Negation of the Object in Sesotho

This article examines constituent negation in Sesotho. It investigates negation of the object argument in clauses with non-copulative verbs. In the case of non-copulative verbs, transitive and di-transitive constructions will be considered. The article argues that Sesotho does not have a direct means of negating clausal constituents but employs clauses such as cleft sentences and clauses with AGRS [ho] as well as contrastive clauses to effect negation. It further examines negation in terms of Haegeman’s (1995) Neg-Criterion, the well-formedness condition that determines the distribution and interpretation of negative elements. It gives a brief overview of grammatical concepts regarding negation and the notion object argument in its focus and post verbal positions, and explains how constituent negation, [the object argument] is realized in Sesotho.


Introduction
Negation as described by Zanuttini (1997) is a syntactic process whereby a language employs negative markers to negate a clause. Haegeman (1995) defines negation in terms of Neg-Criterion, the well-formedness condition that determines the distribution and interpretation of negative elements. This criterion asserts that a negative operator must be in a spec-head configuration with a negative head. Based on the Neg-Criterion, negative sentences are constructions which minimally have a negative feature, associated with a functional head of the extended projection of the verb. This implies that negative elements are identified as Negative Phrase headed by a negative head, where the head is in geometrical position with its specifier. In Sesotho the negative phrase will have negative morphemes /ha/, /sa/ and /se/ as heads.
Negation is further distinguished into sentence negation and constituent negation. According to Klima (1964), sentence negation is realized when negative pre-verbal particles are present in the sentence. Haegeman (1995) goes on to state that negative constituents that trigger subject inversion, have sentential scope and are referred to as 1. The notion object Anderson (1984) defines an object as a noun or pronoun denoting a person or a thing affected by the action expressed by the verb, where a verb becomes transitive because its action goes over to an object. The notion object also refers to that NP which is immediately dominated by VP. Anderson (1984) goes on to state that objects are a variable subset of the non-subjective absolute phrases and there may be more than one object in a simple sentence. According to Moravcsik (1984), objects refer to some controlling participants of a two-participant event that constitutes in its entirety the primary target for the event.
According to Collinge (1984), objects are identified through their categorical operators, semantic diagnostics and control properties. With regard to categorial operations, an NP is a simple operand noun and it can be placed in second highest position. They can be controlled by three semantic features such as definiteness, affectedness and animacy, and control properties are considered to be accessibility, omission and deletion. Moravcsik (1984) argues that direct objects have features that are general NP properties. These general properties include internal composition, head constituent relations and pragmatic function. In the case of internal composition, the head of a direct object NP may be a noun, adjective, quantifier, pronoun, infinitive or a clause. An object phrase with a nominal head can accommodate a wider range of satellites such as determiners and quantifiers. Objects may also be represented by pronominals. With regard to constituent relations, Moravcsik (1984) maintains that the semantic relation between verbs and objects is a headmodifier relation in that the verbobject construction is a subtype of the predicate expressed by the verb itself, where there is a unidirectional relation between the verb and the object, e.g. a verb may require the presence of an object complement for semanticsyntactic well-formedness. Finally, objects may take basic pragmatic functions of topic, focus and pragmatic neutrality. NP objects are however unique NP's on the basis of specific argument function they fulfill, coupled with association with a particular case marker.
On the other hand, Gil (1984) stipulates specific characteristic properties of direct objects in a basic sentence: a) NP object must have a thematic role property of patienthood. b) An NP object should have pragmatic reference property of lesser referential strength, i.e. direct objects are of low referentiality, they occur at the bottom of a grammatical relation quantifier scope hierarchy. c) An NP object must have pragmatic reference property of indefiniteness. Direct object position is a preferred position for introducing indefinite arguments. According to Lazard (1984), nominal categories of objects are variations such as humanness or animacy and definiteness. There are three grammatical persons which are not logically symmetrical, viz. third person may be anybody or anything but mostly non-human, first and second persons rank higher in the hierarchy of humanness and are mostly human. d) Direct objects occur to the right of subjects in the basic word order of almost all languages. In other languages, the relative word order of the NP's in a clause is quite free. e) Direct objects must have a clausal property of transitivity, which serves as a diagnostic for direct-objecthood, since the presence of a direct object is a characteristic feature of transitive clauses. With regard to the structural position of the object and case licensing, Haegeman (1995) states that there is an agreement projection responsible for checking agreement features and accusative case of the object which is Agr0P. The agreement features of objects are licensed at LF if the features are weak and before spell out if they are strong. The object argument has two structural positions; viz. the base position within the VP where it obtains its theta-role and the derived object position above the VP, the [Spec, Agr0P], the landing site of the object after movement responsible for accusative case licensing. According to Ouhalla (1999), accusative case is assigned under Spec-head agreement with a complex The idea of a two structural object position can be clearly illustrated by a transitive Sesotho sentence as adopted by Du Plessis (1995) for African languages as shown in (2) and the diagram in (1) below.
(2) Ngwana o bala buka If the object clitic is present, which is possible in Sesotho, the object will be raised overtly before spell out. After all the categories have been raised, the structure will be as in (2) below:

The object in Sesotho
Sesotho is a SVO language where the object always follows the verb. It has both direct objects and indirect objects which may appear as overt categories or as null objects in the sentence. Overt objects are lexical NP objects which are phonologically represented in post verbal positions. As nominals, objects in Sesotho appear with nominal modifiers and they are closely associated with verbs. It will therefore be proper to deal with them in relation to various kinds of verbs, such as intransitive, monotransitive and ditransitive verbs.
2.1.1 Intransitive verbs While objects normally appear with transitive verbs, in Sesotho there are certain NP objects which appear with intransitive verbs as non-arguments. Such NP's, according to Du Plessis and Visser (1995), are cognate objects and idiomatic objects. Cognate objects may be non-argument NP's which are referred to as adjuncts: ( 3)  There are however cognate objects which are full arguments as indicated by a sentence in (4) below: With regard to idiomatic objects, Du Plessis and Visser (1995) state that there are also intransitive verbs which appear with adjunct phrases that give rise to idiosyncratic meaning. These verb phrases are usually found in idiomatic expressions as in (5) (1995), ditransitive verbs subcategorize for two internal arguments viz. the indirect object and the direct object. According to Ramone (1992), these two objects are grammatically and semantically dissimilar. Structurally they both occur after the verb, i.e. one immediately after the verb and the other in the position further away from the verb. a) Direct Objects According to Ramone (1992), in double object constructions, direct objects appear on the periphery of the sentence and they mostly have the semantic interpretation of theme: c) Applied objects Ramone (1992) describes applied object as the extra argument of the applied form of the verb. The applied object is dependent on the affix [-el -]. It is adjacent to the verb and has theta roles of beneficiary, purpose or cause: (13) (1995) objectival concords occur as prefixes of the verb in morphology without lexical NP objects. In such cases the objects are referred to as null or empty NP objects and they are represented phonologically by empty pronominals indicated as [pro]. Idiophones can also appear with an empty [pro]. The sentences in (14) with the diagram in (3) will illustrate: The woman she him divorced [pro] [ In passive sentences, the object may be moved to the subject position as indicated in (18)

Negation of the object with non-copulative verbs
Negation of the object argument will be considered within transitive and ditransitive verbs. Three possible ways of negating the NP object will be examined, viz. negation of the object in postverbal position, negation of the object by cleft sentences and negation of the object by sentences with the locative AGRS [Ho].
3.1 Postverbal negation As indicated above, the object argument may be negated in its original postverbal position in transitive constructions.

Transitive verbs
The object argument with feature animacy, which is to be negated, is indicated in (19)  In (19) above, the object argument appears next to the verb and it is not moved. To establish negation of such an NP object, the negative morpheme [ha] is employed and it will appear with the negative suffix [-e.]. A contrastive clause is also necessary to highlight negation of the object as shown in (20)  In (20), the object argument is negated within the verb phrase as the objects form part of the transitive verb, i.e. the negation of the verb also affects its arguments, which in this case is the objects [bana].

Ditransitive verbs
As in the case of transitive verbs, the object arguments i.e. direct and indirect objects, may be negated in their original postverbal position in ditransitive constructions. The object arguments to be negated are indicated in the sentences in (21)  In (22) above, the indirect and the direct objects are negated within the verb phrase as they are subcategorized by a ditransitive verb. To isolate negation on these objects, a contrasting clause is necessary either to put emphasis on the direct or the indirect object. To negate the indirect animate object [bana], a contrasting clause with an animate NP will be employed as indicated in (23)  In the case where the indirect object is not human, a contrasting clause will also be non-human as illustrated in (25) From the sentences in (27) and (28), one may deduce that it is possible to use cleft sentences in negating the object arguments with featured animacy.

Ditransitive Verbs
The object argument may be moved to the sentence's initial position through clefting. In the case of ditransitive constructions, only the moved NP object will be negated. In (29) below, the indirect object is moved to appear in a focus position: To strengthen negation on the moved indirect object, it is also acceptable to add a contrasting clause as illustrated in (31)  When the animate direct object is moved to appear in a focus position, the sentence becomes ambiguous as in (32)  From the sentences in (31) and (32)  To strengthen negation of the object argument, a contrasting clause is added as indicated in (35)  From the sentences in (34) and (35) above, one may conclude that it is also possible to negate the object arguments in transitive constructions by means of locative AGRS [Ho] clauses. The animate and inanimate NP objects may be negated by cleft sentences and sentences with locative AGRS [Ho], where the object is negated within a larger sentence scope. It is also possible to negate the NP object within the VP.

Ditransitive Verbs
It is possible for the indirect object argument to be moved to the sentence initial position where it will be introduced by the locative AGRS [Ho]  In negating the object argument, the negative morpheme [ha] will appear with the locative AGRS [Ho] with the object as the head of the relative clause. In the case of the constituents within sentence constructions. The paper also examined object negation in non copulative verbs within transitive and ditransitive verb constructions. The following conclusions can be made.
Sesotho has no direct means of negating clausal constituents, but employs negative clauses such as cleft sentences, clauses with the locative morpheme /ho/ and clauses with subject inversion. In order to negate a single clausal constituent, Sesotho follows a route comprising of three steps. Firstly, the constituent to be negated, in this case the NP object, needs to be placed in focus position by means of copulative verbs, the locative morpheme /ho/ or the existential /ho/. The second step is characterized by implementing negative morphemes /ha/ and /se/ to effect negation. In this step, the object argument is negated within the larger scope resulting in sentential negation. In the last step, contrastive clauses are employed. In other words, constituents are negated in context.
In terms of Haegeman's (1995) Neg-Criterion, this study intends to suggest that Sesotho does not have non-negative operators that will constitute local negation. It conforms to the Neg-Criterion in terms of negative operators which trigger subject inversion, in that it is only after sentential negation has occurred that local negation is achieved through contrastive negation. This is because Sesotho negative morphemes, in this case /ha/ and /se/, are bound negative morphemes. They are verbal prefixes which always appear with their verb stems. In order to negate a clausal constituent, these negative morphemes have to move with their verb stems, hence negative clauses are employed. As verbal prefixes, they cannot be generated from nouns and adverbs as it is the case with English. It is for this reason that Sesotho does not have direct means of negating clausal constituents.
Furthermore, post-verbal negation is possible in transitive verbs where the object arguments are negated within the verb phrase because NP objects form part of the transitive verb. i.e. the negation of the verb also affects its arguments, which in this case are the objects [ Preverbal negation is possible in Sesotho in order to achieve constituent negation. This is possible by using the cleft sentences in both transitive and ditransitive verbs. However, it would seem that with ditransitive verbs, only indirect objects can appear in a focus position, and subsequently only indirect objects can be negated by means of cleft sentences. As it is the case with post verbal negation, contrastive clauses are also necessary to achieve constituent negation: Pre verbal negation is also possible in Sesotho by using clauses with locative AGRS [ho] in both transitive and ditransitive verbs. As it is the case with cleft sentences, only indirect NP objects may be negated by clauses with locative AGRS [Ho].