Focus position below TP in Standard Arabic Open Access

There have been several recent studies on various aspects of Standard Arabic (SA) syntax which have adopted the split complementizer phrase (CP) hypothesis (Alsager, 2017; Alazzawie, 2018, 2019, 2020; Alnajadat, 2020). However, previous studies have not dealt with displaced pronominal objects of SA in much detail, particularly within the split CP hypothesis. This study aims to investigate the topic of dependent pronominal objects within Rizzi's split CP hypothesis.

The term displacement used in this paper refers to a syntactic process where a category, specifically a full determiner phrase (DP) object or a pronominal object moves to either clause-initial or clause-internal position. The typical object position in SA is the sentence-final position:

The object al-ʔarḍ-a follows both the verb and the subject in the sentence above. SA also allows for another variant in which the object is placed in a position structurally above the subject as (2) illustrates:

The optional displacement of the object is associated with the rhetorical effect of attracting attention and prominence to the object DP ^[1]. While the displacement above is optional, it is obligatory when the subject DP is a host for a pronominal clitic in co-reference relation with the object:

In contrast with the sentence (3), sentence (4) where the object is not displaced leads to an ungrammatical sentence. The ungrammaticality of the sentence in (4) could be attributed to binding theory – the object DP that is the antecedent of the pronoun –haa must be fronted to bind the co-referential pronoun. As it stands, the pronoun included within the subject DP is not c-commanded by its antecedent al-ʔarḍ-a “the land”.

It should be noted at the outset that the term clitic is used here to refer to a maximal projection (XP) pronominal object, not X° head along the lines of cliticization in Romance. As a syntactically independent XP, it does not cliticize to the verb inside verb phrase (VP) but remains a separate argument and undergoes movement one step higher to the specifier position of Foc. In other words, the clitic does not merge with the verb during sentence derivation. Cliticization occurs in T after verb raising. As such, the clitic is not an affix marking agreement. This will be justified later in section (3).

The following example also involves re-ordering of the object DP al-ḥariiq-a “the fire” to a position to the left of its canonical position under the “heaviness” of the subject DP in the sense of Ross' (1967) heavy noun phrase (NP) shift:

In this example, the grammatical complexity of the subject triggered the re-ordering of the constituents.

In addition to the optional displacement of a fully referential object DP (3), SA also allows for another context (which is the primary focus of this paper) where object displacement obligatorily applies. This context is seen in examples where the complement is reduced to a non-nominative pronominal clitic. Thus, while the displacement of the full DP al-ʔarḍ-a in (3) above is optional, it is obligatory in the context of a pronominal object clitic:

As shown in (6) the non-nominative clitic must be bound to the lexical verb ya-sqi ^[3].

Of relevance to the present analysis is the treatment of clitics in Romance. Hence, a brief discussion of this construction is in order (cf. also Sections 3.3 and 3.4) In Romance languages, Kayne (1975, 1989, 2000) observed that pronominal objects may surface in a position between the subject and the object ^[4]:

Working within the Barriers framework (Chomsky, 1986), Kayne explains the correlation between clitic climbing and null subjects, proposing the following cross-linguistic generalization on clitic climbing:

According to Kayne, clitic climbing is linked to the null-subject parameter operative in the language. Thus, the phenomenon is permitted in Italian but blocked in French since the former licenses null subjects, an option unavailable in the latter.

The phenomenon is referred to as cliticization, meaning that clitics undergo syntactic movement at the same level as full lexical noun phrases from their base-generated complement position under VP to a higher position, specifically [Spec, Foc] (cf. 3.1). For example, in (11), it can be said that the pronominal object –hum “them” is the reduced form of the complement of the verb sarra “pleased”. It appears attached to sarra, giving rise to the verbal complex sarra-hum. Note that the form –hum appears following the verb, yielding the verb-clitic-subject order. (11b) shows that the clitic cannot stand alone and be used as an independent complement:

Note that in contrast with lexical DP displacement which is optional, pronominal clitic movement is obligatory.

Regarding the status of object affixes in SA, Teeple (2011) argues they are indeed pronouns and not agreement markers. The arguments are reviewed below as it is important to the present analysis.

The first argument is based on the fact that the accusative pronouns of SA are in complementary distribution with lexical DPs:

The accusative pronoun cannot occur together with the full accusative DP. Teeple (2011) argues that this is a “principal criterion for differentiating pronominal affixes from the verbal agreement, which can occur alongside co-referential full DPs” (p. 144).

The second argument for the pronominal status and against verbal agreement status is the fact that they are found suffixed to other Case-assigning categories as well, such as complementizes (13), and active participles which are verbal derivative nominals (14): The abbreviation ACT. PRT stands for active participle.

A third argument for treating the SA accusative affixes as pronouns is based on their weak and stressed or strong forms. The stressed form is suffixed to a dummy stressable unit, ʔiyya-:

The pronoun is contrastively focused, foregrounding the referent as new information in contrast to the prosodically weak pronoun signaling old information. As Teeple (2011) observes, “[a]greement affixes are not generally subject to this sort of strong-weak alternation; even under contrastive focus, it would be unusual to find a strong form of an agreement affix that differs in any way other than stress” (p. 145).

The conclusion is that SA accusative affixes are pronouns and not markers of agreement. The present study adopts this position.

Working within Phase Theory (Chomsky, 2001, 2005) where CP and vP are considered phases whose heads initiate the Agree operation, Musabhien (2009) proposed an analysis of the Case feature of the pre-verbal VP in SA. Of relevance to the present study is his analysis of the verb-object-subject (VOS) word order allowed under certain conditions, where the object DP is fronted to a position preceding the subject. According to this analysis, the VOS order is derived through object movement which is, as he convincingly argued, distinct from object shift found in Scandinavian languages, and from scrambling in German. As indicated, Musabhien's analysis utilizes the notion of edge feature associated with the head of the vP phase as in Chomsky (2001, 2005) work. That is, the trigger of object movement to the pre-subject position is the edge feature of v, the head of the vP phase. Accordingly, the object DP moves to an additional vP specifier position.

Regarding the status of the moved object, he argues that this pre-subject position is a focus position serving as the landing site for the moved object DP. The relevant portion of the structure is diagrammed below:

As indicated, object movement targets the additional specifier position of the light v made available at the phase head v.

Adopting and building on Chomsky (2005), Soltan (2007) proposes that the moved object DP targets an additional specifier of v*P, which means the DP remains within the projection of v. Soltan argues, based on data from SA, that the periphery is not limited to the Phase heads, C and v as in Chomsky (2005) but may include the periphery of other functional heads like T, Neg(ation), or Mod(ality). He further posits that the periphery is not a position but a “syntactically salient” zone, a concept attributed to Uriagereka (2006).

Alazzawie (2020) treatment of focus and object displacement in SA is also relevant to the present study. The author argues that “the clause in SA has two designated focus positions – one projected in the upper layer of the clause (clause-initial) and the other in the lower part above vP” (p. 2). It is proposed that focus projection replaces the extra layer of v*P, and can fulfill the focus-feature-raised function of object movement in SA.

The present study builds on the above proposal and extends it to accusative pronouns referred to as clitics, using Rizzi's framework instead of Chomsky's Phase theory with its edge feature, and outer specifier. In particular, the present analysis posits the existence of a focus position above the vP layer as the landing site for moved accusative pronouns on a par with the full accusative DPs. The present account, thus, diverges from Musabhien's (2009) account as it assumes, following Rizzi, the existence of a focus position below TP as an independent category. Thus, in the case of VOS word order, the object lands in the specifier of Foc, the head of FocP located immediately above vP. The current proposal unifies both processes applying to full accusative DPs and accusative pronouns as instances of movements to the same specifier of FocP.

Adopting the split CP hypothesis, the paper aims at providing an account for object cliticization in SA as an instance of object displacement. The proposed analysis builds on and extends the object displacement analysis (Alazzawie, 2020) of lexical DP's to object cliticization. Kayne's (1975) proposal on cliticization argues that clitics are base-generated, syntactically-independent arguments in French, and this is the proposal we use here to account for the type of object displacement in SA.

The study attempts to answer the following questions:

1. What is the landing site of the moved object pronouns in SA?

2. What triggers the movement of object pronouns in SA?

We claim that object clitics in SA, like full DPs, obligatorily move from their base position as independent complements of the verb to the specifier of Foc first before attaching to the verb under the tense node ^[5].

This section states the research design, followed by a description of the data collected and analyzed. The last section discusses the relevant aspects of the syntactic model adopted for the analysis of the data.

This research adopts a qualitative interpretive research design (Creswell, 2007; Creswell and Plano, 2010). It does not attempt to test or measure interventions, but rather produces a rich description of the reality it observes. Strauss and Corbin (1990) define qualitative research findings as the selection of themes occurring in real-world settings where the interest area is pronounced naturally (Patton, 2002). Altheide (1987) and Morgan (1993) state that qualitative content analysis is a strategy of choice, which is a dynamic form of analysis of verbal and visual data oriented toward summarizing the informational content suggested by the data. Qualitative content analysis is therefore data-driven, that is, codes are not only systemically applied, but also generated from the data themselves in the course of analysis (Sandelowski, 2000).

The majority of the samples chosen for the study involve dependent pronominal objects obligatorily attached to the verb. The grammaticality judgment on the examples provided and analyzed in this study is based on the authors' judgments as native speakers of Arabic. The examples were also presented in the form of a questionnaire to other native speakers of Arabic from Jordan, Iraq, and Algeria, who concurred the judgment on the data.

The model adopted for the analysis of SA data is Rizzi's split CP hypothesis (1997, 2004) which suggests that the CP category should be split into a number of distinct functional projections. For example, C is analyzed as a Force marker heading a ForceP (Force Phrase) because C indicates clause type such as declarative, interrogative, comparative, imperative. Rizzi also maintains that “Focalization” – fronting of focused constituents – should be positioned in FocP (Focus Phrase) headed by Foc. He argues for another projection positioned between FocP and TP, namely FinP (Finiteness Phrase) headed by Fin which serves to specify the clause as ±finite (finite or non-finite). For example, for serves as a nonfinite Fin head in English. Topic Phrases, on the other hand, are positioned within TopP whose head is Top, and they host topicalized phrases.

The overall structure of the CP domain in this model consists of a Force-Finiteness system interacting with a Topic-Focus system with ForceP being at the top for selectional reasons, and a Topic-Focus system sandwiched between ForceP and FinP. The split CP then consists of the following layers ^[6]:

• 12. ForceP (TopicP) FocusP (TopicP) FinP IP/TP VP

Of crucial relevance to the analysis of pronominal objects in SA is the Focus position as the scope of the paper is restricted to this set of pronouns ^[7]. The proposal put forward is that pronominal objects undergo displacement as independent syntactic complements from their base position inside VP into the focus position in the same manner as fully specified DP ^[8]. It is proposed that the pronouns do not cliticize into the verb in the V⁰ position but higher in the tree structure, specifically in T(ense) as they are able to move out as full syntactic XP complements into the specifier position of Foc without attaching to V⁰ under VP. This is followed by a morphological process of cliticization onto the verb which has raised into the T(ense) positon.

Adopting the model outlined above, this section addresses the objectives of the study to propose a minimalist analysis for object cliticization in SA in relation to the derivation of the process, particularly, the surface position occupied by the moved pronominal object and the motivation triggering the movement.

Below are sentences illustrating various pronouns which function as objects. The pronouns are boldfaced and must appear attached to the verb, resulting in VOS word order:

The verb in each of the sentences hosts a pronominal object clitic. The cliticized object surfaces to the right of the verb preceding the subject DP. However, the base position of the object is inside the VP to the right of the verb in the assumed VSO order of SA and has moved out of its base position along the verb, resulting in a VOS surface order. The pronominal object necessarily undergoes displacement, resulting in an OS order. This is supported by the fact that the pronominal object intervenes between the verb and the subject. The verb originates within the VP, and then raises to T (Ouhalla, 1994; Benmamoun, 2000), thus ending up in front of the object and the subject, rendering the VOS order. The subject of the clause remains in the specifier of vP, assuming the internal subject hypothesis (Koopman and Sportiche, 1991), and the notion that T in SA does not attract subject raising (Soltan, 2007). The object pronoun movement, the verb movement and the positions targeted by movement are discussed in the next section.

A reasonable hypothesis is that the displacement operation targets the specifier of Foc situated above vP as diagrammed (22) ^[10]:

This structure claims that the object clitic originates (is merged) as a complement of V. It moves out of VP to the specifier of Foc across the subject DP. The clitic movement is triggered by the edge feature of Foc which assigns to it a functional status of foucs. The focus-raised clitic undergoes a merger with T hosting the lexical verb given that SA is a verb raising language (Ouhalla, 1994; Benmamoun, 2000). The derivation is assumed to proceed as sketched in (23):

As depicted, the structure involves two operations: the first one involves the movement of the verb from the head V (assumed to comprise the consonantal root) into the next head position, namely the v node; and then the movement from the v position into the next higher head position, namely the head T (assumed to include the vocalic inflection) of TP. The direct object is a pronoun, which must undergo movement. Hence, the second operation involves the movement of the pronominal object clitic from the complement position of V into the specifier of FocP as indicated by the broken arrow followed by its merger with the verb under T. As clitics in general “require a host to attach to” (Spencer and Lu, 2012, p. 18), the clitic in SA, obviously by its nature, having the intrinsic property to attach itself to a host (in this case, a verbal host), is attached to the right as an enclitic ^[11]. The view adopted here is the internal subject hypothesis (Koopman and Sportiche, 1991) whereby the subject remains inside the vP. T in SA lacks an extended projection principle (EPP) feature, and thus does not yield a specifier position.

The claim that object clitics originate internally as full arguments inside the VP, and from there move into the specifier of Foc can be supported by the following data. The data in (24–25) demonstrate that SA has displacement of full DPs. The sentences in (a) reflect the unmarked VSO word order in SA whereas the (b) sentences reflect the marked VOS order:

The unmarked word order, generally assumed for Arabic, contains the object DP inside the VP thus eliminating the possibility of a focus interpretation apart from an in situ focus, which is excluded from the present discussion. The object DP which originates as a complement inside VP in (24b and 25b) has been focused, i.e. moved outside VP to a position between the verb and the subject in order to receive special emphasis.

One way of describing and motivating the process is to suppose that the functional head Foc carries an edge feature which attracts the object DP to move to the specifier of Foc (Ouhalla, 1997; Alazzawie, 2018, 2020) ^[12]. The advantages of this view is that the landing site of pronominal objects and full DPs can now be understood as the same specifier position. An analysis which conceives of pronominal objects as head-movement has to treat object displacement of lexical DPs in the language differently, that is, they do not result from the same process although they share common properties. There are good reasons (to be discussed further in section 3) not to consider pronoun object displacement as X⁰ head movement along the lines of cliticization in Romance languages. In the proposed account, the pronoun does not cliticize to the head verb inside the VP but rather moves out of the VP as an argument (XP) targeting the specifier position of Foc ^[13].

The focus function of the repositioned DP can be strengthen further by the exclusive restrictive focus style consisting of the focus particles [the negative element maa … exclusive element ҁilla]:

Introducing the [maa … ҁilla] strategy, the meeting activity is solely confined to the repositioned focused object DP, thus excluding any possibility that the manager might have met someone else.

A test for focus that has been commonly used is the question answer pair (Kasimir, 2005, p. 12) ^[14]. The wh-question in (27a) triggers the two congruent answers in (27b and c) with DP preposing:

If focus is defined as new information, the DP faaṭimat-a, used to replace the wh-word “man” is focused. It is also possible to answer the question with the canonical VSO order with a pitch accent (capitalized) on the direct object:

The sought-after knowledge required to fill in the gap of our understanding of the event is in focus ^[15].

SA allows for the use of strong, stress-bearing, independent accusative pronouns suffixed onto apparently dummy form ҁiyyaa-. Using this ҁiyyaa-as a base produces ҁiyyaa-ya “me”, ҁiyyaa-ka “you”, ҁiyyaa-hu “him”, ҁiyyaa-hum “them”, etc. which assigns the accusative pronoun a contrastive focus. The focused pronominal counterpart of the sentence in (26) above requires expletive ҁiyyaa-support, so that the pronoun has the form ҁiyyaa-ha as illustrated (EXPL stands for expletive):

But how does ҁiyyaa come to appear in this example and why? Before being repositioned, the pronominal clitic -ha is generated in its canonical object position in the VSO order as we have already seen. Since it is unattached and thus requiring a host, and since it is being focused, i.e. attracted by the edge feature of Foc, it moves to the specifier of Foc. The result is that the pronominal remains stranded which triggers the insertion of the dummy form ҁiyyaa to attach to. The insertion process can be conceived of as simply the spellout of a stranded pronominal, occurring in the phonological component rather than in the syntax. The phonetic form (PF) attachment and lexical support operation applies spelling out the unattached pronominal as a fully formed item, viz., ҁiyyaa-ha.

As expected, other dependent pronominals, such as –hum, -hunna, -kum and –kunna, may undergo preposing as XPs from the verbal complement position to the specifier of Foc in order to highlight these pronouns, and thereby surface in front of the subject. Once again, these dependent pronominals cannot remain stranded and accordingly ҁiyyaa-support applies to provide a host with the result that the pronominals are spelled out appropriately as ҁiyyaa-hum, ҁiyyaa-hunna, ҁiyyaa-kum and ҁiyyaa-kunna, respectively, as exemplified below:

The dummy/meaningless ҁiyyaa appears left-adjoined to the pronouns, an operation which has to apply, as indicated, spelling out the unattached pronouns as an appropriately formed unit.

According to the grammarians, in contexts where it is possible to use a dependent pronoun, the independent form should be disallowed. Thus, since it is possible to have the sentence in (31a), where –ka is directly attached to the verb, the sentence in (31b) where ҁiyyaa is the host should not be allowed or rather considered marginally acceptable ^[16]:

However, preposing ҁiyyaa-ka to sentence initial position is perfectly acceptable:

In keeping with the focus-driven displacement analysis for object pronouns proposed and schematized in the structure above, it is reasonable to postulate that the same focus-triggered syntactic operation takes ҁiyyaa-ka from a lower [Spec1, Foc] to a higher focus position [Spec2, Foc] to obtain a higher degree of attention and prominence. This operation appears to serve the same function of ensuring a strong focus interpretation for full DP's ending up at the beginning of the sentence, i.e. in the upper [Spec2, Foc].

As already supposed, Foc has an edge feature enabling it to attract the object pronoun to move to its specifier [Spec1]. The question is why is the sentence in (31b) ḍaraba ҁiyyaa-ka Zajd-un considered unacceptable as an instance of focus? There is a possible explanation for this. Suppose that, being a clitic, dependent and prosodically weak, the pronoun cannot be marked for focus in this position unlike lexical DPs, even when strengthened in some sense through hosting by the dummy formative ҁiyyaa. Suppose further that the higher Foc is associated with a stronger edge feature so that the lower Foc hence attracts the pronoun in the lower Foc, thus deriving the marked emphatic sentence and assigning it the focus function with more weight and emphasis. The utterance is used to exclude all alternatives, that is to deny any suggestion or possibility and to remove any doubt that Zajd did not hit anyone other than the addressee. It is possible to argue, on the basis of this case, that focus comes in degrees and that sentence initial focus of the left periphery is stronger than focus in lower positions.

It is instructive to draw a parallel between the focused pronominal objects discussed above and contexts where the focused DP is lexical (non-pronominal):

Both full DP objects and pronominal objects may undergo object displacement out of VP without incorporating into the head V as discussed earlier. The following sentence is parallel with (34b). The pronoun takes initial position, it is now strong in the sense that it is stressed and can be modified and coordinated (Holmberg, 1986, p. 209):

While both DPs in (36) below are focused, the sentence initial DP receives more emphasis, i.e. it is more prominently stressed than the sentence internal DP. When contextualized, what is considered new information and what is being contrasted and focused depends on the preceding and succeeding discourse. It is evident that the second example encodes more focus and more exclusivity:

The ҁiyyaa meaningless dummy is used to aid focus marking in sentences which would otherwise be unfocused as the pronoun would be stranded without a host. The ҁiyyaa strategy strengthens the pronoun in its morph-phonological nature and thus enabling to function as an independent focused pronominal object.

To sum up, the analysis takes object pronominals as clitics having independent syntactic status and function in SA, and the clitics invariably move in “audible syntax”, like fully specified DP complements, to the specifier position of Foc. Both pronominal objects and full DP objects may undergo object displacement in SA. The movement is triggered by the edge feature on Foc attracting the pronoun to its specifier with the difference having to do with the obligatory nature of object movement of pronouns vs the optional object displacement of full DPs.

Arguments supporting the claim that cliticized pronominals are initially merged (i.e. enter the derivation) as independent arguments and therefore act as lexical DPs can be based on the Binding Theory, conjunction, passivization, substitution and clitic clustering. The arguments are presented in the following Sections (3.3–3.7).

To illustrate the argument status of the dependent object pronouns, examples (20a) and (25b) are repeated as (37a) and (37b), respectively:

Note that –hu cannot have the DP Zajd-un as a binder inside the TP domain, a piece of evidence for its pronominal status. As per principle B of the binding theory, it is not allowed for a pronominal to co-refer with a DP in its minimal binding domain:

• 38. *qatala-hu_i Zajd-un_i

The minimal binding domain here is the TP containing the pronoun the verb and the subject. The sentence is unacceptable with the intended reading since the pronoun is in fact bound, not free as required. Moreover, the pronoun fulfills the internal argument role of the predicate qatala much the same way as the fully specified DP in (37b). These examples are minimal pairs. The omission of the pronoun renders the sentence ungrammatical. The inclusion of the pronoun along with the lexical DP creates two slots for the one-place predicate:

This, however, poses an important question of implementation about what mechanism ensures that the clitic does not merge with the V inside the VP. In this connection, we follow Kayne (1991, 2000) who argues that cliticization is not VP-internal but rather it is an “IP-internal cliticization”. Under this view, cliticization unto the verb in SA does not take place inside the VP, but rather outside of the VP. In other words, the dependent pronominal object would have to be able to move one step as an XP to [Spec, Foc] without merging with V°. The order verb-object clitic-subject results from the verbal consonants' having moved from V to T via v and the subsequent merger of the clitic with the verb.

The argument status of the clitics can also be supported on the basis of conjunction such as,

The cliticized object pronoun –haa “her” is conjoined with a non-pronominal lexical NP l-muҁalim-a “the teacher” by the conjunction element wa “and”. The object pronoun –hu “him” is conjoined with a full NP l-mudiir-a “the principal”. There are two DPs in each sentence, one pronominal and the other is non-pronominal, which appear to fulfill the same syntactic and semantic functions, namely patient and theme, respectively. This coexistence supports the idea of treating object pronominal clitics as syntactically-independent elements. The two DPs provide answers to the questions that follow:

It is worth noting that an accusative pronominal clitic and a nominative DP (-u marks nominative) are unconjoinable, as in the following which is a minimal pair with (40a) above.

Like a full DP conjunction, the clitic and the full DP must match in grammatical function, suggesting their shared syntactic behavior as arguments. In particular, the data suggest attached accusative pronominals are clitics occupying a complement position, and are not agreement-marking affixes.

It is not possible to conjoin two clitics (Kayne, 1975, 1989). Two attached pronominal objects cannot be conjoined, neither can two attached subjects, thus exhibiting clitic properties:

It is also possible to conjoin a cliticized object pronoun with a free standing object pronoun:

The two pronominals have the same syntactic function – both are direct objects. Pronominal objects and non-pronominal DPs behave alike with respect to conjunction. Conjunction, a standard constituency test, provides some evidence since only alike constituents can be conjoined ^[17].

Sentence (45a) is active and (45b) is the passive version:

The initial DP ^[18] ҁaṭ-ṭullaab-u “the teacher” is linked with the pronoun –hum “their” inside a construct state (genitive) DP complement with a patient theta-role. The verb ḍaraba “hit” agrees in person and gender with the lexical subject al-mudiir-u “the principal”, not with the sentence-initial DP. When the sentence is passivized, the verb agrees with the patient theta-role which is now turned into subject marked for the nominative Case. This is further enhanced by the following active and passive contrast:

The accusative pronoun -hum in the active sentence which is the direct object thematic complement (patient) of the verb has become the structural nominative subject -uu of the same clause in the passive version. This suggests that clitics are part of the argument structure of the verbal predicate denoting the activity, i.e. -hum and -uu represent the real participants in the relevant event.

Facts gleaned from coordination and passivization processes suggest that attached object pronouns in SA are less likely to be mere inflection markers. Rather, the examples strongly support the hypothesis that clitics have an independent status as genuine arguments generated in the base, following the reasoning of Kayne (1989).

Dependent object pronouns can substitute for a full DP complement of a verb:

The above sentence can be rephrases using the pronoun –hu “it”:

Dependent object pronouns may also substitute for the genitive DP in a construct state:

Like full DP's, they may also substitute for a genitive complement of a preposition:

Spencer and Lu (2012) give a number of diagnostics for clitics, chief among them are:

Clitics must attach to a host, may basically attach to any host “show low selectivity towards their host” (p. 37), show a rigid order in the way they cluster and attach to their host, and pronominal clitics often function as arguments. It has also been observed (Zwicky and Pullum, 1983) that only clitics may be integrated consecutively to a host in the manner shown below ^[20]:

Three pronouns are integrated to the same head verb ҁarsal “send” in this double object construction: 1st person pronoun –tu, second person singular masculine pronoun –ka and third person singular feminine pronoun ^[21] –ha, suggesting their status as clitics. The following clustering is also possible with the verbs ҁahdaa “to present as a gift” and ҁaʔṭaa “to give”:

In the following case, two other pronouns –uu and -nii exist cliticized to the same host ḍarab “hit”

We take –uu and –nii to be pronominal clitics functioning as subject and object respectively.

The pronouns may also lean to some complementizers, among them ҁinna glossed as C:

The clitic nature of the dependent pronouns is supported by their conjoinability with fully-specified DPs, passivization, clustering, integration into a variety of hosts: verbs, nouns, prepositions, and complementizers. As illustrated, they “attach promiscuously … they do not select words of a particular class” (Spencer and Lu, 2012, p. 18).

This article provided an account of pronominal objects in SA, using the FocP projection of (Rizzi, 1997, 2004). It is proposed that such objects move out of VP and target a structural position in a FocP projection whose head is Foc.

As indicated, in line with Rizzi (1997, 2004), this article proposed that the left periphery feature of Foc attracts the accusative pronouns of SA from their base position as complements of V to the specifier of Foc. However, In Rizzi's system the Focus Phrase is projected above TP by virtue of belonging to the CP system. We extended Rizzi's idea of FocP projection, which is standardly higher than TP, and assumed that object pronouns in SA target an intermediate specifier position available between TP and vP in the clausal structure of the language. Thus, a new FocP projection is introduced and placed lower than TP. In other words, the proposal put forth is that there are two projections of FocP in SA mediated by TP, one higher and one lower than TP. The periphery feature associated with Foc in this intermediate projection triggers the accusative pronoun displacement, resulting in the object – subject word order.

The V-clitic complex appears under T below the split-CP layer as an instance of head movement. Regarding this subject, We have adopted the internal-subject hypothesis with the view that the subject does not raise from within vP to the specifier of TP, given that T in SA lacks the raising edge features responsible for subject raising in languages like English (Soltan, 2007).

An opportunity for further research is extending the proposal of the additional intermediate FocP projection to the analysis of double object structures in SA. One possibility is to position the indirect object in the specifier position below TP in structures where the indirect object precedes the direct object.Tables 

The analysis proposed that, like lexical DPs, pronominal objects originate as separate XP constituents and move from their base position as verbal complements to the focus position [Spec, Foc], positioned immediately above vP within TP. In other words, both are able to move out of VP, targeting the same specifier position of a functional projection. This movement is focus-driven, that is, triggered by the edge feature of Foc. This proposal allows for object movement of full accusative DPs and of accusative pronouns if the notion of FocP is adopted, following Rizzi (1997, 2004). The motivation for this movement to the landing site follows from the edge feature of Foc.

Pronominal objects at a later phase, crucially higher than V⁰ (possibly in PF), get cliticized to the verb which has adjoined to T. Alternatively, ҁiyyaa-support has to apply to save the unattached pronoun which may also take sentence initial position for the same pragmatic function of prominence in the specifier of Foc. Their clitic nature and argument status has been justified on the basis of conjunction, passivization, binding and clustering.