Np-Trace in Ekegusii

International Journal of Language and Linguistics Vol. 2, No. 4; October 2015 Np-Trace in Ekegusii Deborah Nyanchama Oyioka Deparment of Linguistics Maseno University Postal Address: 186-40200, Kisii-Kenya. Dr. Benson Oduor Ojwang Department of Linguistics Maseno University Postal Address: 333, Maseno-Kenya. Dr. David Ogoti Ongarora Department of Linguistics Maseno University Postal Address: 333, Maseno-Kenya. Abstract One of the prominent cross-linguistic properties of natural language is the presence of phonetically null elements in the sentence referred to as empty categories. An empty category (EC) occurs because of a fundamental assumption which dictates that the thematic structure is directly mapped onto phrase structure rules (PSRs) at every level of syntactic representation. This paper offers an in-depth study of EkeGusii NP-movement within Chomsky's Government and Binding framework which holds that an NP-trace is left when A-movement occurs. The study argues that overt NPs leave traces behind during movement whereas non-overt NPs replace the null subject. The data for this paper is collected from essays and utterances by people who are native speakers of EkeGusii. This paper hopes to enrich the knowledge on the study of theoretical linguistics, specifically Bantu linguistics, and contribute towards a theoretical understanding of syntactic phenomena. Key words: Empty category (EC), NP-trace, A-movement, Binding Theory, Universal Grammar (UG), agglutinating 1. Introduction Ekegusii is a Bantu language spoken by Abagusii people of Kenya. It is classified as a Central Bantu language (Guthrie, 1971). According to Elwell (2008), further notes that EkeGusii has been classified as an E. 10 language spoken in South Western Kenya, the present Kisii and Nyamira counties. EkeGusii has a rich concordial system and the study of both overt and covert NPs in EkeGusii will help investigate deeper into the syntactic behavior of the nucleus of the sentence. This paper studies NP-trace in EkeGusii in the light of Government and Binding Theory and also brings out the differences between NP-trace and its overt counterpart. This study endeavours to show that Chomsky’s (1981) theory accounts for the linguistic phenomenon in EkeGusii. The paper addresses NP-trace and shows that a moved NP leaves a trace behind while a null NP replaces a null subject. 2. Background According to Chomsky's typology of NPs, overt NPs are divided into three types: anaphors (which include reflexives and reciprocals), pronouns and referring expressions (Chomsky, 1982). In the light of Chomsky’s theory, anaphors are subject to Principle A of the binding theory, pronouns are subject to Principle B and referring expressions are subject to Principle C. 103 ISSN 2374-8850 (Print), 2374-8869 (Online) © Center for Promoting Ideas, USA www.ijllnet.com Radford (1997) defines ECs as syntactic structures that contain empty or covert or null categories. Accordingly, such categories have no overt phonetic form and hence are inaudible or silent. This study thus considers an EC as a position with an unarticulated determiner phrase (DP). The basic assumption is that when a lexical item is moved in a sentence, it leaves a copy of itself in the extraction site referred to as a trace (Chomsky, 1981; Radford, 1997; Otsuka, 2001). Moreover, covert NPs ECs are divided into four types namely NP-trace, wh-trace, PRO, and pro (Chomsky, 1981, 1982; Otsuka, 2001). According to Holmberg (2005, p. 535) “the grammatical properties of ECs are ‘functionally determined’, that is, they are determined by the syntactic relations that it enters into, particularly the binding relation.”Therefore, an empty NP qualifies to be awh-trace if it is A' bound, an NP-trace if it is A-bound from a non-ϴ-position, PRO if it is A-bound from a Ө-position, and pro if it is governed by rich inflectional morphology (Holmberg, 2005; Ouhalla, 1999). The typology of ECs mirrors that of overt ones. Therefore, four types of ECs namely, NP-trace, WH-trace, PRO and pro satisfy the four types of expressions predicated by Chomsky’s binding conditions. However, this paper only focuses on the first condition which is relevant to the discussion on NP-movement. Condition (a) states that: ‘An anaphor must be bound in its governing category (GC)’ is satisfied by an NP-trace because it is a pure anaphor that must be bound in its GC following condition A of Binding theory. An NP-trace is therefore, the empty counterpart of lexical anaphors, for example, ‘each other’ and ‘herself’ 3. The syntax of NP-Trace When a DP moves from an argument position (A-position) to another A-position, for example, from object to subject position, a gap is left in the position from which a category has been moved (Aroonmanakun, 1997; Radford, 1997, Carnie, 1999; Haegeman, 1998). The gap created by NP-movement is referred to as an NP-trace. In practice, an NP-trace is represented as t, and the empty node as e. Koot and Neeleman (2010) hypothesize that the NP-trace plays an important role in A-movement. This is in tandem with Chomsky’s argument that the subject of a verb in the passive voice moves to the position that was initially occupied by the subject of the active sentence (Chomsky, 1981). Secondly, Chomsky (1981) and Carnie (1999) hold that all subjects move from a thematic position to a designated position external to a verb phrase (VP). That is, the VP’s external Ө-role is always introduced by the verb in active constructions, but in passive and raising constructions it is introduced by an NP-trace. Since the subjects of passive verbs satisfy the relevant external Ө-role of the VP, they are equivalent to subjects of active verbs. An NP-trace arises only at surface structure (SS) (Haegeman, 1992). Further, it can be found either in a passive sentence or in a sentence with a raising verb (Radford, 1997). In NP-movement, one constituent is substituted for another empty one; therefore, a category must be moved to a landing site which has the same property as the moved category. Carnie (1999) observes that an NP-trace results from a transformation that moves an NP to the specifier position of an inflectional phrase that is [Spec, IP]. Carnie (1999), Lasnik and Uriagereka (1988), and Cowper (1992) posit that in A-movement, an NP moves from a position where an NP receives a Ө-role to one in which no Ө-role is assigned due to the fact that an NP-trace cannot satisfy a Ө-role. Harford (1985) notes that Bantu languages permit NP-movement from the subject position of a tensed clause. Furthermore, Bantu languages license NP-trace in either passive orraising constructions as analyses of NPmovement in generative syntactic theories and the Bantu passive is realized by a suffix which is an argument reducing morpheme (Zeller, 2011). Schadeberg (1995) notes that the suffix has been reconstructed in Proto-Bantu as /*u/ after consonants and /*ibu/after vowels. In addition, Ngonyani and Githinji (2006) observe that the preferred position for suffix placement is the final position as shown in the examples in (1) from Kikuyu cited in from Ngonyani and Githinji (2006, p. 49): 1. (a) Mũ-geni a- ra- gũ- r- ire cianamū-bira 1-guest 1SM-PRG-buy-AP-PF 8-child 3-ball ‘The guest bought children a ball.’ (b) Cianaci- ra- guȓ- ĩ- w- o mũ-biranimũ-geni 8-child 8SM-PRG-buy-AP-PASS-FV 3-ball by 1-guest ‘The children were bought a ball by the guest.’ 104 International Journal of Language and Linguistics Vol. 2, No. 4; October 2015 The examples in (1) show that the logical object is realized as the subject of a sentence in NP-movement because the thematic subject no longer functions as a grammatical subject but is realized as the complement of a preposition. In LuGanda, NP-trace results from a moved overt NP (Abudonia, 2014). This is illustrated in the example (2) from Abudonia (2014, p. 44): 2. A-ba-jaasibaagabwa e- bi- kwekweto IV-2-soldiers 3SMpl-PST-distribute- PASS IV- 8- group [IP A bajaasii [VP baaga- [I´-bwat (i)ebikwekweto]]] ‘The soldiers were distributed in groups.’ In example (2), the derived subject that is the moved NP, c-commands the trace which is assigned a theta role by the verb. Such illustrations on NP-trace based on other languages such as Kikuyu and Luganda offer insight into the general characteristics of Bantu languages and are beneficial in the present study to find out if EkeGusii bears the same characteristics as well as defining the passive marker in the language. The examples given from Kikuyu and LuGanda illustrate that NP-traces are considered as anaphors because they bear the feature [-p, +a] just like anaphors and also occupy an A-position in agreement with Chomsky’s 1981 proposal. The foregoing discussion on the comparison of the properties of NP-trace and its overt equal thus offers a pre-analysis into NP-trace in EkeGusii and it is projected to bear almost similar characteristics because it is a Bantu language too. It is therefore important to bring data on EkeGusii on board so as to evaluate the universality of the overt NPs and their covert counterparts. 4. Theoretical Framework Data was analyzed for properties of NP-Trace in EkeGusii using Government and Binding theory (GB). This is because GB tries to explain a Universal Grammar (UG), which describes all languages since it assumes that a large portion of the grammar of any language is common to all languages (Black, 1999). The GB Theory of syntax is modular in nature (Chomsky, 1981, 1982, 1986; Epstein, 1991; Haegeman, 1994; Carnie, 1999; and Ouhalla, 1999), but specifically five of its subsystems namely, X-bar theory, Government theory, Binding theory, Trace theory, and Theta theory. The modules of GB interact with one another as posited by Chomsky (1981). Therefore, the researcher employs them concurrently in the presentation and analysis of the data for the study. For instance, in a passive construction, by using either a parse tree or bracketed notation for the construction, X-bar theory comes at play since each clause has a head category. In addition, by co-indexing the anaphoric arguments, Binding theory, Government theory and Trace theory come to play. Going further to explain the semantic role of arguments in the given construction such as agent, theme, subject, and object is licensed by Theta Theory. Ideally, the GB theory is intertwined so no single discussion for the licensing conditions of a given construction will rely entirely on one module of GB theory. 5. Methodology The main objective of this paper is to examine the properties of NP-Trace in the EkeGusii sentence. The specific objectives that guided the analysis of this paper were to: 1) identify and analyze properties of NP-Trace in the EkeGusii sentence and (2) compare and contrast the properties of NP-Trace in the EkeGusii sentence and those of its overt counterpart (anaphor). Data for the study was elicited through recording radio call-in programmes and crosschecking with native speakers by subjecting them to essay writing. Besides, the researcher used native speaker intuition to generate further sentences containing ECs. The researcher engaged multiple methods in data collection so as to attain more valid, reliable and diverse construction of realities in this paper. The paper deals with non-overt NPs, particularly the NP-trace to show how far the Principles and Parameters approach is applicable in this respect. In this section, NP-trace in EkeGusii is discussed in depth. 6. Results and Discussion The purpose of this study was to identify and discuss the properties of NP-Trace in the EkeGusii sentence. Having the features [+anaphoric, -pronominal], NP-trace is subject to Principle A of the Binding Theory. Various theories intervene in the analysis of the NP-trace as an EC since it is left behind by a moved overt or null NP it is studied within the framework of binding theory, bounding theory, theta theory and government theory. 105 ISSN 2374-8850 (Print), 2374-8869 (Online) © Center for Promoting Ideas, USA www.ijllnet.com The analysis revealed that NP-movement occurs in various constructions namely, the passive of transitive constructions, the applicative, the causative, and the raising constructions. Therefore, the purpose of this section is to present an analysis of ECs created by NP-movement in such constructions so as to discuss their syntactic properties, which is the focus on objectives of this paper. 6.1 NP-movement in EkeGusii In this section, the occurrence of NP-Trace in each of the named construction is presented in detail. 6.1.1 The Passive in EkeGusii The study maintains that an NP-trace is the result of a moved NP and that the moved NP satisfies the Subjacency condition. A passive construction in EkeGusii is characterized by the presence of the morpheme /-u/ which is realized in various phonologically conditioned forms such as /bw/, /gw/ and /kw/ (Elwell, 2008; Ongarora, 2008). Moreover, this passive construction is derived by NP movement which leaves a trace at the extraction site. 6.1.2 NP-Trace in Transitive Constructions Deen (2005) notes that an NP-trace does not alternate with overt NPs because it occupies a ϴ-marked position. Further, he concurs with Haegeman’s (1992) opinion that an NP-trace must be antecedent-bound so as to satisfy the ECP which states that: All traces must be properly governed (Chomsky, 1982, p.21; Haegeman, 1992, p. 22). Haegeman (1994) observes that NP- movement creates an A-chain resulting in co-indexation of the moved NP with its trace. The antecedent NP, rirube, therefore A-binds the trace, the NP, ti. The traces of NP-movement are referred to as NP-traces (Haegeman, 1994; Radford, 1997; Carnie, 1999).This occurs in EkeGusii as indicated in (3): 3. Ri-rubeiri- a- riik- w- ati 5-Letter 3SG-PST-write-PASS-FV ‘A letteriwas writtenti.’ In agreement with the position about movement of NPs taken by Radford (1988), that is, from object to subject position, in the examples in (3) the NP, rirube, originates from the post verbal object NP position and is subsequently moved into a preverbal subject NP position. Through this, an empty category which is indicated by tiis created. In example 4, the moved NP c-commands the trace which is assigned a theta role by the verb thus they form a chain whose head that is the derived subject is case marked. Therefore, the moved NP does not have a theta role since passive verbs cannot assign a theta role to their external arguments (Burzio, 1986). Moreover, the movement does not violate the Subjacency Condition because movement of NPs in EkeGusii satisfies the theta criterion as the chain formed by the moved NP and its trace is assigned one theta role. Consequently, the moved NP is assigned the nominative case whereas the trace is assigned the role of the patient. Active sentences have corresponding passive derivations (Haegeman, 1994; Carnie, 1999 and Radford, 1997). Therefore, a passive construction is derived by transformational rules from a corresponding active sentence. Passive sentences are hence generated by phrase structure (PS) rules as complements of the verbs that Ө-mark them. The complements are then moved to occupy the subject position of the passive verbs. As reported in Mchombo (2005), the suffix changes the status of a subject NP to oblique status and introduces a preposition into the construction. This occurs in EkeGusii, for example, in (4): 4. (a) Marube a- Ø- gor-a chi-sani Marube 3SG-PST-buy-FV 10-plate ‘Marube bought plates.’ (b) Chi-saniichi- a- gor- w- atina Marube 10-plate 3PL-PST-buy-PASS- FV by ‘Marube‘Platesi were boughtti by Marube.’ The tree diagram of (4b) is drawn in example (5) below: 5. 106 International Journal of Language and Linguistics Vol. 2, No. 4; October 2015 The tree diagram in (5) shows that sentence (4b) is derived via movement of the NP, chisani, from the complement position in the VP into [Spec, IP] so as to function as the specifier to the passive participle, gorwa. The moved constituent, therefore, leaves a trace (t) which must be bound by an argument, chisani, in the same governing category.In the above examples, the moved NP c-commands the trace which is assigned a theta role by the verb thus they both form a chain. The head of the chains, the moved NP, is case-marked but it does not have a theta role. Haegeman (1994) argues that according to Transformational Generative Grammar that the transformation of a sentence from active to passive status inverts the logical function of the object and subject to grammatical subject and object respectively. The study thus observes that active and passive sentences do not mean exactly the same thing EkeGusii as illustrated in (6): 6. (a) Nyanchera o- Ø- gor- aebi-tabu Nyanchera 3SG-PST-buy-FV 8-book ‘Nyanchera bought the books.’ (b) Ebi-tabui bi- a- gor- w- atinaNyanchera 8-book 3PL- PST-buy- PASS-FV by Nyanchera ‘The booksiwere boughtti by Nyanchera.’ Sentence (6a) is about the subject NP, Nyanchera, while (6b) is about ebitabu. However, they both describe the same participants in an event hence the act of buying is described whereby the buyer (agent) is Nyancheraand the item bought (theme) is ebitabu. Hence, the surface structure (SS) of the two sentences involves the same thematic information and argument structure. Moreover, Williams (1995) posits that the passive induces internalization of the subject ϴ-role (argument) as evident in example (6). 6.1.3 NP-Trace in EkeGusii Applicative Constructions The applicative suffix in EkeGusii has been reconstructed as /-er/. As is the case in other Bantu languages, the applicative suffix allows the verb to take two unmarked objects. This thus licenses introduction of an object which functions as the locative, instrumental or beneficiary in EkeGusii. A widely attested strategy in Bantu languages is that of inserting an indirect object resulting in alteration of the argument structure of the verb since it increases the number of NPs that can occur in a sentence (Pylkkänen, 2002; 2008). Consequently, an example of a mono-transitive verb in EkeGusii, gora, has been used applicativized with the suffix /-er/ as in (7b): 7. (a) Nyanchera o- Ø- gor- a e-ganchwa Nyanchera3SG- PST-buy-FV 9-dress ‘Nyanchera bought a dress.’ (b) Nyancherao- Ø- gor- er- a Moraa e-ganchwa Nyanchera 3SG- PST-buy-APPL-FV Moraa9-dress ‘Nyanchera bought a dress for Moraa.’ Sentence (7a) has a mono-transitive structure since it contains only one object argument, eganchwa. It has however been made di-transitive by the introduction of an applicative suffix, /er/ so as to license a second object argument, Moraaas illustrated in (7b). 107 ISSN 2374-8850 (Print), 2374-8869 (Online) © Center for Promoting Ideas, USA www.ijllnet.com There are two types of applicatives as hypothesized by McGinnis (2001) and Pylkkänen (2008) namely, high (E) applicatives, and low (I) applicatives. The Bantu applicative denotes a relation between an event and an individual, hence it is a 'high' applicative because the Bantu ApplH consists of a VP complement and a DP specifier as in 9 (Ongarora, 2008, p. 64): 8. Moreover, the EkeGusii construction in (7b) represented as (9) is in support of the Ngonyani and Githinji (2006) finding that the Bantu applicative can be analyzed as a containment whereby the higher vP selects and contains the low VP (Ongarora, 2008). 9. Ongarora (2008) observes that in some Bantu languages, both objects behave alike when subjected to object order, passivization and object marking.EkeGusii is thus a symmetrical language because either object can raise to subject position in the process of passivization.For instance, the applicative argument in (7b) can be passivized in EkeGusii as in example (10): 10. Moraai o- Ø- gor- er- wati e-ganchwanaNyanchera Moraa 3SG- PST-buy- APPL-PASS- FV 9-dress by Nyanchera ‘Moraa was bought a dress by Nyanchera.’ As shown in (9), the movement of the beneficiary to subject position and from there to the specifier position creates a chain in which the NP at the specifier binds the trace of the benefactive in its D-structure position. In this case, the internal argument, that is, the indirect object of the ditransitive verb is promoted to subject position thus the trace that occupies the empty node, ti, and the subject NP, Moraa, are co-indexed since they co-refer. In the passive of a benefactive applicative, either the beneficiary or the theme moves to subject position as in (11a) and (11b), respectively. 11. (a) Omo-ng’inai o- Ø- riker- w- atiriruben’omwana 1-woman 3SG- PST- write- APPL-PASS-FV letter by 1-child ‘The womani was writtenti a letter for by the child.’ (b) Rirubeiri- a- riker- w- a omo-nginatin’omwana 5-Letter 3SG-PST-write-APPL- PASS- FV 1-woman by 1-child ‘The letteri was written for the womanti by the child. 6.1.4 NP-Trace in EkeGusii Causative Constructions The causative is a valence changing morpheme which introduces the ‘causer’ or ‘agentive argument to the verb which attaches to an intransitive verb thus altering the number of arguments it can license by introducing either the Causer NP or one bearing the agent ϴ-role (Pylkkänen, 2002). 108 International Journal of Language and Linguistics Vol. 2, No. 4; October 2015 The causative suffix in EkeGusii is construed as /-i/. For instance, in example (12), the non-causative verb minyoka takes the suffix /-i/ so as to license the agentive NP, esese, in the construction. Additionally, introduction of the causative suffix licenses an additional argument, omogeni, in example (12). 12. (a) Omo-geni o- Ø- minyok- a 1-visitor 3SG-PST- runFV ‘The visitor ran.’ (b) E-seseya- a- minyook- i- a omo-geni 9-dog 3SG-PST- run-CAUS-FV 1-visitor ‘The dog made the visitor run/chased the visitor.’ In EkeGusii, the passive morpheme can occur in constructions with causatives. For example, the sentences in (12b) can be passivized as in (13): 13. Omo-genii o- Ø- minyook- igw- a tina e-sese 1-vistor 3SG- PST- run- CAUS- PASS- FV by 9-dog ‘The visitori was made to run awaytiby the dog.’ The external argument (Agent), esese in (12b) is dropped in passivization as in (13) and its position is filled by the internal argument (Theme), omogeni. In the present study, the researcher adopts Pylkkänen’s (2002; 2008) theta role analysis because when constructions with a Causer ϴ-role are passivized, the external argument may be implied but not diagnosed through a by-phrase in the case of short passives (14b). Besides, causative constructions may be passivized such that the external argument is usually implied via a by-phrase as is the case in long passives as in (14a). 14. (a) E-saniiya- a- sib- i- gw- atina Mosota 9-plate 3SG-PST- wash-CAUS-PASS- FV by Mosota ‘The platei was washedti by Mosota.’ (b) E-saniiya- a- sibigwati 9-plate 3SG- PST-wash- CAUS- PASS- FV ‘The platei was washedti.’ In a nutshell, EkeGusii takes two forms; the long passive, an instance where the agent is present as illustrated in (14a), and the short passive whereby the agent is absent as in (14b). To account for the passive as NP-movement, the study concludes that the EkeGusii passive morpheme /u/ receives accusative case of the Theme object of an active sentence. For this reason, it removes the ability of the passive participle to Ө-mark its subject as opined by Radford (1997), thus, allowing the NP to move into [Spec, IP]. The Agent argument is then realized as part of an adjunct prepositional phrase (PP) headed by the passive morpheme /u/. Example (15) shows a passive construction in EkeGusii which involves the movement of a null object. 15. (a) Ba- a- rager- igw- a ba- a- igot- a 3PL-PST-eat- PASS- FV 3PL-PST-fill up- FV ‘They were fed and got filled up’ (b) Tu- a- ee- gw- a 2PL-PST-give-PASS-FV ‘We were given’ In example (16), it is observed that the object markers are dropped and moved to the position which was occupied by the subject markers in the deep structure (DS). 16. proba- ba- rager- ia proba- a- igot- a 3PL-PST- eat- CAUS- FV 3PL-PST- fIll up- FV ‘They made them eat and were filled up’ A moved null object does not leave a trace behind, rather the null object moves to replace the subject marker within the verb phrase and the moved null object is assigned nominative case by INFL and the theta role is incorporated in the verb phrase due to the fact that passive verbs cannot assign a theta role to their external arguments. Moreover, movement of null NPs results in the replacement of the subject marker and null subject with the object marker and null object respectively. 109 ISSN 2374-8850 (Print), 2374-8869 (Online) © Center for Promoting Ideas, USA www.ijllnet.com 6.1.5 NP-trace in EkeGusii Raising Constructions Otsuka (2001) observes that raising is a syntactic operation whereby an NP moves from one argument position to another so as to be assigned Case. Therefore, an NP moves from a non-Case position to a Case position so as to check its Case feature. Thus, raising is restricted to subjects because other arguments such as direct objects would involve movement from a Case position. As observed by Radford (1997), the subject of a raising construction originates in a ϴ-marked specifier position within VP, and then moves into a Case-marked specifier position within IP. Therefore, such a subject or nominal originates as the subject of the complement clause and is then raised to become the matrix clause subject by application of raising. This leaves behind an EC trace to serve as the subject of the complement clause. In Carnie (1999), a raising verb is one that allows NPs to move from positions where they are assigned Ө-roles to those in which they acquire Case instead of Ө-roles. Such movement results in presence of a trace in the empty node. This is illustrated by the EkeGusii example (17): 17. (a) Eko-rorekana aba-isekeba- a- nch- et- e chi-nyange Thing-seem 2-girl 3PL- PST- like- PFV- FV 10-bicycle ‘It seems that girls like bicycles.’ (b) Aba-isekeiba- kororekanatiba- a- nch- etechi-nyange 2-girl 3PL-seem 3PL- PST- like- PFV- FV 10-bicycle ‘Girlsi seem notti to like bicycles.’ In example (17a), the predicate holds the proposition that the Extended Projection Principle (EPP) can be satisfied with an expletive when the embedded clause is finite while (17b) is a derivation of (17a) because (17b) is subject to the ECP which requires that movement takes place from the lower COMP to the higher COMP ([Spec, CP]) in a non-finite clause, leaving a trace in the extraction site which is co-indexed with its antecedent. Moreover, Harford (1985) opines that most Bantu languages license hyper raising constructions where the logical subject of an embedded finite clause is realized as the matrix subject or object. 18. 110 International Journal of Language and Linguistics Vol. 2, No. 4; October 2015 From the illustration in (18) it is evident that the subject, abaisekeoriginates within the subject of the embedded predicate where Ө-marking takes place and is further raised to [Spec, IP]. Consequently, the NP raised to [Spec, IP] still bears the already assigned Ө-role because such a position only assigns Case to an argument. Moreover, the moved NP had already been Ө-marked before movement. Such movement is thus referred to as raising because the NP moves from a lower to a higher clause. Considering the foregoing discussion, it can be concluded that in both passivization and subject raising, an NP is moved from one A-position into another. 6.2 Conclusion The paper has examined NP-trace in EkeGusii which conforms to Chomsky’s typology of non-overt NPs in particular the NP-trace and its overt counterpart (anaphor). The paper has shown that NP-trace in EkeGusii is realized through passivization of various constructions namely, transitives, applicatives and causatives. The NPtrace is as well realized in raising constructions in EkeGusii. The passive construction in EkeGusii is characterized by the presence of the morpheme /-u/ which is realized in various phonologically conditioned forms; /bw/, /gw/ and /kw/. The study concludes that a moved overt NP leaves a trace in the base position whereas a moved null NP replaces the null subject and that the extraction site is assigned nominative case by INFL. Besides, anaphors in EkeGusii are non-overt NPs which are encoded in the morphology of the verb. Lastly, EkeGusii is an agglutinating language which allows dropping of the overt object because there are object markers in the verb phrase that bear the nominal features. References Abudonia, R. (2014). Overt and non-overt NPs in Luganda. International Journal of Language Studies 8(1), 3348. Aroonmanakun, W. (1997). Referent resolution for zero pronouns in Thai. In A. S. Abramson (Ed.), Southeast Asian linguistic studies in honour of VichinPanupong (pp. 11-24). Chulalongkon University Press. Black, C. A. (1999). A Step-by step Introduction to the Government and Binding Theory of Syntax. Summer institute of Linguistics. Available:http://www.silorg/mexico/lingE002-Introgb.pdf(September 2, 2011). Burzio, L. (1986). Italian syntax. Dordrecht: Reidel. Carnie, A. (1999). Sentence structure:A generative introduction. Tucson: University of Arizona. Chomsky, N. (1981). Lectures on government and binding. Dordrecht: Foris Publications. Chomsky, N. (1982). Some concepts and consequences of the theory of government and binding. Cambridge, Massachusetts: MIT Press. Chomsky, N. (1986). Barriers. Cambridge, MA: MIT Press. Cowper, E. (1992). A concise introduction to syntactic theory: the government and bindingapproach. Chicago: CUP. Deen, K. Ud. (2005). The acquisition of Kiswahili. Amsterdam: John Benjamins Publishing Co. Elwell, R. (2008). Finite-state methods for Bantu verbmorphology. In N. Gaylord, A. Palmer and E. Ponvert (Eds.), Computational Linguistics for less-studied languages (pp 55-57). Stanford: CSLI. Available: http://www.robertelwell.info/ papers/ elwellekegusiifst. pdf (February 4, 2011). Epstein, D. S. (1991). Traces and their antecedents. New York: Oxford University Press. Guthrie, M. (1971). Comparative Bantu: an introduction to the comparative linguistics prehistory of the Bantu languages.Vol 2. Farnborough: Gregy Press. Haegeman, L. (1992). Theory and description in generative syntax: A casestudy of West Flemish. New York: Cambridge University Press. Haegeman, L. (1994). Introductionto government and binding theory (2nded.). Oxford: Blackwell Publishing Ltd. Haegeman, L. (1998). Elements of grammar. Dordrecht: Kluwer. Harford, P. C. (1985). Aspects of complementation in three Bantu languages (Unpublished doctoral thesis), University of Wisconsin. Holmberg, A. (2005). Is there a little pro?: Evidence from Finnish. Linguistic Inquiry, 36(4), 533-564. Available: http://muse.jhu.edu/journals/lin/summary/ v036/36.4holmberg. html (October 12, 2011). Koot, H. &Neeleman, A. (2010). A local encoding of syntactic dependencies and its consequences for the theory of movement. Syntax, 13(4), 337-372. Available: citeseerx.ist.psu.edu/viewdoc/download?doc=10.1.1.110… (January 12, 2012). 111 ISSN 2374-8850 (Print), 2374-8869 (Online) © Center for Promoting Ideas, USA www.ijllnet.com Lasnik, H. &Uriagereka, J. (1988). A course in GB Syntax: Lectures on binding and empty categories. Cambridge: MIT Press. McGinnis, M. (2001). Phases and the syntax of applicatives. In Minjoo Kim and Uri Staruss (Eds.), Proceedings of NELS 31. Amherst, MA: GLSA. Available: www.uni. Leipzg.de/~georgi/Lehre/FolienSoSe11/McGnnis00.pdf (November 3, 2011). Mchombo, S. A. (2005). The syntax of Chichewa. Cambridge: CUP. Ngonyani, D. &Githinji, P. (2006). The asymmetric nature of Bantu applicative constructions. Lingua, 116 (1), 31-63. Ongarora, D. O. (2008). Bantumorphosyntax: A study of EkeGusii (Unpublished doctoral thesis), Jawaharlal Nehru University. Otsuka, Y. (2001). Empty categories in Tuvaluan: NP-Trace, PRO, pro or variable? Oceanic Linguistics, 40(2), 342-365. Available: http://www.jstor.org/stable/3623446 (October 21, 2011). Ouhalla, J. (1999). Introducingtransformational grammar: From principles and parameters tominimalism. Oxford: Oxford University Press. Pylkkänen, L. (2002). Introducing arguments (Doctoral thesis, MIT). Available:http://www.people.fas.harvard.edu/~ctjhuang/NTNU/Pylkkanen_2002.pdf (January 10, 2012). Pylkkänen, L. (2008). Introducing arguments, linguistics inquiry monograph, vol. 49. The MIT. Radford, A. (1988). Transformational grammar: A first course. Cambridge: CUP. Radford, A. (1997). Syntax: A minimalist introduction. New York: Cambridge University Press. Schadeberg, T. C. (1995). Object diagnostics in Bantu. In: E. N. Emenajo and O. Ndimele (Eds.) Issues in African languages and linguistics: Essays in honour of Kay Williamson (pp 173–180). Aba: National Institute for Nigerian Language. Williams, E. (1995). Thematic structure in syntax. Cambridge: MIT Press. Zeller, J. (2011). The syntax of African languages: A review. Available: http:// www.jzeller.de/ pdf/Afri%20Syn.pdf (September 8, 2011). 112