On the syntactic distribution and morphological form of resumptive pronouns in Esan

I argue in this paper that in Esan (ISH) [Edoid, Benue-Congo, Niger-Congo: Nigeria], the co-indexed pronouns in constructions such as the following represent resumptive pronouns. (a) Post-nominal pronoun construction (PNPC): Mẹ i ọ i rẹ khian giẹgie do igho mu bhi ibank 1.SG 3.SG although FUT quickly steal money carry LOC bank “I was going to quickly steal money from the bank, (but...)” (b) Relative clause (RC): eni awa i [ni e i kpọlọ] [ni e i mọnsẹ] [ni e i gian] DEF.PL dog REL 3.PL be big REL 3.PL be beautiful REL 3.PL be red “the beautiful big red dogs” In the constructions above, a pronoun is found in a lower clause, following a nominal with which it is co-referential. This pronoun is invariably third person, whose semantic contribution is not apparent. Starting from such data, this investigation has two main goals: (1) a basic description of personal pronouns and resumptive pronouns (RPs), and (2) an analysis of the syntactic distribution and morphological form of these resumptive pronouns. In the analysis, I argue that resumptive pronouns are true nominal arguments and not a manifestation of a predicational agreement system (i.e. an agreement marker). Evidence for this comes from (1) the pronoun is not obligatory in all finite clauses which would be highly anomalous if it were a part of any verbal conjugation, (2) a contrast between a full/strong form and a reduced/weak form of a pronoun is still available, and (3) in post-nominal pronoun constructions, for some speakers the resumptive pronoun provides an implication of particular attention given to the co-referential nominal, suggesting topicalization. Topicalization is also suggested from the left peripheral position of the co-indexed nominal (Rizzi 1997, Ermisch 2007). I formalize this evidence in a series of diagnostics, using the related Edoid language Ivie, which manifests an agreement system, as a basis for comparison (Emuekpere-Masagbor 1997). After establishing these pronouns as arguments rather than agreement markers, I argue that they occur when a nominal is extracted from subject position (i.e. spec-IP) to a position in a higher clause (e.g. spec-TopP), such as in a post-nominal pronoun construction (a type of topicalization) or in a relative clause, leaving a trace (Chomsky 1995). This movement creates an A-bar chain between the extracted nominal and its trace (Cinque 1990). When this trace is in subject position, it is realized as a co-indexed resumptive pronoun (i.e. an overt trace; Koopman & Sportiche 1986), resulting from a structural subject requirement. I formalize this requirement as a particularly strong manifestation of the Extended Projection Principle (a la Chomsky 1995: 232; Adesola 2005: 102). This structural subject requirement is corroborated by independent evidence in the language against a phonologically empty subject position including (1) a lack of pro-drop (2), the use of expletives and dummy subjects, (3) the presence of an impersonal subject in negative imperatives, and (4) the raising of objects to subject position in certain causative/existential clauses involving ri bhi ‘to put at/to be at’. The form of the resumptive pronoun is dependent upon the morphosyntactic featural composition of the topic with which it is referential. It is shown that only third person plural nominals are co-referential with e ‘they’. All other nominal or pronominal topics co-occur with singular ọ ‘he/she/it’, including seemingly plural mhan ‘we’ and bha ‘youpl’ (e.g. Bhai oi gbikhiẹn “You(all)i hei did dance”). I argue that this distribution falls out from the feature geometry of the pronominal inventory (following Harley & Ritter 2002, among others). Under the [Number] feature node, non-singular pronouns mhan ‘we’ and bha ‘youpl’ are specified as [Mass], whereas non-singular pronoun e ‘they’, as well as plural nominals, are specified as [Group]. This division will ensure that e ‘they’ resumes only the position of those nominals extracted from subject position with the feature [Group], i.e. plural nominals. Because mhan and bha do not have this feature, they are therefore resumed with the most unspecified default form ọ 3.SG. The insertion of the form of the resumptive pronoun is formalized under the constraints of distributed morphology (DM) (Halle & Marantz 1993). Such an approach “ensures that the Vocabulary item that matches the most features of the node will be inserted” and that a pronoun may be “underspecified for the morpho-syntactic feature complexes that they realize” (Halle & Marantz 1993: 121-122). Finally, I discuss why resumptive pronouns in Esan are invariably third person, that is, why there is feature matching with respect to number but not for person, and how this compares to feature matching in resumption viewed cross-linguistically. Crucially, if we understand these resumptive pronouns as spelling out the syntax, then it entails that there are no [Person] features present in the trace position left by the extracted nominal, otherwise they would be pronounced. This suggests that in Esan, extraction/movement of a nominal neutralizes any person specification (i.e. all extracted nominals are treated as third-person).

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.