Arabic Semantics Peter Hallman, University of Vienna

1.

INTRODUCTION

The development of model-theoretic semantics in the late 1960’s, particularly the work of Montague (1970, 1973 and elsewhere), has lead to striking progress in the description of the derivation of meaning in natural language. This framework implements the principle of compositionality, usually attributed to Gottlob Frege, which states that the meaning of each constituent in a syntactic structure is derivable from the meanings of its own constituent parts by regular rules. This framework has proven powerful in describing the relationship between form and meaning in natural language. Since the development of this framework, Arabic has not been subject to substantial formal semantic analysis, meaning that Arabic still stands to make substantial new empirical and theoretical contributions to the development of the field of semantics. Some issues that have received attention are described in section 2 below, while section 3 describes some aspects of Arabic that appear to represent fertile territory for future semantic inquiry.

2.

ISSUES

The issues that have received the most attention in the literature on Arabic semantics relate primarily to the semantic correlates of syntactic and morphological patterns. Literature on the syntax-semantics interface deals primarily with the interpretation of A’-chains, the dependency between a clause-peripheral operator such as a wh-phrase

or topicalized phrase and the semantic variable that it binds. Literature on the morphology-semantics interface deals primarily with the interpretation of verbs, focusing on tense and aspectual distinctions as well as derivational verb morphology. Another nacent area of semantic inquiry on Arabic is pragmatics, which deals with the relationship between linguistic content, discourse context and communicative intent. In this section, I describe the current state of the art on the interpretation of syntactic displacement in Arabic, tense and aspect, derivational verb morphology, and pragmatics.

2.1

Interpretation of Displacement

A variety of observations indicate that some syntactic displacement operations affect the interpretation of the expression they apply in, while others do not. More specifically, when a constituent is displaced from position A to position B in a given construction, the construction may require it to be interpreted in position B (the derived position), or it may require it to be interpreted in position A (the base position), or it may allow either interpretational option. Another dimension to this pattern is that the displacement itself may be covert. That is, the constituent may be interpreted in position B while occurring overtly in position A. The first subsection below reviews a body of literature on A’-chains in this connection. The subsequent section augments this literature with some observations about covert quantifier raising, an operation bearing a closer resemblance to Amovement.

2.1.1

Topicalization and Reconstruction

In Arabic, a sentence-initial topic may be linked to a pronoun in the comment, as the Lebanese examples in (1) show, from Aoun and Benmamoun (1998). Non-subject pronouns are enclitics morphologically bound to their governor. As the examples in (1) demonstrate, the dependency between a topic and the associated pronoun is not obviously island-sensitive, crossing over the adjunct island boundary in (1b), a constituent from which movement is illicit in English.

(1)

a.

təәlmiiz-a ʃ-ʃitaan btaʕrf-o [ʔəәnno kəәll mʕallme ʔaaaṣaṣəәt-o]. student-her the-naughty know.2S that every teacher.F punished.3SF-

him ‘Her naughty student, you know that every teacher punished him.’

b.

təәlmiiz-a ʃ-ʃitaan fallayto [ʔablama kəәll mʕallme t-ʔaaaṣəәṣ-o]. student-her the-naughty left.2P before every teacher.F 3SF-punished-

him ‘Her naughty student, you left before every teacher punished him.’

In each example, the phrase təәlmiiz-a ʃ-ʃitaan (her naughty student) stands in a syntactic dependency with the object inflection -o (him), which is contained in a complement clause in (1a) but in an adjunct clause in (1b) (the bracketed constituent in each case). Generally, adjunct clauses but not complement clauses are syntactic islands, constituents from which nothing may be extracted (Ross 1967). Aoun and Benmamoun observe that the topic təәlmiiz-a ʃ-ʃitaan may be construed within the semantic scope of the quantifier kəәll mʕallme in (1a) but not in (1b). Specifically, (1a)

may be read as asserting that you know that each teacher punished her respective naughty student. That is, the naughty students vary with the teachers. This reading is not available when the bound object inflection -o is contained in an island in (1b), which may only be understood to assert that there is one particular naughty student who every teacher punished, and you left before this happened. Aoun and Benmamoun conclude on the basis of this and other evidence that on the relevant reading of (1a), the topic occurs in the backeted constituent containing the object inflection -o at the level of representation at which the structure of the sentence is ‘fed’ to the interpretative component, which is the representation that reflects the meaning of the sentence. Topicalization and insertion of the inflection are superficial, ‘post-semantic’ processes that do not impact quantifier scope. Another derivational source is available where the topic is base generated in the clause-initial position prior to the point at which the structure is interpreted, and binds a null pronoun in the position indexed by the -o inflection. The island in (1b) prohibits the movement derivation, so that the only derivational source for (1b) is base generation of the topic in the clause-initial position. This dependency does not show a semantic scope ambiguity because it is derivationally unambiguous. Guilliot and Nouman (2006) note that Jordanian Arabic shows scope ambiguities over island boundaries, but only when the resumptive element is morphologically weak. They develop an analysis that links the possibility of resumption without movement to the morphological weakness of the resumptive pronoun. Cross-linguistic variation in the availability of this reading indicates that the interpretational properties of A’-chains discussed above are subject to variation.

2.1.2

Quantifier Raising

In this section, I expand on the observations described above on scopal interactions in A’-movement/topicalization configurations by describing local (intra-clausal) scopal interactions between quantifiers. The judgments described below represent relative trends in speakers’ attitudes toward specific interpretations of various word order configurations in standard Arabic. Variation in these judgments exists, but contrasts between pairs of sentences in the relative availability of a scopal configuration are consistent. For example, in the VSO word order, a quantificational subject easily scopes over an indefinite object, illustrated in (2a). However, a quantificational object may not readily take scope over an indefinite subject that precedes it, as in (2b). (2b) is inappropriate in a context in which each vase is broken by a different child. That is, inverse scope (where two terms show a scopal order that is the inverse of their surface linear order) is restricted in Arabic. The examples below reflect the fact that the quantifier kull (every) is more natural when its nominal associate is modified, for example by a min (of) phrase below, especially in object position. The -n declension (‘tanwiin’) on indefinite nouns below is glossed N.

(2)

a.

kassar-a kull-u walad-i-n min l-ʔawlaad-i mazhariyyat-a-n broke-3MS every-NOM child-GEN-N of the-children-GEN vase-ACC-N ‘Every one of the children broke a vase.’

b.

kassar-a walad-u-n kull-a mazhariyyat-i-n min al-mazhariyyaat-i broke-3MS child-NOM-N every-ACC vase-GEN-N of the-vases-GEN ‘A child broke every one of the vases.’

The impossibility of inverse scope in (2b) is unlike English, where an object may scope over a preceding subject; the English translation of (2b) admits a reading by which each vase is broken by a different child. This reading is said to be derived by covert movement of the object to a position preceding the subject, a syntactic transformation known as ‘quantifier raising’ (May 1985). Arabic allows such a permutation in its surface word order, shown in (3). On this word order, a quantificational object readily scopes over following indefinite subject; (3a) may be understood to assert that each vase was broken by a different child. In addition, though, a quantification subject may scope over a preceding indefinite object; (3b) may be understood to assert that each child broke a different vase. That is, in the VOS order, the surface scopal order is available, as in the VSO order, but also the inverse scopal order is available, unlike the VSO order. Therefore, a reading in which the subject scopes over the object is available regardless of the surface order of the two arguments, but a reading in which the object scopes over the subject requires linear precedence of the object. This in turn suggests that the syntactic prominence of the subject over the object is structurally basic, and therefore inherited by various word order permutations, whereas prominence of the object over the subject is derived.

(3)

a.

kassar-a kull-a mazhariyyat-i-n min l-mazhariyyaat-i walad-u-n broke-3FS every-ACC vase-GEN-N of the-vases-GEN child-NOM-N ‘A child broke every one of the vases.’

b.

kassar-a mazhariyyat-a-n kull-u walad-i-n min l-ʔawlaad-i broke-3FS vase-ACC-N every-NOM child-GEN-N of the-children-GEN ‘Every one of the children broke a vase.’

The generalization described above is identical to facts reported by Mohammad (1984) about coindexation in Arabic. When an object occurs with a pronominal enclitic, the clitic can be coindexed with the subject regardless of word order (4a-b). Both are grammatical on the reading where -hu refers to the subject ʔaħmad.

(4)

a.

rama-a ʔaħmad-u-n kurat-a-hu threw-3MS Ahmad-NOM-N ball-ACC-his ‘Ahmadi threw hisi ball.’

b.

ramaa kurat-a-hu ʔaħmad-u-n threw-3MS ball-ACC-his Ahmad-NOM-N ‘Ahmadi threw hisi ball.’

However, when the subject bears an enclitic pronoun, the pronoun may be coindexed with the object only when the object precedes it. In (5a), therefore, where the subject (including the pronoun) precedes the object, the pronoun -hu cannot be interpreted as referring to ʔaħmad, while in (5b), where the object precedes the subject, it may.

(5)

a.

*ḍarab-a ṣadiiq-u-hu ʔaħmad-a-n hit-3MS friend-NOM-his Ahmad-ACC-N (‘Hisi friend hit Ahmadi.’)

b.

ḍarab-a ʔaħmad-a-n ṣadiiq-u-hu hit-3MS Ahmad-ACC-N friend-NOM-his ‘Hisi friend hit Ahmadi.’

This parallelism suggests that in Arabic, the mechanism that establishes co-reference is related to that which establishes quantifier scope. This mechanism is tied closely to syntactic prominence. A constituent has scope over, and may identify a pronoun in, a constituent that is structurally subordinate to it. See Reinhart (1976), Chomsky (1981) and May (1985) on this notion of prominence, where it is identified with the syntactic ‘c-command’ relation. In Arabic, syntactic prominence for the purposes of quantification and coreference may be established with reference to a basic structure in which the subject is more prominent than the object, or the surface structure, in which the object may be more prominent than the subject (as in (3), (4b) and (5b) above). Since the subject has structural prominence over the object in the base structure, prominence of the object over the subject must be derived by displacement of the object to a higher structural position. In Arabic, this displacement must be overt, while English allows covert displacement.

One clear interpretational parallel between Arabic and English, though, is the ‘scope freezing’ effect found in double object constructions in both languages (see Larson 1990, p. 605 on English). The verb ʔaʕṭaa (give) has two internal objects, a theme and a possessor who comes into possession of the theme by virtue of the action described by the verb. As in English, the two internal objects may either appear in the ‘double object’ frame [V possessorACC themeACC] or in the ‘prepositional’ frame [V themeACC li-possessorGEN], where the theme is introduced by the preposition li- (to). The scope of the two objects is ‘frozen’ in the surface order in the double object frame but not the prepositional frame. Like its English translational equivalent shown below, the double object frame in (6a) may not be read as asserting that each toy in the cupboard

was given to a different child, but rather only that there is a particular child who was given every toy. In the prepositional frame, however, the possessor may have inverse scope over the theme that linearly precedes it. An interpretation is available to (6b) that asserts that every child was given a different toy, again like its English equivalent. Arabic mirrors other languages in the conditions that govern the availability of inverse scope in multiple-object constructions.

(6)

a.

ʔaʕṭay-tu walad-a-n kull-a luʕbat-i-n fii l-xazaanat-i gave-1S child-ACC-N every-ACC toy-GEN-N in the-cupboard-GEN ‘I gave a child every toy in the cupboard.’

b.

ʔaʕṭay-tu luʕbat-a-n li-kull-i walad-i-n fii ṣ-ṣaff-i gave-1S toy-ACC-N to-every-GEN child-GEN-N in the-class-GEN ‘I gave a toy to every child in the class.’

The possibility of inverse scope in (6b) extends to prepositional arguments in general, not just those related to the double object frame. A quantifier in a locative prepositional phrase may scope above the direct object, as in (7a), which asserts that I put a different book on each shelf.

(7)

waḍaʕ-tu kitaab-a-n fii kull-i raff-i-n min l-rufuuf-i put-1S book-ACC-N in every-GEN shelf-GEN-N of the-shelves-GEN ‘I put a book on every one of the shelves.’

Therefore, English and Arabic are largely similar except for the unavailability of an inverse scope reading of the object over the subject in Arabic when the subject precedes the object. ‘Inverse scope’ in this case must be accompanied by actual surface inversion of the two arguments in Arabic. This might be taken to suggest that the relative restrictiveness of covert scope inversion in Arabic is a by-product of its relative freedom of word order, while the relative freedom of covert scope inversion in English is a by-product of its relatively rigid surface word order.

2.2

Tense and aspect

Another area in which semantic inquiry has touched on Arabic relates to the interpretation of temporal and aspectual verb morphology. Though traditional Arabic grammarians characterize the imperfective (8a) and perfective (8b) verb forms in Arabic as signifying the present and past tense respectively (Sibawayhi 2009, vol. 1, p. 69 and elsewhere), some contemporary authors have characterized the distinction as an aspectual ‘incompletive’ vs. ‘completive’ opposition (Cohen 1924, Cantineau, 1953, Wright 1981, and others).

(8)

a.

qaraʔ-a l-walad-u l-kitaab-a readPERF-3MS the-boy-NOM the-book-ACC ‘The boy read the book.’

b.

ya-qraʔ-u l-walad-u l-kitaab-a 3ms-readIMPF-IND the-boy-NOM the-book-ACC ‘The boy is reading the book.’

The aspectual view receives some preliminary support from the fact that the perfective morphology occurs in contexts other than past tense, and the imperfective in contexts other than present. The perfective may occur, for example, in the context of a future copular auxiliary, expressing the future perfect, as in (9). If tense relates the time of the eventuality described directly to the speech time, this fact is unexpected.

(9)

saufa ya-kuun-u qaraʔ-a l-walad-u l-kitaab-a FUT 3MS-beIMPF-IND

readPERF-3MS the-boy-NOM the-book-ACC

‘The boy will have read the book.’

However, Comrie (1976) and Fassi Fehri (2003a, 2004) argue that examples such as (9) demonstrate instead that tense interpretation is locally relative in Arabic, meaning that each verb form locates its eventuality time with respect to a reference time established by its local syntactic context, rather than to the speech time directly. The auxiliary complex saufa yakuunu (will be) shifts the reference time into the future. The perfective verb qaraʔa (read) is interpreted as past with respect to this future reference time. Consequently, perfective morphology systematically signifies past with respect to a reference time. This reference time is the speech time only in the default case.

The relative tense view is supported by the fact that a perfective verb cannot express that the event it describes is simultaneous with a reference time, even when the reference time is itself in the past. The hypothetical simultaneous reading in such

contexts is called the ‘present under past’, or ‘sequence of tense’ reading in languages where it is available such as English (Prior 1967, Ladusaw 1977, Dowty 1982, Enç 1987, Ogihara 1995, and many others). In Arabic, simultaneity is expressed by the imperfective. For example, the imperfective verb yaktubu in (10a) (the imperfective indicative form of write) describes an event that is in progress at the past time established by the perfective matrix verb qaala (say) (Fassi Fehri 2004). It asserts that he said to me at a past time: “I am writing the letter”. The English translation expresses this temporal relation with the past tense progressive verb was writing. The past tense morphology in was writing is not interpreted in English. It is a morphological reflex of the temporal subordination of be writing to the matrix past tense verb said. As expected in light of (10a), the perfective subordinate verb in (10b) has only a reading in which it describes a time that is in the past with respect to the reference time established by the past tense matrix verb, which is itself in the past with respect to the utterance time. It can only mean that he said to me at a past time: “I wrote the letter”. Thus, the past/present distinction expressed by the perfective/imperfective morphological distinction is relative to a reference time in Arabic, which is the utterance time by default but may be shifted forward or backward with respect to the utterance time by various temporal operators.

(10)

a.

qaal-a l-ii ʔinna-hu ya-ktub-u r-risaalat-a saidPERF-3MS to-me that-he 3MS-writeIMPF-IND the-letter-ACC He said to me that he was writing the letter.

b.

qaal-a l-ii ʔinna-hu katab-a r-risaalat-a saidPERF-3MS to-me that-he 3MS-wrotePERF-3MS the-letter-ACC

He said to me that he wrote the letter

It should be noted that the possibility of a “present under past” reading of the past tense in English is limited to stative verbs. Consequently, English actually patterns like Arabic in (10b), whose English translation also asserts that the letter was finished at the time he told me that he wrote the letter. Crucially, the “present under past” interpretation of the past tense in English is not available in Arabic for stative predicates, which include the progressive interpretation of the imperfective seen in (10a), as well as basic stative predicates like mariiḍ (sick) in (11) (Fassi Fehri 2004; see Vlach 1981 on the stativity of the progressive). Without the past tense copula, (11) asserts that he said to me at a past time that he was sick at that time. With the past tense copula kaana (was), (11) can only assert that he said to me at a past time that he had been sick prior to that time.

(11)

qaal-a l-ii ʔinna-hu mariiḍ-u-n / kaan-a mariiḍ-a-n. saidPERF-3MS to-me that-he sick-NOM-N / wasPERF-3MS sick-ACC-N He said to me that he was sick / had been sick.

The tense analysis of the perfective/imperfective contrast maintains that tense in Arabic functions largely as in other languages, except for the absence of sequence of tense effects. One point in which certain modern dialects of Arabic are clearly unlike European languages, and in which Arabic stands to make a empirical contribution to semantic theory, concerns the interpretation of active participles. The pattern in question is well documented in Syrian Arabic (Cowell 1964, Boneh 2010), Palestinian (Wild 1964), Egyptian (Woidich 1975, Eisele 1990), Libyan (Mitchell 1952), Kuwaiti

(Al-Najjar 1984, Brustad 2000), and Najdi (Ingham 1994), and probably occurs in other dialects. In these dialects, the active participles show a durative or perfect reading depending on the lexical aspectual type of the underlying verb. Vendler (1957) identifies four basic aspectual types in English: states are compatible with forphrase duration adverbials (e.g. for an hour) but not compatible with the progressive (12); activities are compatible with for-phrases and the progressive (13); accomplishments are compatible with in-phrase duration adverbials (e.g. in an hour) and the progressive (14); achievements are compatible with in-phrases but not the progressive (15).

(12)

(13)

(14)

(15)

a.

Mary saw the star for an hour.

b.

*Mary was seeing the star.

a.

Mary drew for an hour.

b.

Mary was drawing.

a.

Mary drew a circle in an hour.

b.

Mary was drawing a circle.

a.

Mary reached the top in an hour.

b.

*Mary was reaching the top.

[State]

[Activity]

[Accomplishment]

[Achievement]

The predicates that are compatible with in-phrases (the accomplishments and activities) have in common that they attribute a logical endpoint to the event they describe (the circle being complete in (14) and Mary being at the top in (15)). These

are referred to as telic predicates (having a goal, or “telos”). The predicates compatible with for-phrases are atelic, lacking a logical endpoint.

Brustad (2000) proposes that in the dialects in question, if the underlying verb is atelic, the corresponding active participle describes an ongoing eventuality of the type described by the underlying verb (the ‘durative’ reading), illustrated in (16) for Syrian Arabic. If the underlying verb is telic, the corresponding active participle describes a post-state of the underlying event (the ‘perfect’ reading, so called because it resembles the English perfect in interpretation), illustrated in (17). These examples are from Cowen (1964).

(16)

a.

ləәssaa-ni məәtradded still-1S vacillating ‘I am still vacillating [undecided].’

b.

ʔana maaliyyan məәʕtəәmed ʕalee-h I financially depending on-him ‘I am financially dependent on him.’

(17)

a.

ʔana ʒaaye ʔaddem ṭalab I coming submit request ‘I have come to submit a request.’

b.

ṭ-ṭaʔṣ ħəәlu wa-ʃ-ʃams ṭaalʕa the-weather beautiful and-the-sun coming.out

[Syrian]

‘The weather is nice and the sun has come out.’

The resemblance between structures like those in (17) and the English perfect is reinforced by the fact that the state resulting from the event the underlying verb describes must still hold at the reference time, a semantic connotation the English perfect has (McCawley 1971, McCoard 1978). Example (17a) entails that the speaker is still present at the utterance time and (17b) that the sun is still out. As Cowell (1964) notes, while the perfect verb labas means to put on, said of clothing (18a), its active participle laabis means to wear (18b). That is, it asserts that whoever put on the clothes still has them on.

(18)

a.

labas tyaab-u.

[Syrian]

put.on clothes-his ‘He put on his clothes.’

b.

laabis tyaab-u. putting.on clothes-his ‘He has put on his clothes.’ (He is still wearing them)

Another similarity between the perfect reading of the Arabic active participle and the English perfect is that the present relevance implication disappears in the context of quantificational adverbs like marra (once). In example (19), the adverb hallaʔ (now) selects the standard present perfect reading with its present relevance connotation, while the adverb marra (once) selects the so-called ‘existential’ perfect reading, which asserts that a cake baking event took place at some point in the past. On this

reading (19) does not imply that the cake in question is still available or in any way relevant to the discourse context.

(19)

ana ʕaamil gaato hallaʔ/marra

[Syrian]

I making cake now/once ‘I have made a cake now/once.’

In contrast to Brustad’s assessment that the difference between (16) and (17) can be traced to telicity, Boneh (2010) develops an analysis of the basic pattern in (16) and (17) that posits a fundamental similarity between activities and accomplishments. She claims that the participle holds of a (post)-state invoked in the underlying verb denotation. If the underlying verb is an accomplishment, the participle describes the post-state of the transition the verb describes, whence the perfect reading in examples like (17). Activities, on her account, are like accomplishments, describing a complex event with a development portion and post-state portion. She supports this contention with the observation that some verbs that function as activity verbs in English, such as sleep, have counterparts in Arabic that describe a transition, and whose participial derivatives describe a post state. Hence, the participle naayim appears at first glance to be synonymous with English sleeping (20a). However, the underlying verb naam in (20b) does not have an activity reading, but only an accomplishment reading analgous to fall asleep, which the progressive construction in (20c) clarifies. The progressive in (20c) does not entail that Sami is asleep yet, unlike the English progressive counterpart of (20b) Sami is sleeping. Hence, naam means not sleep but fall asleep and the participle naayim means not be sleeping but have fallen asleep, the usual perfect reading of the participle.

(20)

a.

saami naayim.

[Syrian]

sami sleeping ‘Sami has fallen asleep.’

b.

saami naam. sami sleepPFTV ‘Sami fell asleep.’

c.

saami ʕam byinaam. sami PROG sleepIMPF ‘Sami is falling asleep.’

This analysis captures the fact that most verbs whose English counterparts are activity predicates have the perfect reading in the participial form in Arabic. If, as Boneh’s analysis requires, a verb like maʃʃaṭ (comb) in (21a) describes a transition of the state of Sami’s hair, then, as expected, the progressive form in (21) locates the listener within that transition (as in (20c)), and the participle in (21c) describes the post-state of that transition.

(21)

a.

saami maʃʃaṭ ʃaʕr-u. sami combed hair-his ‘Sami combed his hair.’

b.

saami ʕam byimaʃʃiṭ ʃaʕr-u.

[Syrian]

sami PROG combIMPF hair-his ‘Sami is combing his hair.’

c.

saami mmaʃʃiṭ ʃaʕr-u. sami combing hair-his ‘Sami has combed his hair.’

However, some activity verbs do not display the pattern in (21). Verbs such as maʃa (walk, go) have a durative interpertation in the participial form (in some dialects in addition to a perfect interpretation, as Brustad 2000 notes) that is more or less synonymous with the corresponding progressive form.

(22)

a.

saami maʃa bi-ʃ-ʃaṭṭ. sami walked on-the-beach ‘Sami walked on the beach.’

b.

saami ʕam yimʃi bi-ʃ-ʃaṭṭ. sami PROG walkIMPF on-the-beach ‘Sami is walking on the beach.’

c.

saami maaʃi bi-ʃ-ʃaṭṭ. sami walking on-the-beach ‘Sami is walking on the beach.’

[Syrian]

While it is not immediately obvious how Boneh’s analysis might extend to these verbs, Pallottino (2013) makes the observation that in Tunisian Arabic, the addition of an endpoint description to an otherwise atelic verb does not effect the interpretation of the corresponding active participle. Example (23a) below asserts without the parenthesized material that Ali is walking—the durative reading typical of participles of atelic verbs like maʃa (walk). The parenthesized prepositional phrase contributes an endpoint to the spatial path associated with the walking event, making the underlying event description telic. The participial phrase meʃi li-d-dar (walking to the house), however, remains durative in interpretation. It does not receive the perfect interpretation typical of telic predicates (cf. (17)). This observation carries over to Syrian, as example (23b) shows (cf. (22c)). This means that such predicates do not receive the perfect interpretation in the participial form even when they are augmented with material that makes them telic. On one hand, this observation reinforces Boneh’s point that the activities and accomplishments pattern the same; activity verbs generally have the perfect reading of accomplishments in the participial form and activity verbs that for some reason have a durative reading in the participial form also have a durative reading when they are made into accomplishments by the addition of a telos. On the other hand, it remains unclear what exactly is exceptional about verbs like maʃa (walk).

(23)

a.

ʕali meʃi (li-d-dar).

[Tunesian]

Ali walking (to-the-house) ‘Ali is walking (to the house).’

b.

saami maaʃi la-ʃ-ʃaṭṭ.

[Syrian]

sami walking (to-the-beach) ‘Sami is walking (to the beach).’

2.3

Derivational Verb Morphology

Another issue that has attracted attention in the semantics of Arabic and other Semitic languages concerns the meaningfulness of the verb templates, or ‘forms’. One clear generalization, articulated by Wright (1981, p. 31ff, vol. 1) and analyzed by Fassi Fehri (2003b) and (in connection with the cognate Hebrew verb forms) Doron (2003), concerns the two causative templates faʕʕala (form II) and ʔafʕala (form IV). Both (24b) and (24c) are causative derivatives of the intransitive verb in (24a). The external argument of the form II causative represents the immediate source of the action described by the verb, for which reason example (24b) implies that the captain acted with the intention to sink the ship. The form IV causative does not carry this connotation, and so is more readily compatible with the inanimate subject it has in (24c).

(24)

a.

ghariq-at s-safiinat-u. sankI-3FS the-ship-NOM ‘The ship sank.’

b.

gharraq-a l-qubṭaan-u s-safiinat-a. sankII-3MS the-captain-NOM the-ship-ACC ‘The captain sank the ship.’

c.

ʔaghraq-at l-ʕaaṣifat-u s-safiinat-a. sankIV-3FS the-storm-NOM the-ship-ACC ‘The storm sank the ship.’

Doron claims that the form II template characterizes the external argument of the underlying verb as what she calls an ‘actor’, while form IV is genuinely causative. Form II may, but need not, actually add an external argument. This accommodates the fact that form II is not always valency increasing vis à vis form I, unlike form IV. A purely intensivizing use of the second form of a transitive verb, that does not introduce an additional argument, is illustrated by verbs like ḍaraba (hit), which in the second form (ḍarraba) is still transitive but means ‘beat severely’. Yet, it is not clear that the intensivizing function of form II can be reduced to the thematic status of the subject. Fassi Fehri (2003b) describes the intensivizing function of Arabic form II as pluractionality, i.e., pluralization of the event argument. While (25a) means that the (implicit singular) subject injured the man once, (25b) asserts that he inflicted many wounds on him. This pluractionality may distribute over a plural object, so that (25c) asserts that he injured many soldiers, inflicting one wound on each (Fassi Fehri 2003b, p. 155).

(25)

a.

ʒaraḥ-a r-raʒul-a wounded-3MS the-man-ACC ‘He wounded the man.’

b.

ʒarraḥ-a r-raʒul-a

wounded-3MS the-man-ACC ‘He inflicted many wounds on the man.’

c.

ʒarraḥ-a l-ʒunuud-a wounded-3MS the-soldiers-NOM ‘He wounded the soldiers.’

It is therefore unclear whether Doron’s characterization of the meaning of form II exclusively in terms of the relation of the subject to the event is adequate for Arabic. Doron makes another claim, however, that holds some promise, in connection with apparent exceptions to the pattern in (24). She claims that exceptions occur only for ‘singleton’ roots—roots that occur in only one form. Here, the form does not do any ‘work’ distinguishing lexical items in causativity and actionality and therefore carries no significance. Testing this hypothesis for Arabic requires further research, but it seems promising that there do not appear to be any verbs in Arabic that show the opposite pattern as that seen in (24), that is verbs whose form II is compatible with an indirect causer but whose form IV requires a direct ‘actor’.

2.4

Pragmatics

Pragmatics is the study of the interaction of language use and the grammar of discourse structure, particularly how the communicative goals of discourse participants affect how they structure their discourse contributions. As in other areas of semantic inquiry on Arabic, little research has focused on pragmatic issues. However, Haddad (2013) presents a thorough investigation of the role of non-

argument dative clitic pronouns in Lebanese Arabic in communicating speaker attitudes toward the propositional content of their utterances, and the role of these attitude attributions in enforcing social mores. Non-argument dative pronouns occur as verb suffixes in Lebanese Arabic. The non-argument pronouns occur in addition to the thematic arguments of the verb. Haddad claims that the presence of a nonargument dative clitic profiles the referent of the pronoun (the speaker in (26)) as a holder of an attitude toward the proposition encoded by the corresponding sentence without the clitic. The attitude need not be negative, but presents a moral evaluation of the content of the proposition. The propositional content of (26) represents a situation that is not considered virtuous in a conservative society. One who utters (26) with the dative pronoun communicates his or her condemnation of Samir’s behavior.

(26)

samiir kill layle byishar-li bi-maṭraħ ʃikil.

[Lebanese]

samir every night stay.out3MS-meDAT in-place different Lit.: ‘Samir goes out to a different place every night on me.’ i.e., ‘Samir parties too much.’

The evaluation is not part of the truth conditional content of the sentence and the listener may not deny the attitude itself, though he or she may criticize the speaker for being judgmental, in terms like (27).

(27)

leeʃ miin-ak inta ta-yi-shar-lak ʔaw ma yi-shar-lak.

[Lebanese]

why who-you you to-3MS-stay.out-youDAT or not 3MS-stay.out-youDAT Lit.: ‘Who are you for him to go out on you or not go out on you?!’ i.e., ‘Who are you to judge him?!’

Omitting the dative clitic from (26) de-profiles the speaker as an attitude holder. Without the clitic, the sentence may still be taken to imply a moral judgment. As before, the listener cannot deny the speaker’s attitude toward the proposition. But without the dative clitic, the listener is also no longer in a position to criticize the speaker for being judgmental, since the speaker’s judgment is no longer profiled by the overt form of the utterance. That is, (27) is an infelicitous response to (26) when (26) lacks the dative clitic. The listener might still take issue with the speaker’s propensity to gossip by responding with the utterance in (28).

(28)

leeʃ miin-ak inta ta-ti-ħki ʕl-e la-z-zalame.

[Lebanese]

why who-you you to-2MS-talk about-him to-the-man ‘Who are you to gossip about the man?!’

Haddad presents a model of these interactions in the cognitive grammar framework, while a followup paper (Haddad 2014) describes syntactic conditions on the reference of the dative clitic.

3.

FUTURE DIRECTIONS

Arabic is underrepresented in the theoretical semantic literature, and therefore carries substantial potential to offer new empirical discoveries and novel theoretical contributions. Some aspects of Arabic grammar that represent potentially fruitful areas of semantic inquiry but that have gone largely unexplored are described below.

3.1

Tense and Modality

Comrie (1976) cites the example in (29), from Wright (1981, vol. 2, p. 9), in support of the notion that perfective signifies relative past tense. The perfective subordinate verb ħmarra (became ripe) describes a time that is in the past with respect to a reference time established by the future interpretation of the imperfective verb ʔaʒiiʔu (I come). Although the relative tense view of Arabic has independent support, it is unexpected that the English translation to (29) contains a present tense verb, rather than a past tense verb on analogy to the parallelism in the interpretation of (10b), where both English and Arabic past tense is interpreted as locally relative, because the verb in question is eventive.

(29)

ʔa-jiiʔu-ka ʔiðaa ħmarr-a l-busr-u 1S-come-you when redden-3MS the-dates-NOM ‘I will come to you when the dates become red.’

This observation suggests that the particle ʔiðaa (if, when) makes a semantic contribution of its own that interacts with the verb tense in Arabic. No compositional semantic account of this interaction has been articulated at the time of this writing. One thing such an account must accommodate is the fact, as Wright notes, that ʔiðaa may optionally be followed by a verb in the imperfective form, as in (30). He does not mention any difference in interpretation contingent on the morphological form of the verb. The particle qad below reinforces the perfect interpretation of the following perfective verb.

(30)

wa-ʔiðaa tu-tlaa ʕalay-him ʔaayaat-u-naa qaal-uu qad samiʕ-naa and-when 3FS-readIMPF.PASS to-them verses-NOM-our saidPERF-3P QAD heard-1P ‘And when our verses are read to them, they said we have heard.’

The occurrence of the perfect with ʔiðaa is presumably related to its occurrence with related particles such as the counterfactual conditional complementizer law (31a) (counterfactual if) and maa (31b) (as long as). Counterfactual law differs from conditional ʔiðaa in that it presupposes the falsity of the underlying proposition. (31b) presupposes that God did not wish to make mankind one nation. The examples below are from Wright (1981, vol. 2, pp. 6 and 17 respectively).

(31)

a.

law ʃaaʔ-a rabbu-u-ka la-jaʕal-a n-naas-a ʔummat-a-n waaħidat-a-n if wishedPERF-3MS lord-NOM-your LA-madePERF-3MS the-people-ACC

nation-ACC-N one-ACC-N ‘If your lord had wished, he would have made mankind one nation.’

b.

ʔan-naas-u maa daam-uu fii l-ħayaat-i d-dunyawiyyat-i ghaafil-uuna the-people-NOM as.long.as remainedPERF-3MP in the-life-GEN the-

temporal-GEN careless-3MP ‘People are careless as long as they remain in the life of this world.’

Wright mentions that as with ʔiðaa, the verb following law may occur in the imperfect, but here notes a difference in meaning. In this case, law has the meaning of the non-counterfactual conditional often expressed by ʔiðaa, one that does not presuppose the falsity of the underlying proposition.

(32)

law na-ʃaaʔ-u ʔaṣab-naa-hum bi-ðunuub-i-him if 1P-wishIMPF-IND injuredPERF-1P-them for-sins-GEN-their ‘If we wished, we could injure them for their sins.’

The interactions between these modal and temporal particles and tense deserve careful investigation in connection with a thorough survey of native speaker judgments of entailment and contradiction in such cases independently of what the historical written record appears to show, which does not provide us with robust evidence of interpretational subtleties. This investigation promises to be fruitful both for the development of a rigorous theory of tense interpretation in Arabic and for the understanding of the semantic similarities between the particles in question, by virtue of which they all allow or require the perfect.

3.2

Scalar Operators

The suggestion that the distinct particles discussed above share a component of meaning is similar to the case of particles that seem to show a semantic uniformity in superficially distinct usages. One example of such a particle is ħattaa, meaning either until or even. In its until use, it occurs either as a subordinating complementizer or as a preposition governing genitive, both illustrated in (33a). Its use meaning even is illustrated in (33b-c).

(33)

a.

saar-uu ħattaa ṭalaʕ-at ʃ-ʃams-u / ħattaa ṣ-ṣabaaħ-i journeyed-3MP until rose-3FS the-sun-NOM / until the-morning-GEN

‘They journeyed until the sun rose / until morning.’

b.

hattaa kariim najaħ-a fii mtiħaan-i l-riyaaḍiyyaat-i even karim succeeded-3MS in test-GEN the-math-GEN ‘Even Karim passed the math test.’

c.

najaħ-a kariim hattaa fii mtiħaan-i l-riyaaḍiyyaat-i succeeded-3MS karim even in test-GEN the-math-GEN ‘Karim passed even the math test.’

These two usages are not identical and have a different distribution. In its use as until, the following proposition or nominal describes what is the case at the endpoint of a scale associated with the verb, the path of the journey in (33a). In its use as even, it describes what is the case at the endpoint of a scale of outcomes ordered by surprisal. (33b) implies that Karim was that last student that anyone expected to pass the math test. This outcome was maximally surprising. Both usages therefore make reference to a scale and combine with a constituent that specifies the endpoint of that scale. A formal semantic analysis is called for that demonstrates what these two clearly related usages of ħattaa have in common and in what respect they differ semantically, and how this is connected to the difference in distribution between them. Such an analysis should also account for the connection evident in (33b-c) between the placement of ħattaa and the scale of comparison. In combination with a preverbal noun phrase in (33b), ħattaa compares the surprisal associated with the noun phrase referent to the surprisal of other values for the subject variable x in the phrase x passed the math test. In (33c), it compares the surprisal associated with the math test with the surprisal

associated with other values of x in the phrase Karim passed x. An analysis of this phenomenon should not only relate the meaning of ħattaa here to its use meaning until, but also account for restrictions on its distribution that distinguish Arabic from English, where the meaning of even is better studied (see Rooth 1985, among others). In particular, although ħattaa may combine with a topic noun phrase, as seen in (33b), it may not combine with a noun phrase inside a prepositional phrase (34a), nor with a post-verbal subject (34b). Nor may it occur by itself preceding a verb-initial clause (34c).

(34)

a.

*najaħ-a kariim fii hattaa mtiħaan-i l-riyaaḍiyyaat-i succeeded-3MS karim in even test-GEN the-math-GEN

b.

*najaħ-a hattaa kariim fii mtiħaan-i l-riyaaḍiyyaat-i succeeded-3MS even karim-NOM-N in test-GEN the-math-GEN

c.

*hattaa najaħ-a kariim fii mtiħaan-i l-riyaaḍiyyaat-i even succeeded-3MS karim-NOM-N in test-GEN the-math-GEN

3.3

Degree Constructions

Another type of scalar construction which is likely to be informative for future research is the syntax and semantics of superlative and comparative constructions in Arabic. The superlative construction in Standard Arabic has a surface format that supports standard analyses of the logical form of the superlative construction in other languages. Superlative adjectives are derived by the prosodic template ʔaC1C2aC3,

where C1-C3 are the root consonants of the adjectival base. I gloss this template as ‘elative’ since it occurs in comparative constructions as well, more on which below. The Arabic sentence in (35) is analogous in form and meaning to its English counterpart, modulo independent differences in the structure of noun phrases. The Arabic superlative adjective follows the modified noun and agrees with it in definiteness, as well as case, though case agreement is coincidentally obfuscated by regular phonology in example (35), which deletes the stem-final glide in the base ðakiyy (smart).

(35)

yasmin hiya ṭ-ṭaalibat-u l-ʔaðkaa fii ṣ-ṣaff-i yasmin she the-student-NOM the-smartELV in the-class-GEN ‘Yasmin is the smartest student in the class.’

Szabolcsi (1986) and Heim (1999) note that the specification of the scale of comparison in superlative constructions extends beyond the host adjective to the NP containing the adjectival host as a whole. (35) does not assert that Yasmin has the property of being both a student and the smartest person (the smartest person might not be a student)—the expected intersective reading of the juxtaposition of noun and adjective in (35)—but rather merely the smartest among the students in the class. They derive this effect by postulating movement of the superlative morpheme, without the corresponding adjective, to the left edge of the NP containing it in the surface structure, deriving a structure that looks schematically like (36). Movement leaves a degree-denoting trace in the base position, deriving a description be a dsmart student in the class in the case at hand, where the predicate d-smart denotes the

set of individuals who are smart to at least degree d. The superlative morpheme then names the individual who exemplifies this predicate to the greatest value for d.

(36)

Yasmin is mostd [a d-smart student in the class]

One curious facet of this analysis is that the derived description is semantically indefinite; we compare individuals with respect to the description be a d-smart student in the class. If we compared individuals with respect to the description be the d-smart student in the class, then for any given value for d we would have the presupposition, associated with the, that there is only one student who meets that description. This uniqueness presupposition would then clash with the fact that, as mentioned above, the predicate d-smart is true of any individual who is at least dsmart. For any student who is smart to a given degree d, all the smarter students are also smart to degree d, so no alternative student can fulfill the uniqueness presupposition introduced by definiteness. Heim and Szabolcsi suggest for this reason that the definite article in superlative constructions is a surface epiphenomenon, and that superlative constructions are fundamentally indefinite. On the basis of (35), where the definite article also appears in the surface structure, it appears this logic applies to Arabic as well.

Movement of the superlative morpheme is not overt in English. However, in Arabic the format in (35) is in free variation with two other formats that show displacement of the superlative morpheme. First, the entire adjective containing the superlative morpheme may be fronted to the left edge of the NP containing it in what is arguably the base structure in (35), deriving (37). This construction has some of the properties

attributed to the logical form of the corresponding English construction. The superlative morpheme is fronted out of its canonical position to the position where it is postulated to occur in the logical form according to the standard analysis of superlatives in English. Further, the construction is morphologically indefinite. There is no definite article and the head noun ṭaalibatin (student) bears the ‘tanwiin’ suffix (-n) typical of indefinites. On the other hand, (37) does not represent exactly the structure posited as the logical form of superlative constructions, since the entire superlative adjective is preposed, not just the superlative morpheme by itself.

(37)

yasmin hiya ʔaðkaa ṭaalibat-i-n fii ṣ-ṣaff-i yasmin she smartELV student-GEN-N in the-class-GEN ‘Yasmin is the smartest student in the class.’

However, even this last mentioned option is available in Arabic. (37) can be reformulated as (38), in which the term ʔakθaru (most) precedes the head noun ṭaalibatin and the adjective occurs in its nominal form (ðakaaʔ = smartness) as an accusative adjunct of the head NP. The term ʔakθaru is itself the superlative derivative of the adjective kaθiir, meaning much, or many, which arguably occurs here as a default morphological base for the superlative morpheme, which is itself a prosodic template.

(38)

yasmin hiya ʔakθar-u taalibat-i-n ðakaaʔ-a-n fii ṣ-ṣaff-i yasmin she muchELV-NOM student-GEN-N smartness-ACC-N in the-class-GEN ‘Yasmin is the smartest student in the class.’

This example has roughly the surface form that Heim and Szabolcsi attribute to the logical form of the corresponding English construction. The Arabic counterparts lend credence to the superlative displacement approach, and suggest as well that the distribution of the definite article is contingent on the surface position of the superlative morpheme. Its occurrence in (35) may represent a surface morphological marking of the covert movement required to derive the proper interpretation of the construction.

The comparative in Arabic is expressed by the same morpheme as the superlative, with the preposition min (from) introducing the standard of comparison. In light of the facts about the superlative described above, it is significant that the comparative never occurs with a nominal head. It may occur as a predicate adjective as in (39a), and in this function may be separated from the associated scalar term as the superlative may, illustrated in (39b).

(39)

a.

yasmin hiya ʔaðkaa min muʕiin yasmin she smartELV than mu’en ‘Yasmin is smarter than Mu’en.’

b.

yasmin hiya ʔakθar-u min muʕiin ðakaaʔ-a-n yasmin she muchELV-NOM than mu’en smartness-ACC-N ‘Yasmin is smarter than Mu’en.’

However, a comparative adjective may not occur in construct with a nominal head, as the superlative may, illustrated in (40a). It may only occur as a modifier of a noun, illustrated in (40b), as a positive adjective would.

(40)

a.

*yasmin hiya ʔaðkaa ṭaalibat-i-n min muʕiin yasmin she smartELV student-GEN-N than mu’en

b.

yasmin hiya ṭaalibat-u-n ʔaðkaa min muʕiin yasmin she student-GEN-N smartELV than mu’en ‘Yasmin is a smarter student than Mu’en.’

That is, the comparative morpheme has the distribution of the associated adjective, while the superlative morpheme shows this distribution, illustrated in (35), in addition to a distribution corresponding to the logical form of the superlative construction posited by Heim and Szabolcsi. These observations suggest that the superlative construction is fundamentally nominal, the case of (35) being a base for covert displacement of the superlative to an adnominal position, while the comparative construction is fundamentally adjectival. These observations demonstrate substantial cross-linguistic uniformity in degree constructions and warrant further investigation.

The exploration of logical and syntactic relationships between the constructions described above and their interaction with their semantic and syntactic context promises to shed light on parallels in structure between different sortal domains in semantics, i.e. individuals, degrees, and times, and the significance of these parallels for a general combinatorial theory of meaning in natural language. These are a few

issues in an extensive range of similar issues relating to conditionals, modality in general, tense, quantification, degree constructions, and other areas of semantic inquiry that are under-described in Arabic and that represent fertile ground for future semantic research.

4.

REFERENCES

Aoun, Joseph and Elabbas Benmamoun (1998). Minimality, Reconstruction and PF Movement. Linguistic Inquiry 29:569-597. Al-Najjar, Balkees (1984). The Syntax and Semantics of Verbal Aspect in Kuwaiti Arabic. Ph.D. Dissertation, University of Utah. Boneh, Nora (2010). Perfect Constructions in Syrian Arabic. In Cabredo-Hofherr, Patricia and Brenda Laca (eds.) Layers of Aspect, CSLI Publications, Stanford, pp. 23-42. Brustad, Kristin (2000). The Syntax of Spoken Arabic: A Comparative Study of Moroccan, Egyptian, Syrian and Kuwaiti Dialects. Georgetown University Press, Washington, D.C. Cantineau, Jean (1953). Manuel élémentaire d’Arabe Oriental (Parler de Damas), C. Klincksieck, Paris. Chomsky, Noam (1981). Lectures on Government and Binding. Foris Publications, Dordrecht. Cohen, Marcel (1924). Le système verbal sémitique et l’expression du temps. Paris: Éditions Leroux. Cowell, Mark (1965). A Reference Grammar of Syrian Arabic. Georgetown University Press, Washington, D.C.

Doron, Edit (2003). Agency and Voice. Natural Language Semantics 11:1-67. Dowty, David (1982). Tenses, Time Adverbs and Compositional Semantic Theory. Linguistics and Philosophy 5:23-55. Eisele, John (1990). Time Reference, Tense and Formal Aspect in Cairene Arabic. In Eid, Mushira (ed.) Perspectives on Arabic Linguistics I, John Benjamins, Amsterdam, pp. 173-212. Enç, Mürvet (1987). Anchoring Conditions for Tense. Linguistic Inquiry 18:633-657. Fassi Fehri, Abdelkader (2003a). Arabic Perfect and Temporal Adverbs. In Alexiadou, Artemis, Monika Rathert and Arnim von Stechow (eds.) Perfect Explorations, Mouton de Gruyter, Berlin, pp. 69-99. Fassi Fehri, Abdelkader (2003b). Transitivity, Causativity and Verbal Plurality. In Lecarme, Jacqueline (ed.) Research in Afroasiatic Grammar II. John Benjamins Publishing Company, Amsterdam, pp. 151-185. Fassi Fehri, Abdelkader (2004). Tense/Aspect Interaction and Variation. In Guéron, Jacqueline and Jacqueline Lecarme (eds.) The Syntax of Time. MIT Press, Cambridge, Mass, pp. 235-257. Guilliot, Nicolas and Nouman Malkawi (2006). When Resumption Determines Reconstruction. In Baumer, Donald, David Montero and Michael Scanlon (eds.) Proceedings of the 25th West Coast Conference on Formal Linguistics. Cascadilla Press, Somerville, MA, pp. 168-176. Haddad, Youssef (2013). Pronouns and Intersubjectivity in Lebanese Arabic Gossip. Journal of Pragmatics 49:57-77. Haddad, Youssef (2014). Attitude Datives in Lebanese Arabic and the Interplay of Syntax and Pragmatics. Lingua 145:65-103. Heim, Irene (1999). Notes on Superlatives. Ms., University of Texas.

Ingham, Bruce (1994). Najdi Arabic: Central Arabian. John Benjamins, Amsterdam. Ladusaw, William (1977). Some Problems with Tense in PTQ. In Schmerling, Susan and Carlota Smith (eds.) Texas Linguistic Forum 6. Austin: University of Texas. Larson, Richard (1990). Double Objects Revisited: Reply to Jackendoff. Linguistic Inquiry 32:589-632. May, Robert (1985). Logical Form: Its Structure and Derivation. MIT Press, Cambridge, Mass. McCawley, James (1971). Tense and Time Reference in English. In Fillmore, Charles and Terence Langendoen (eds.) Studies in Linguistic Semantics. Holt, Rinehart and Winston, New York, pp. 96-113. McCoard, Robert (1978). The English Perfect: Tense-Choice and Pragmatic Inferences. North-Holland Publishing Company, Amsterdam. Mitchell, T.F. (1952). The Active Participle in an Arabic Dialect of Cyrenaica. Bulletin of the School of Oriental and African Studies 14:11-33. Mohammad, Mohammad (1984). Stylistic Rules in Classical Arabic and Levels of Grammar. Studies in African Linguistics, Supplement 9:228-232. Montague, Richard (1970). Universal Grammar. Theoria 36:373-398. Montague, Richard (1973). The Proper Treatment of Quantification in Ordinary English. In Hintikka, Jaakko, Julius Moravcsik and Patrick Suppes (eds.) Approaches to Natural Language. Reidel, Dordrecht, pp. 221-242. Mughazy, Mustafa (2005). Rethinking Lexical Aspect in Egyptian Arabic. In Alhawary, Mohammad and Elabbas Benmamoun (eds.) Perspectives on Arabic Linguistics XVII-XVIII. John Benjamins Publishing Company, Amsterdam, pp. 133-172.

Ogihara, Toshiyuki (1995). The Semantics of Tense in Embedded Clauses. Linguistic Inquiry 26:663-680. Pallottino, Margherita (2013). The Interpretation of Participial Sentences in Tunisian Arabic. Talk delivered at the International Conference on Linguistics, 19-21 November, Petra, Jordan. Reinhart, Tanya (1976). The Syntactic Domain of Anaphora. Ph.D. Dissertation, MIT. Prior, Arthur (1967). Past, Present and Future. Oxford: Clarendon Press. Ross, John Robert (1967). Constraints on Variables in Syntax. Ph.D. Dissertation, MIT. Rooth, Mats (1985). Association with Focus. Ph.D. Dissertation, University of Massachussetts, Amherst. Sibawayh, Abu Bishr ‘Amr ibn ‘Uthman ibn Qanbar al-Bisri (2009). Al-Kitaab. Dar Al-Kotob Al-Ilmiyah, Beirut. Edited by Emil Badi’ Ya’qub. Originally published ca. 800. Szabolcsi, Anna (1986). Comparative Superlatives. In Fukui, Naoki (ed.) MIT Working Papers in Linguistics, vol. 8. MIT Press, Cambridge, Mass, pp. 245265. Vendler, Zeno (1957). Verbs and Times. The Philosophical Review 66:143-160. Wild, Stefan (1964). Die resultative Funktion des aktiven Partizips in den SyrischPalästinischen Dialekt des Arabischen. Zeitschrift der Deutschen Morgenländischen Gesellschaft 114:239-254. Woidich, Manfred (1975). Zur Funktion des aktiven Partizips im KeirenischArabischen. Zeitschrift der Deutschen Morgenländischen Gesellschaft 125:273293.

Wright, William (1981). A Grammar of the Arabic Language. Librarie du Liban, Originally published 1858.

5.

FURTHER READING

Boneh, Nora and Ivy Sichel (2010). Deconstructing Possession. Natural Language and Lingusitic Theory 28:1-40. This article discusses the connection between the choice of preposition used to express possession in Palestinian Arabic and the animacy and definiteness of the subject and the alienability of the possessive relation. Elghamry, Khaled (2004). Definiteness and Number Ambiguity in the Superlative Construction in Arabic. Lingua 114:897-910. This article treats the partitive superlative construction in Standard Arabic and documents parallels to quantifier constructions. McNabb, Yaron and Christopher Kennedy (2011). Extraction and Deletion in Palestinian Arabic Comparatives. In Bloselow, Ellen and Hamid Ouali (eds.) Perspectives on Arabic Linguistics XXII-XXIII. John Benjamins Publishing Company, Amsterdam, pp. 149-166. This article treats syntactic constraints on ellipsis in comparative constructions in Palestinian Arabic.

Word Count: 8,673