The Problem of Word Order and Verbal Movement in Moroccan Arabic

This paper attempts to investigate word order and verbal movement in Moroccan Arabic in the Minimalist framework. We observe that the unmarked word order in MA is SVO while the derived structure is VSO. SVO follows an English-like derivation where the subject moves from [Spec, vP] to [Spec, TP] whilst the verb moves from v to T. This paper raises the issue of the verbal movement when it comes to VSO order in languages that have VSO as the derived order and SVO as the underlying order. To derive VSO, we propose that the verb moves from T to Focus based on pragmatic reasons: verbs positioned in the left-periphery denote new information that is focused compared to SVO. We also test our new proposal against the marginal word orders OSV and OVS and propose that object topicalization is the result of the object moving to [Spec, TopicP] which dominates FocusP. Moreover, we go back to the issue of verbal movement and trace the verbal cyclic movement. We argue that the verb moves from V to v based on the position of the adverb. The verb further moves to T based on the quantifier evidence and feature checking: Focus and T form a complex and probe into v to check [TNS] and [V] features. Moreover, T-to-Focus occurs in wh-constructions except when /lli/ ‘that’ is present. In WH-VO (WH as a wh-subject), the verb stays in T while the wh- subject stays in [Spec, TP]. If /lli/ ‘that’ is present, then the wh-subject is forced to move further to [Spec, FocusP]. In WH-SV, the wh-elements move to [Spec, FocusP] while the subject moves to [Spec, TopicP] and the verb moves to Topic. In WH-VS, the wh-elements move to [Spec, FocusP] while the verb moves to Focus.

upwards (Chomsky, 1995). The operation of Move will be restricted in later Minimalism, which is Agree-based (Chomsky, 1998(Chomsky, , 1999(Chomsky, , 2001a and it will only be concerned with checking the [EPP] feature while the rest is operated under either Merge or Agree. Finally, Agree is an operation that can be defined as follows (Chomsky, 2000(Chomsky, , 2001a: (1) Standard Agreement α can agree with β iff and only iff: a. α carries at least one unvalued and uninterpretable feature and β carries a matching.
β is the closest goal to α. d. β bears an unvalued uninterpretable feature.
Agree entertains the interaction between two elements: Probe and Goal. The elements are activated when an uninterpretable feature exists (Chomsky, 1999, p. 4). For example, verbs agree with their subjects. Both the verb and the subject share phifeatures (i.e., gender, person, and number features). While phi-features are interpretable on the nominal subject, they are uninterpretable on the verb (i.e., they are not important in interpreting the meaning of the verb); therefore, the verb is the Probe while the subject is the Goal and this results in the phi-features on the verb being checked and deleted before they reach the Spell-Out.
The outcome is that the structure is spelled out (in a stage called the Spell-Out) and shipped to LF and PF in order to meet the principle of Full Interpretation (Chomsky, 1986a), which demands that all items receive appropriate licensing and interpretation or else the derivation will crash at the LF (Chomsky, 1995). After giving a general overview of MP, we introduce some descriptive generalizations on word orders in MA.

Some Descriptive Generalizations on SVO and VSO orders
A simple sentence in Moroccan Arabic (henceforth MA) contains a subject and a predicate: (2) ħməd kla t-tffaħa Ahmed ate.PAST.3SG.M the-apple 'Ahmed ate the apple.' (2) is a simple SVO structure in MA: /ħməd/ 'Ahmed' is the subject of the sentence, /kla/ 'ate' is the verb, and /t-tffaħa/ 'the apple' is the complement of the verb /kla/ 'ate'. A simple sentence in MA can also have a VSO order: (3) kla ħməd t-tffaħa ate.PAST.3SG.M Ahmed the-apple 'Ahmed ate the apple.' (3) is a VSO structure where the verb /kla/ 'ate' precedes the subject /ħməd/ 'Ahmed' and the direct object /t-tffaħa/ 'the apple'. Unlike Standard Arabic (henceforth SA), MA has SVO as the unmarked structure. This has implications on not only the initial structure of MA (i.e., the D-structure from a GB 2 perspective) but also its derivation (i.e., the Spell-Out structure from a Minimalist perspective). This means that VSO is the derivational result of the unmarked structure SVO. It should be noted that the Dstructure being SVO is not new and not only argued for in MA but also in SA. For example, Elesseily (1985, p. 2) argues that "the D-structure of SA is SVO, and that this becomes VSO at S-structure via verb movement." This is against the traditional belief that the underlying order of SA is VSO (see Aoun, 1979;Emonds, 1980;Zughoul, 1980;Marshad and Suleiman, 1991;Fassi Fehri, 1993;Belnap and Haeri, 1997;Al-Shorafat, 1998, as cited in Elesseily, 1985 andBtoosh, 2010). Musabhien (2008) also contends that Jordanian Arabic has SVO as its unmarked structure. In fact, it seems that Arabic dialects use SVO as its underlying structure (see El-Yasin, 1985 for Jordanian Arabic; Aoun, Benmamoun, and Sportiche, 1994 for Lebanese Arabic; Shlonsky, 1997 andMohammad, 2000 for Lebanese Arabic; Benmamoun, 2000b for Egyptian Arabic; Mahfoudhi, 2002 for Tunisian Arabic; Fassi Fehri, 1993 and Benmamoun, 2000b for Moroccan Arabic, as cited in Musabhien 2008, p. 22).
In addition to that, there is another distinction between MA and SA 3 when it comes to word order. While MA shows full agreement whether SVO or VSO, SA is more conservative and shows partial agreement when it comes to the VSO order. Consider the following examples from SA 4 : (4) ʔal-ʔawla:d-u qaraʔ-u: ʔal-dars-a SV + full agreement the-boys-NOM read-3PL.M the-lesson-ACC 'The boys read the lesson.' (5) qaraʔ-a ʔal-ʔawla:d-u ʔal-dars-a VS + partial agreement read-3SG.M the-boys-NOM the-lesson-ACC 'The boys read the lesson.' (6) *ʔal-ʔawla:d-u qaraʔ-a ʔal-dars-a *SV + partial agreement the-boys-NOM read-3SG.M the-lesson-ACC 'The boys read the lesson.' (7) *qaraʔ-u: ʔal-ʔawla:d-u ʔal-dars-a *VS + full agreement read-3PL.M the-boys-NOM the-lesson-ACC 'The boys read the lesson.' In (4), we have a case of SVO structure. The verb /qaraʔ-u/ 'read' agrees in person, number, and gender with the subject /ʔal-ʔawla:d-u/ 'the boys'. The verb contains the element /-u/ which marks the third person, plural, masculine and fits the features to that of the subject /ʔal-ʔawla:d-u/ 'the boys'. However, once VSO structure is applied (see e.g., 5), a full agreement no longer applies, and we see that the verb /qaraʔ-a/ 'read' has a different marker /-a/ and only agrees with the subject partially in terms of person and gender while the number agreement no longer holds. We further observe that in SV order, we see that full agreement is tolerated (see e.g., 4), but the sentence is ungrammatical once we apply a partial agreement to the SV order (see e.g., 6). On the opposite situation, we see that if full agreement on VS structure is applied, the sentence is ungrammatical (see e.g., 7). This shows that full agreement has to apply to SV structure only whilst partial agreement has to apply to VS structure only. We need to speculate whether MA follows the same process as its descendent SA: (8) l-wlad qra-w d-dərs SV + full agreement the-boys read-3PL.M the-lesson 'The boys read the lesson.' (9) qra-w l-wlad d-dərs VS + full agreement read-3PL.M the-boys the-lesson 'The boys read the lesson.' (10) *l-wlad qra d-dərs *SV + partial agreement the-boys read-3SG.M the-lesson 'The boys read the lesson.' (11) *qra l-wlad d-dərs *VS + partial agreement read-3SG.M the-boys the-lesson 'The boys read the lesson.' Unlike the SA data, MA requires full agreement in both SV and VS structures (see e.g., 8 and 9 respectively). As soon as a partial agreement in any of SV and VS structures is applied, the sentences are ruled out (see e.g., 10 and 11, respectively). This shows that MA demands a full agreement between the verb and the subject regardless of word order. Musabhien (2008, p. 17) reports that Jordanian Arabic and many other local dialects of Arabic show this full agreement pattern, whether preverbally or postverbally.
Another distinction between SA and MA when it comes to word order is the nature of the preverbal DP. It has been argued that the preverbal DP in SA is actually a topic and the evidence comes from the fact that subjects before the verb cannot occur as indefinite DPs as shown in the following example (Soltan 2007, p. 51): (12) *waladun kasara l-ba:b-a boy-NOM broke.3SG.M the-door-ACC (13) kasara walad-un l-ba:b-a broke.3SG.M boy-NOM the-door-ACC 'A boy broke the door.' As we can see from the examples above, while it is possible to have an indefinite DP postverbally (see e.g., 13), the sentence becomes ill-formed once it is left to the verb (see e.g., 12). This is a piece of evidence that when preceded by the verb, subjects in SA are, in fact, topics. We see a different case when it comes to MA data. Let us consider these examples: entail distinct analyses, derivations, and representations. This is another reason why one also needs to focus on the different dialects of SA as distinct languages, which might offer other perspectives on linguistic phenomena. 4 SA data is taken from (Soltan, 2007) with slight modifications of the transcription to fit the one used in this paper. (14) rʒal tay-xədm-u bəzzaf men IMPV-work. PRES-3PL a lot 'Men work a lot.' (15) r-rʒʒala tay-xədm-u bəzzaf the-men IMPV-work.PRES-3PL a lot 'The men work a lot.' Example (14) shows that MA allows the appearance of indefinite DPs in the preverbal position. This shows that, unlike in SA, MA subjects are not topics. In this subsection, we introduced the two common word orders SVO and VSO. First, we stated that SVO is the natural order in MA. Second, we observed that MA disallows a partial agreement and demands full agreement between the verb and the subject. Finally, we concluded our description by stipulating that preverbal DPs in MA are subjects and not topics. In the next section, we will provide a general analysis of SVO and VSO in terms of their representation and derivation.

Representing and Deriving SVO and VSO from a Focus-based Perspective
After looking at some basic properties of the two common word orders SVO and VSO in MA, we need to ask how they are represented and derived. We have already said that the unmarked structure in MA is SVO while VSO is the derived one. Since the former is the more 'natural' word order, we need to represent the structure of MA as a postverbal language first. Let us repeat the following sentence: Assuming VP-internal subject hypothesis and VP Shell (Chomsky, 1955(Chomsky, , 1995Koopman and Sportiche, 1991;Larson, 1988Larson, , 1990 among others; also see section 3 in order to see the reasoning behind adopting this assumption), it is proposed that a basic SVO structure (e.g., like in English) has the subject moving from [Spec, vP] to [Spec, TP] and the verb moving from v to T. Noting that SVO order is derived by the subject moving from [Spec, vP] to [Spec, TP] for edge and case-related matters and the verb moving to T for tense-related reasons, we will explore the reasoning behind these choices in more details in the third section of the paper. For now, we focus on the issue of word order in MA. Let us look at the following structure: (17) As we can see from Figure 17, the verb /kla/ 'ate' moves in a successive cyclic manner from V to v and then to T (see section 3 to see arguments for verbal movement in MA). As for the subject, it moves from [Spec, vP] to [Spec, TP] 5 . Just like English, we see a normal SVO structure represented in Figure 17. The issue starts with how to derive a less natural VSO structure. This implies that the verb has moved to a projection higher than TP. The question is what is this projection and why should we adopt it? 5 We use, in accordance with MP, the Copy-Delete operation. A basic assumption of the Copy-Delete is that elements in the Spell-Out do not leave a trace. In fact, as soon as they are 'moved', their copy gets deleted as well (henceforth ħməd).  VSO structure gives a certain emphasis to the act of eating. This means that, in comparison to the SVO order which does not emphasize any element in the structure, the VSO structure implies a certain attention to the action of eating. Moreover, by fronting the verb /kla/ 'ate', the structure implies the presentation of new information, which is that of eating. Based on all of these assumptions, it seems appropriate to propose that the verb actually moves to a Focus head 6 in order to check the [Focus] feature. We will then have the following structure: (20) In this subsection, we showed that SVO is the natural order, which has the same derivation as that of English: the subject moves from [Spec, vP] to [Spec, TP] in order to check the [EPP] feature (i.e., [Spec, TP] is an unfilled position) and the verb moves from T to Focus in order to check the [Focus] feature (see 3.1 for more on evidence for V-to-v and v-to-T movement). In the next subsection, we consider an alternative analysis for the VSO order which tries to incorporate the principle of economy. 6 Note that we are not analysing the left periphery in terms of a one-layer CP, but we are adopting Rizzi's (1997Rizzi's ( , 2001) split CP hypothesis. Therefore, instead of using C, we will refer to the movement of the verb to a Focus head, noting that Rizzi (1997)

Word Order and Economy: Considering an Alternative
The Minimalist Program aims at realizing the 'simplest grammar', which is the main aim of the Generative framework since its emergence in the 1950s. In fact, Chomsky (2015) claims that the aim of any 'normal' science is to catch the simplicity of nature. This simplicity is what Chomsky embodies in the form of a Universal Grammar (UG). Indeed, it is the goal of MP to reduce UG to the minimum. One problem that might be raised in our proposed analysis is that it does not conform to the economy of derivation and simplicity. In order to derive VSO order, MA takes a further step than English and moves its verb to the Focus head. One might claim that the proposed analysis does not only have theoretical problems (i.e., too many movements and new projections), it also has empirical problems. Consider the following example: with-you 'The one that ate the apple would like to talk to you.' In (21), the nominal clause /lli kla t-tffaħa/ 'that ate the apple' is headed by the complementizer /lli/ 'that'. What is interesting about this sentence is the existence of both /lli/ 'that' and the verb /kla/ 'ate'. /lli/ 'that' is a complementizer that usually shows focus when it is inserted in the sentence; therefore, the natural occurrence of the complementizer is that of C or, in our updated split-CP version, Focus. However, we run into a serious empirical problem. We have stated that the verb, if moved, moves to the Focus head. We cannot have both the verb and the complementizer be situated in Focus. To solve this, we might speculatively say that the verb did not actually move since we cannot have a subject postverbally (i.e., we cannot produce a VSO order): kla ħməd t-tffaħa bɣa yəhdər mʕa-k That ate Ahmed the-apple want talk with-you 'The one that Ahmed ate the apple would like to talk to you. ' We might propose that the reason why both /lli/ 'that' and /kla/ 'ate' appear simultaneously is because the verb does not move (i.e., it stays in T) whilst the complementizer /lli/ (that) is base-generated in Focus. The evidence comes from the fact we cannot have the verb /kla/ 'ate' be positioned before the subject /ħməd/ 'Ahmed'. However, one might say that further data shows that we cannot have the subject /ħməd/ (Ahmed) preverbally as well (i.e., before the subject) as seen in the following examples: (23) *lli ħməd kla t-tffaħa bɣa yəhdər mʕa-k That Ahmed ate the-apple want talk with-you 'The one that Ahmed ate the apple wants to speak with you.' If we would like to have the subject in the sentence along with the verb and the complementizer, the correct construction is in the following example: (24) ħməd lli kla t-tffaħa bɣa yəhdər mʕa-k Ahmed that ate the-apple want talk with-you 'Ahmed that ate the apple wants to speak with you.' Assuming that /lli/ 'that' heads a C or, in our proposal, a Focus head, this means that the subject will be situated in [Spec, FocusP] whilst the complementizer /lli/ 'that' heads the Focus Phrase. However, that does not mean that the verb /kla/ 'ate' has to be situated in Focus head. We propose that, in this situation, the verb /kla/ 'ate' is simply positioned in T 7 . Therefore, we arrive at the following structure 8 : 7 The proposal that the complementizer /lli/ 'that' and the verb /kla/ 'ate' are in Focus and T heads respectively is strengthened more based on empirical data in section 3 on verbal movement. 8 It is worth to note that the structure is not complete, and it makes us wonder how would we derive a full structure like 'Ahmed that ate the apple is nice'? Where would the Focus Phrase be positioned? Is it a specifier to a [Spec, TP]? Aoun and Li (2003, p. 6) describe head-initial relative constructions (like English and MA) as "selected by (a complement of) a determiner D: [DP D CP]." We leave the issue of relative construction aside since our main focus is to discuss word order and its implication on verbal movement, and we propose that the FocusP is the specifier of [Spec, TP]. This goes in line with the fact the fact that relative constructions that define a noun and wh-questions that appear with the complementizer have similar semantic and pragmatic implications; this entails that they should have a similar analysis as well (see section 4 for more on the interaction between wh-elements, complementizers, and verbal movement). (25) It should be noted that the proposed analysis is also pragmatically natural. The complementizer /lli/ 'that', when present, provides more focus to the subject /ħməd/ 'Ahmed' whilst the act of eating is just secondary; consequently, it is /lli/ 'that' that will head the Focus projection. One might try to claim that we can provide a simpler and more economical representation of the SVO/VSO alternation without positing the need for a Focus projection. Indeed, one might say that the subject can stay in [Spec, vP] whilst the verb stays in v; this will give us the natural SVO structure of MA. Moreover, when we need to derive VSO structure, the verb can simply move from v to T as seen in the following two figures respectively: This analysis creates more theoretical problems than it solves them. If we posit that T is strong in MA. Then, the verb has to move from v to T all the time. If we say that T is weak, then the verb will stay in-situ all the time. Indeed, theoretically speaking, we cannot posit that the T head is strong at times (see section 3 for more on feature checking and empirical arguments on verbal movement) and weak at other times (i.e., no movement happens). Therefore, we keep the proposal that VSO is the outcome of the verb moving from T to Focus. In the next subsection, we consider other marginal word orders apart from SVO and VSO and see if our current proposal fits the other word orders.

Some Considerations on the Other Marginal Word Orders
We stated that the SVO order is the unmarked structure in MA, which goes through the same process as its English counterpart: the subject moves from [Spec, vP] to [Spec, TP] whilst the verb moves from v to T. We also said that the VSO order is an additional step where the verb moves from T to Focus head due to pragmatic reasons (i.e., to add emphasis and present new information). Now we need to present other word orders that are less commonly used in MA and see if we can retain the analysis proposed in 2. In both cases, the object /t-tffaħa/ 'the apple' is fronted to the left periphery. Since it can be positioned before the verb /kla/ 'ate', then it must move higher than Focus Phrase. In terms of the interpretation of the sentence, the object /t-tffaħa/ 'the apple' presents an old information that is known to the addressee. According to Rizzi (1997Rizzi ( , 2001, this projection is Topic Phrase. Indeed, the object /t-tffaħa/ 'the apple' is topicalized to [Spec, TopicP] in order to show emphasis. One question is how does the structure keep the interpretation of the object as a complement of the verb? The answer lies in the existence of the resumptive pronoun /-ha/ which agrees in person, number, and gender with the topicalized object. Syntactically speaking, the object moves to [Spec, TopicP] and forms a chain with the resumptive pronoun /-ha/ and binds it. Note that removing the resumptive pronoun deems the sentence ill-formed, which means it is necessary to have the resumptive pronoun: (30) *t-tffaħa kla ħməd the-apple ate.PAST.3SG.M Ahmed 'the apple, Ahmed ate.' We will get the following representation for (31): (31) In (31), this is a case of an OVS structure: the verb moves in a successive cyclic manner from V all the way to Focus; the subject moves from [Spec, vP] to [Spec, TP]; most importantly, the direct object, moves from the complement of the verb to [Spec, TopicP] and leaves behind a resumptive pronoun. OSV is the least natural word order in MA. In this situation, the verb stays in T instead of going an extra step from T to Focus; the reason behind this is that, in this situation, the verb does not have any function of emphasis or adding any new information. In the next section, we investigate evidence of verbal movement starting from V.

Evidence for Verbal Movement in Moroccan Arabic
In this section, we explore arguments that support verbal movement in MA. In 3.1, we investigate the evidence for the movement of the verb from V to small v using the adverbial position evidence. After that, we argue for the movement of the verb from v to T based on the evidence of floating quantifiers and feature checking. Finally, in 3.2, we summarize the verbal movement from T to Focus based on empirical reasons (i.e., word order).

Evidence for v-to-T Movement
Before we discuss v-to-T movement, let us argue for the movement of the verb from V to small v. Pollock (1989) posits that verbal movement is based on the position of verbs with regard to adverbs. When the latter follows the verb, we say that the verb has moved. When the adverb precedes the verb, we say that the verb has not moved. We can see the following differences between English and French: The construction in (32) is deemed ill-formed when we force verbal movement in English. In (33), the verb /mange/ 'eat' moves leaving the adverb /souvent/ 'often' in [Spec, VP] (see also Emonds, 1978 for more on V-to-T movement in French). In MA, we can also explain verbal movement based on the position of the adverb: In (34), the verb /tajlʕəb/ 'plays' precedes the adverb /dima/ 'always'. This provides evidence that the verb has moved leaving the adverb phrase in [Spec, VP]: (35) In (35), the adverb /dima/ 'always' is situated in [Spec, VP]. Since the verb /tajlʕəb/ 'plays' is preceded by the adverb /dima/ 'always', then the verb has moved from V to small v. Now, let us see the evidence of the further movement of the verb from v to T.
Let us look at the following examples in French to introduce the theory of floating quantifiers we adopt (Sportiche, 1988 (36) shows the quantifier in its default position (i.e., close to the noun). However, in (37), /tous/ 'all' is not immediately close to the noun. In fact, it seems to be inside the vP. According to Sportiche (1988), the quantifier in (37) While the QP moves to [Spec, TP] as seen in (40), only the DP /d-drari/ 'the boys' moves to [Spec, TP] leaving the quantifier /kulhum/ 'all' in the Q head as seen in (41). The latter Figure is clear, empirical evidence that v actually moves to T.
In the next subsection, we explain v to T movement on theoretical grounds: feature checking.

v-to-T Movement Revisited: Feature Checking
Throughout the paper we argued for the verbal movement using the adverbial position (for V-to-v movement), floating quantifiers (for v-to-T movement), and VSO/SVO order (for T-to-Focus movement; see section 3.3. for one last look at the interaction between this movement and word order) as three pieces of evidence. Now, we shall offer one last evidence stipulated by Chomsky (1995). Feature strength and interpretability also provide reasons for such movement 10 . Before we speak about the application of verbal movement from the perspective of MP featural checking, we introduce some key concepts on features and the operation of Move. Let us first look at features: As is known to all, a lexical item is a collection of phonological features (relevant at the PF), semantic features (relevant at LF), and morphosyntactic features, which include formal features (relevant at syntax). More importantly, features account for parametric variation (Chomsky, 2015, p. 62) 11 . Formal features that we need include categorial features; strong and weak 10 Although VSO/SVO order, floating quantifiers, and adverbial position can be considered empirical evidence, feature strength and interpretability are theoretical evidence that give motivations for the movement of the verb. 11 It should be noted that the reason why MP is concerned more with formal features than phonological or semantic features is that if we take a word like 'boy,' we are more interested in the fact that it is a noun than that it starts with a plosive consonant or is a When it comes to the categorial features, the categorial distinction between nouns, verbs, adjectives, and prepositions can be looked at in terms of two sets of categorial features. These are [±V] (i.e., verbal or non-verbal) and [±N] (i.e., nominal or nonnominal) (Chomsky, 1970;Jackendoff, 1977). Therefore, we can describe the syntactic categories in the following categorial features ( When it comes to strong and weak features, a strong feature is a feature that needs to be immediately eliminated as soon as it is introduced in the overt syntax and this is done by bringing a relevant feature next to it. These kinds of features force overt movements. As for a weak feature, it is a feature that does not need to be eliminated as soon as it is introduced in overt syntax; in fact, it is deleted after the Spell-Out. These features concern themselves with covert movement 12 . For example, auxiliary verbs in English, in sentences like 'he did not understand', carry a strong tense feature; therefore, they need to move overtly to T in order to check their features. However, in English, main verbs carry a weak tense feature; therefore, they move after the Spell-Out (see Pollock, 1989 for more information on V-to-T movement in English).

When it comes to interpretable and uninterpretable features, the distinction between interpretable and uninterpretable features (henceforth IF and UF respectively) is the most important. Chomsky (2015) states in the following statement:
There is a much more important distinction that has so far been overlooked. Evidently, certain features of FF(LI) enter into interpretation at LF while others are uninterpretable and must be eliminated for convergence. We therefore have a crucial distinction ±interpretable. Among the Interpretable features are categorial features and the ϕ-features of nominal. (p. 255).
IF are features that have an effect on semantic interpretation; therefore, IF must reach the LF whilst UF are features that do not have an effect on semantic interpretation. Consequently, UF get prevented from reaching the LF. Therefore, UF are relevant at the PF whilst IF are relevant at the LF. For example, in a sentence like /ħməd kla t-tffaħa/ 'Ahmed ate the apple', the verb /kla/ 'ate' has the feature [+N] in order to indicate that it requires the object /t-tffaħa/ 'the apple'. However, the [+N] feature on the verb is uninterpretable and, if kept, the derivation will 'crash' at the LF (Chomsky, 2015, p. 255). The object /t-tffaħa/ 'the apple' has an interpretable [+N] feature in order to ensure a 'nominal' interpretation of the object. The interpretability of features is the most important in realizing well-formed derivations. In fact, the Full Interpretation Principle can be summarized in the following statement (Chomsky, 2015): The principle FI is assumed as a matter of course in phonology; if a symbol in a representation has no sensorimotor interpretations, the representation does not qualify as a PF representation. This is what we called the 'interface condition.' The same condition applied to LF also entails that every element of the representation have a (languageindependent) interpretation. There can, for example, be no true expletives, or vacuous quantifiers, at the LF level. (p. 24).
This means that any element that is present at any interface (i.e., PF and LF) must receive a relevant interpretation or licensing. To clarify, at the LF, an element must contain all and only the semantic features that would determine its meaning, whilst, at the PF, an element must contain all and only the phonetic features that would determine its pronunciation (Radford, 1997, p. 261). As for the operation of Move 13 . It is constrained by an economy principle named the Last Resort principle 14 , which is "(a) condition of movement (which) requires that movement is permitted only to satisfy some condition" (Chomsky, 2015, p. 41). This principle offers motivations for movement. In GB, the reasons for movement were ambiguous. However, in MP, we only move an element in order to check features 15 .
In section 3.1., we stated that the verb moves v to T based on the quantifier evidence. Now, we will give a grammatical reason why a verb needs to move from v to T. To answer this question, we need to refer to Chomsky's (2005) feature inheritance model. It is simplified in the following statement (Musabhien, 2008): T is finite only when it is selected by C. C, the head of the CP phase, is the source of the tense feature and the I-features on T. This means that, as Chomsky assumes, it is C that probes down via T and agrees with the goal subject. Put differently, C and T form a complex that functions as a probe. The valuation of Case on the subject as nominative is an outcome of the Agree relation between C-T and the subject. (p. 112).
Since we are using Rizzi's (1997Rizzi's ( , 2001  After we talked about the workings of the verb on T, the next section revisits the movement of the verb from T to Focus using the word order evidence.  (Chomsky, 2000(Chomsky, , 2005. In this paper, we will not focus too much on the distinctions between these versions (i.e., early Minimalism, Agree-based approach, and Phase-based approach). We retain the main idea of analysing word order and verbal movement using the premises of feature checking. 14  According to Greed, we cannot move 'John' to matrix [Spec, TP] since the latter has already checked its Case feature. 15 Feature checking is an operation which entails that two interrelated elements share a given feature but only one element needs that feature to interpret its meaning; therefore, the 'unneeded' or uninterpretable feature gets deleted (Chomsky, 2015). For example, when a verb and a subject share a number feature, only the subject needs it. Therefore, the [+singular] feature that is uninterpretable on the verb gets checked and deleted against the [+singular] interpretable feature on the subject.

T-to-Focus Movement Revisited: Word Order
In figure (42a), the subject stays in [Spec, VP] forming a VSO order whilst, in (42b), the subject moves to [Spec, TP] forming an SVO order. In order to have an SVO/VSO variation, the verb must always move to T leaving the option of the subject. If there is no V-to-T movement at all, we can only have the SVO order (i.e., whether the subject moves or not, it will still generate SVO order). Alotaibi (2013) also treats the movement of the verb from V to T as a result of VSO/SVO alternation in Modern Standard Arabic. In addition to that, he posits that MSA has an optional [EPP] feature on T. The [EPP] feature optionality can also be applied to MA since it allows both orders. When the [EPP] feature is present in T, the subject moves to the specifier of TP; however, when the [EPP] feature is absent, the subject stays in the specifier of VP 17 (Alotaibi, 2013; see also Chomsky, 2000).
Since Fassi Fehri's (1989) analysis applies to SA which has VSO as its unmarked structure and SVO as the derived one, his work argues for V-to-T movement analysis. I apply the same proposal to argue for the movement of the verb from T to Focus based on word order alternation. Therefore, SVO is the outcome of the verb staying in T whilst the subject is positioned in [Spec, TP]. VSO is the outcome of verb moving further to Focus. If we choose not to move the verb to Focus, we will risk having only the SVO order.
We know that, in the literature of movement, T-to-C (in our case T-to-Focus) movement has implications on the pragmatic force of the sentence (e.g., if it is a yes-no question or not). In the next section, we explore the interaction between T-to-Focus movement and interrogativity, specifically wh-questions.

Is there a T-to-Focus Movement in MA Interrogative Structures?
In this section, we verify a possible T-to-Focus movement in wh-question constructions empirically reaching the conclusion that such movement happens but with some exceptions.

Wh-subjects and T-to-C Movement
It is established in the literature that wh-subjects move to [Spec, CP] (Chomsky, 1995). Let first us consider an example where the complementizer /lli/ 'that' does not occur. First, the wh-subject is initially situated in the specifier of vP where it moves to the specifier of TP in order to check the [EPP] feature. After that, the wh-subject will move to the specifier of FocusP in order to check the [Q] (or [WH] 18 ) and the [EPP] features as well (i.e., the specifier of FocusP is an unfilled position) (Chomsky, 1995). Note that the stipulation that wh-questions move to [Spec, FocusP] is not new and many have argued for it in many Arabic dialects (e.g., Gad, 2011 for Egyptian Arabic, who state that wh-phrases bear contrastive focus feature by virtue of supposing the existence of a set of entities present in the discourse; Al-Momani & Al-Saidat, 2010 for Jordanian Arabic, who argue for contrastive focus in the case of wh-movement and that focus is checked via an intonational morpheme if wh-elements are in-situ). Second, the verb forms a Focus-T complex and checks the [TNS] and [V] features. Since the feature is strong, it pied-pipes along with it the verb before the Spell-Out. The problem that 17 Optionality has long been a controversial issue in generative grammar, and especially with Minimalism's feature checking. It might be considered that the presence and absence of features, which either dictate movement or not seems to be an ad-hoc reasoning. We leave this theoretical issue aside as it falls outside the scope of our paper. 18 We refrain from discussing the differences between [  arises here is whether the verb /kla/ 'ate' actually stays in T or moves further to Focus in the case of wh-questions. The following two figures illustrate the two possibilities where Figure (44a) shows that the verb stays in T whilst, in (44b), the verb moves from T to Focus (i.e., T-to-Focus movement) (wh-subject movement is irrelevant here): (44) Nouhi (1996, p. 116) posits that, in MA, the verb moves from T to AgrS and then to C as a consequence of the interaction between verb movement and wh-movement. Theoretically speaking, we cannot allow for the verb to sometimes sit at T and then at other times move at Focus unless there is a strong reason for doing so (e.g., an element already fills that position). In addition to that, we speculate that a feature such as [Focus] can either be strong or weak, which automatically either forces the displacement or not, respectively. Moreover, and empirically speaking, there are situations where the verb clearly stays in T, but that is only because there are other items which satisfy featural needs of Focus. For example, another item that can fill the position of Focus is the complementizer /lli/ 'that'. Consider this example: (45) škun lli kla t-tffaħa who that ate the-apple 'Who is the one that ate the apple?' MA is a language that resists Doubly Filled COMP Filter (Chomsky & Lasnik, 1977), which states that we cannot have the presence of both a wh-element in [Spec, FocusP] and an overt complementizer. In MA, and as seen in (45) [Spec, FocusP] when there is no complementizer /lli/ 'that'. This raises the issue of the tension that happens between uniformity and economy. On the one hand, if we want to achieve the former, all whelements will move to [Spec, FocusP] in order to provide a contrastive reading of the question. On the other hand, if we want to achieve economy, the wh-subject does not have to move to [Spec, FocusP] since there is no complementizer /lli/ 'that' and the verb is not really in the left-periphery. We propose that if we have wh-subjects that appear in WH-SV order, they are derived the same way as their declarative SVO counterpart: the verb stays in T and the wh-subject stays in [Spec, TP] (see Figure 46 below); however, when the wh-subject is present with /lli/ 'that', then it further moves to [Spec, FocusP] to get a contrastive focus reading whilst /lli/ 'that' stays in Focus (see Figure 47 below): In this subsection, we looked at the status of wh-subjects in relation to word order. The next section discusses the problem of wh-object construction in relation to word order.

Wh-objects and T-to-C Movement: the Problem of Word Order
This subsection explores the claim that there exists a T-to-Focus movement in wh-object construction while taking into consideration the problem of word order. In section 4.1, we stated that there was a theoretical implication for the movement of the verb from the T position to Focus when it comes to wh-subjects. Wh-subjects move to [Spec, TP] while the verb moves to T in order to satisfy pragmatic and featural needs. We discussed that when a complementizer is present, it takes the Focus head whilst the verb stays in T and the wh-subject moves to [Spec, FocusP]. However, let us explore, in this section, if there is any empirical evidence that will show the non-movement in wh-object construction. Let us look at the following example:  Alotaibi (2013) notices that, in MSA, it is not possible to have wh-objects in an SV order. He states that the ungrammaticality of wh-objects in SV order is a failure of subjects to move. He explains this ungrammaticality based on agreement adopted from Chomsky (1999, p. 9) who states that the EPP can only attract the subject if T is φ-complete (i.e., carries a complete set of agreement features: number, person and gender). However, if T is φ-incomplete (i.e., when MSA only displays partial agreement as in number agreement), then the subject won't move. However, Alotaibi (2013) admits that this analysis will predict the following sentence as grammatical: (i)  As we can see from (50), it is possible to have a WH-SV order. As discussed earlier in the paper, SVO order is the result of the movement of the subject from [Spec, vP] to [Spec, TP] whilst the VSO order is the result of the verb moving from T to Focus. We see a problem here. In order to derive a WH-SV order, we are forced to leave the verb in T. To solve this issue, it should be noted that WH-SV is less natural to Moroccan speakers than WH-VS. In fact, some speakers might be puzzled and even question whether it is grammatical in the first place. To clarify, when we have WH-SV order (and WH being a wh-object), it seems like this is known information to the speaker and it is topicalized to show emphasis. Indeed, we argue that the subject /ħməd/ 'Ahmed' moves to [Spec, TopicP] whilst the verb will move to Topic to check the [Topic] feature, noting that the verb here does not denote any contrastive focus reading; consequently, FocusP will dominate [Spec, TopicP]. We reach the following representation when it comes to WH-SV order (where WH is a wh-object): (51) As we have said, we find that the verb stays in T in the case of the complementizer, just like wh-subjects as well. Let us look further at the following example: As we can see in the figure above, the wh-object moves to [Spec, FocusP] whilst /lli/ 'that' heads Focus, which gives us a contrastive focus reading. Since /lli/ 'that' fulfils the pragmatic needs of Focus, the verb does not have to move and stays in T. The next subsection deals with the final type of wh-elements: wh-adverbials.

Wh-adverbials and T-to-Focus Movement
In this subsection, we look at T-to Focus movement when it comes to wh-adverbial construction. Just like wh-subjects and whobjects, we use word order consideration in order to argue for the movement analysis of the verb from T-to-Focus in whadverbial construction. Consider the two examples below: (54) fin qra ħməd? where studied.PAST.3SG.M Ahmed 'Where did Ahmed study?' (55) fin ħməd qra? where Ahmed studied.PAST.3SG.M 'Where did Ahmed study?' Just like in wh-object constructions, it is possible to have both Wh-VS order (see 54) and Wh-SV order (see 5(). We propose the same analysis discussed in 4.2, and we posit that, in wh-adverbial constructions, the wh-adverbial moves to [Spec, FocusP] whilst the subject moves to [Spec, TopicP] which hosts the verb as the head of Topic in the case of WH-SV order. In the case of WH-VS order, the wh-adverbial will move to [Spec, FocusP] whilst the verb will move from T to Focus in order to check the [Focus] feature. Wh-adverbials do not have the option to appear with the complementizer /lli/ 'that' and that is due to the fact that complementizers, in general, relate to DPs and nominals and never adverbials.

Conclusion
In this paper, we explored the correlative relationship between word order and verbal movement in MA using the Minimalist framework. In the first and second sections, we observed that the natural order in MA is SVO; therefore, this will follow an English-like derivation where the verb moves from v to T while the subject moves from [Spec, vP] to [Spec, TP]. Since VSO is the derived version of SVO, it is only natural to stipulate that the verb has moved to a higher projection than TP. We proposed that this projection is that of the Focus Phrase. The reason behind this is pragmatic or discursive. The verb in the left-periphery denotes new information that is focused; automatically, the verb, in VSO, moves to a Focus head. We also tested this new proposal to other marginal word orders: OSV and OVS; we proposed that the object topicalization is the result of the object moving to [Spec, TopicP], which leaves behind it a resumptive pronoun that binds it. In the third section, we went back to the issue of verbal movement and traced the successive cyclic movement of the verb: [1] V to v, [2] v to T, and [3] T to Focus. For [1], the verb moves from V to v using the position of the adverb: adverbs are in [Spec, VP] while we see the verb is positioned to the left of adverb, hinting that it went beyond VP. For [2], the verb moves to T based on theoretical and empirical evidence. Theoretically, the movement is motivated by the need to form a Complex with Focus and check the features of [TNS] and [V] on the goal v. Empirically speaking, the verb moves to T based on the fact they can move over quantifiers that usually head the

FocusP
Quantifier Phrase that is situated in [Spec, vP]. Finally, we went back to the issue of T-to-Focus stating that this movement is essential in order to derive VSO order. In the last section, we explored the interaction between T-to-Focus movement and whquestions. We reached the following conclusions: First, and when it comes to wh-subject constructions, these wh-elements are derived in WH-SV order the same way as in the declarative SV structure: the verb is in T whilst the wh-subject stays in [Spec, TP]; however, when the wh-subject is situated with the complementizer /lli/ 'that', the former moves further to [Spec, FocusP] which is headed by /lli/ 'that'. Second, the less natural WH-SV derivation for both wh-objects and wh-adverbials is different: the subject moves to [Spec, TopicP], the verb moves to Topic, and the wh-elements move to [Spec, FocusP]. In a normal WH-VS, all whelements move to [Spec, FocusP] while the verb moves to Focus. We see the following order of projections in the Split CP: Topic Phrase1 > Focus Phrase > Topic Phrase2. The first Topic Phrase hosts topicalized objects whilst the second Topic Phrase hosts topicalized subjects. There is still further research to do when it comes to word order and verbal movement, especially when we consider other constructions such as yes-no questions, negation, embedded clauses, to name but a few.
Funding: This research was funded by the National Center for Scientific and Technical Research, grant number 2UIT2020.