have already been blocked', says a female caller. On their part men agree that they fear commitment because they don't see 'wife material' in Kenyan women. Through the cues/traces of slayer and slay queen one man says 'these are not women you can take home to your people as a wife, they cannot cook or wash, what they do better is keep long nails and apply make-up; we can only use them for sexual satisfaction and move on'. This discourse is supported by a 2009 research which indicated that 60% of men run from a relationship when the partner attempts in a cunning way to make them commit/marry. # d) Contradicting Discourses/Discourses Challenging Traditional Discourses Sunderland (2004) describes contradicting discourses as those that challenge traditional ideologies. They criticise conservative discourses, challenge patriarchal values and are progressive in nature; they show a change or subversion from the norm. While studying Chola gang-girls, Mendoza-Denton (1996) found that the girls used make up to exert a form of power and reject sexualised femininity. The eyeliner, lipstick and hair styles worn by the members of the gangs portrayed them as hard, tougher and more masculine. The girls use makeup not to entice men nor for men's admiration (male gaze) as seen in Kosetzi & Polyzou (2009) and Atanga (2012) studies but to establish themselves as powerful, intimdidating and ready to fight. This challenge the traditional 'makeup-asa-form-of-beauty discourse' which has always portrayed made up women as beautiful and soft (objects of gaze), as is the case in Kosetzi's (2007) work. Challenges to traditional discourses in the circumcision data occur when, for instance, subjects are presented in a manner which would, in normal circumstances, be interpreted as going against the norm. Men are often stereotypically represented as strong and powerful whereas women are always weak and powerless. However, there are instances in the data where this is the opposite. # e) Men have no Backbone This discourse is anchored on the argument that the modern Kenyan man is not afraid to live off a woman's sweat. It contradicts the age old belief that men are providers thus challenging the 'men as providers and men as protectors' discourses (see Wambura 2018). Through the cues/traces such as kupe, stingy, rider and dependent. Most female callers, and some men too, construct the Kenyan man as one who has lost his ego. He is not embarrassed to sit in the house and flip through the TV channels as his woman goes to work or business Other traces of contradicting discourses in the data include the metaphorical use of the wanderer, silent murderer and slayer while referring to women. These traces are associated with freedom, and 'murderer' relates to destruction which is not a feminine trait. Destruction is associated with masculinity and it is men who show destructive attributes through traces such as strong, macho among others. However, when these metaphors are used for women, there are objectives to be achieved. The reference of wanderer, murderer and slayer to women shows that the modern Kenyan woman is evolving into a form of 'manness' and can now possess a form of power that only lies with men hence challenging the timid, naïve and subordinate attributes initially associated with women. As stated earlier, it presumed that it is the sole responsibility of a man to provide for his family. African and Kenyan men go out to bring home what the family needs and women stay at home to prepare, cook and serve what has been brought. However, in the data women have been constructed as doing more chores and duties than their husbands on average. They have also been constructed as providers of material needs with men being lazy and just depending on them. With the traces such as 'kupe' (pest) dependent, empty pocket. I view this as a contradictory discourse as it presents a deviation from the norm. Satia (2014) found a similar situation while studying letters of inmates in Kenyan prisons. In his case women constructed themselves as breadwinners. They shouldered the responsibility of providing for their families and parents while their brothers remained unconcerned. Traces of this discourse in the data include instances where women are constructed as working harder than men hence enacting the 'woman as hard worker' discourse which challenges the 'women as dependent' discourse. Here, women are portrayed as the main source of labour in their matrimonial homes, as seen in the description of what a woman does daily. # f) Woman as Wanderer This discourse challenges the dominant discourse of women confinement within the walls of the house. Apart from being hard-workers, women are also constructed as uncaged. The main trace of this discourse lies in the 'wanderer' lexical item whereby discussers state that women do a lot of things in the day ranging from dropping kids off to school, going to work, shopping, picking kids up, cooking and performing wife and mother-related chores, they are always on the move trying to get what the family will eat while most men go to waste time and drink in bars, 'men have turned to john walker literally, they frequent bars after work every day only to go home in the evening drunk and asking for their food'. They have left all the responsibilities to women in and out of the home. # VII. # Summary The first objective was to find out 'which lexical items and metaphors are employed in the Classic 105FM morning drive. I categorised lexical items and metaphors identified in the data into two categories: those that referred to men, and those referring to women. Categorisation aimed at giving a broader picture of how the two groups that form a binary hierarchical gender structure in the Kenyan community, were being portrayed and what the perception was about their position in relation to each other and to the society in general. The results showed that men's position is no longer higher; they are not dominant. They are represented as being inactive and dependent, they are always away from home drinking and not performing their manly duties. They are no longer involved in activities that require action, strength and courage; they have left these to their women. Women on the other hand are no longer subordinated; they are represented as being active both in and outside the domestic spheres. This aspect of the findings, I argue, challenges the broader picture of existing gender relations and positioning in most societies. Men are no longer active in doing things and women are no longer waiting for things to be done to them or done for them. Women are out to compete and beat men in activities that were initially male-oriented. They are the executors and they direct actions while men wait for food on the table. This has been achieved through the analysis of linguistic traces (micro level), the gendered discourses (meso level) and the existing social structures in Kenya (macro level). Again, women are constructed as being no longer interested in cooking, serving, bearing children and minding about the interests of others (they are not other-centred). They are not confined to chores within domestic spheres. These findings are a critique to previous studies on representation of gender (see Kosetzi 2007;2008;Wambura 2018) where women have been constructed as dependent on men and as domestic (Atanga 2010; Ellece 2012). The non-binary representations in the data shows that a Kenyan woman no longer stays at home so that when her husband comes back he finds she has cooked his food. In fact, women now demand for an equal share in doing house chores. They are no longer naïve nor do they bow to pressure or accept to be hit/caned as a form of love. Women are no longer concerned about what their husbands would eat neither do they throw tantrums when he has slept out. This is supported by a common phrase that 'sharing is caring' and 'he is only yours when he is with you'. Viewed critically, such constructions still legitimate unequal gender and power relations and disadvantage women, in the sense that they have to do both what was conventionally considered to be 'men duties' together with their duties, hence being disadvantaged even more. 1Linguistic FeaturesExperiential ValuesRelational ValuesExpressive ValuesTextualHow is reference allocated?What relational values do wordsWhatexpressiveanalysishave?valuesdowordsWhat is the classification scheme ofWhatsocialrelationsarehave?Lexical itemsword categories?produced/affirmed/ challenged?What social identitiesAre there rewordings?are constructedAre there absent lexical items?Are there meaning relations betweenwords?What knowledges and beliefs arere(produced)/confirmed?MetaphorsWhich metaphors are used?WhatsocialrelationsareWhat social identitiesHow are they allocated to the socialreinforced/legitimised?are constructed?actors? ILexical Items and MetaphorsLexical Items and Metaphors ReferringReferring to Mento Women1.Simba (lion)Slayer2.Woman eaterSlay queen3.EaterPoor Cook4.Hit and runMonster5.Kupe (pest)Silent murderer6.RiderDifficult7.AbsenteeNot a Slave8.Empty pocket manFailures (married women)9.StingyStingy10.HyenaMadam boss11.BoxerMsupa (beautiful)12.John walkerWandererVI. * Gender, Discourse and Power in the Cameroonian Parliament LAtanga Bamenda: RPCIG 2010 * African feminism? LAtanga Gender and Language in Sub-Saharan African Contexts: Tradition, Struggle and Change LAtanga SEEllece LLitosseliti &JSunderland Amsterdam and Philadelphia John Benjamins 2013 * Sexed Texts, Language, Gender and Sexuality PBaker 2008 London: Equinox * Gender Trouble: Feminism and Subversion of Identity JButler 1990 Routledge New York * Gender, language and discourse: a review essay DCameron Signs: Journal of women 23 4 1998 Culture and Society * Discourse in Late Modernity: Rethinking Critical Discourse Analysis LChouliaraki NFairclough 1999 Edinburgh University Press Edinburgh * Be a fool like me': gender construction in the marriage advice ceremony in Botswana-a critical discourse analysis SEllece Agenda: Empowering Women for Gender Equity 25 1 2011 * Discourse and Social Change NFairclough 1992 Polity Press Cambridge * Language and Power NFairclough 2001 Longman Group London nd edition * Analysing Discourse: Textual Analysis for Social Research NFairclough 2003 Routledge London * Critical Discourse Analysis: The Critical Study of Language NFairclough 2010 Pearson Education Limited Edinburgh * Language and Power NFairclough 2015 London: Routledge rd edition * Introduction: perspectives on discourse analysis AJarwoski NCoupland The Discourse Reader AJarwoski & NCoupland London Routledge 1999 * Discourse Analysis as Theory and Method MJorgensen LPhillips 2002 London: Sage * Representations of women in terms of gender roles and sexual practices in ?????? ????? (???) ('almost never'): the text and the audience KKosetzi 2007 UK Lancaster University PhD Thesis * Challenges to) Hegemonic masculinity in Greek fictional television KKosetzi Culture, Society and Masculinities 4 2 2012 * The perfect man, the proper man' -construals of masculinities in Nitro, a Greek men's lifestyle magazine -an exploratory study KKosetzi APolyzou Gender and Language 3 2 2009 * Language and Woman's Place RLakoff 1975 Harper and Row New York * Gender and Language: Theory and Practice LLitosseliti 2006 Hodder Arnold London * LLitosseliti JSunderland Gender Identity and Discourse Analysis 2002 * SMills London: Longman. Amsterdam: John Benjamins 2004 nd edition * Language and the construction of identity in legal contexts ESatia 2014 Kenya Moi University PhD Thesis * Words and Phrases: Corpus Studies of Lexical Semantics MStubbs 2001 Blackwell Publishers Oxford * Gendered Discourses JSunderland 2004 Palgrave Basingstoke * Current research methodologies in gender and language study: key issues JSunderland LLitosseliti Gender and Language Research Methodologies KHarrington LLitosseliti HSauntson &JSunderland New York Palgrave Macmillan 2008 * JKSwim LLCohen CovertOvert SubtleSexism Psychology of Women Quarterly 21 1997 * Detection and Use of Sexist Language JKSwim RMallet CStangor Sex Roles 5 3 2004 * The Earrings and the Shields': Metaphor and gender discourses in female genital mutilation songs in Kuria JBWambura Kenya. Gender and Language 12 1 2018 * AWeatherall Gender, Language and Discourse. Hove: Routledge 2002 * Discourse Theory and Practice. A Reader MWetherell STaylor SYates 2002 Sage London * Gender and Discourse RWodak 1997 London: Sage