Argument-Structure-Reducing Operations in Standard Arabic : An Exploratory Comparative Study

This paper attempts to examine some argument-structure-reducing operations in Standard Arabic (SA for short). It is proposed here that some affixes (viz. prefixes and infixes) can decrease the argument structure (or valence) of the subclass of change-of-state (COS for short) verbs in the language under study. More specifically, these affixes function as unaccusativizers or decausativizers in that they can derive unaccusative COS verbs from causative COS verbs by suppressing the external argument of the latter verbs and syntactically promoting the direct object to subject position. Crucially, the ability of these affixes to affect the argument structure and the morphosyntactic realization of arguments is not limited to SA, but it has been attested in some other languages, such as Italian, Russian, Chichewa, Spanish, French, Eastern Armenian, West Greenlandic, and Tzutujil, among others.


Introduction
This paper examines some argument-structure-reducing processes in Standard Arabic (SA for short), compared to other languages.(Note 1) It is thus proposed that some affixes (viz.prefixes and infixes) can alter the argument structure (or valence); i.e., the number of arguments a predicate takes, and argument realization of the subclass of causative COS verbs in the language under study.The main claim defended in this paper is that these affixes function as unaccusativizers or decausativizers in this language.Interestingly, the ability of these morphological entities to affect the argument structure of verbs is not limited to SA, but it has been attested in some other languages, such as Russian, Spanish, French, Italian, Eastern Armenian, West Greenlandic, Chichewa, and Tzutujil, among others.Published by IDEAS SPREAD subclass of unaccusative COS verbs in SA which is derived from causative COS verbs by adding some affixes and making a slight vocalic change.
To support his argument, Hallman claims that the verb ħazana 'make sad', for instance, is derived from ħazina 'be sad' by changing the second vowel /i/ into /a/.He adds that the verb hadama 'ruin' is, similarly, derived from the verb hadima 'fall to ruin,' and the verb ħarama 'prohibit' is derived from the verb ħaruma 'be prohibited'.Hallman also argues that the causative verb sammana 'fatten' is derived from the unaccusative verb samina 'be fat' by doubling the consonant /m/ and changing the vowel /i/ into /a/.For lack of space, I will not address Hallman's analysis in detail here.

Verbs Beginning with the Prefix n-
I would claim that in SA unaccusative COS verbs which begin with the prefix n-are derived from causative COS verbs by the addition of this prefix at the beginning of the latter verbs.This claim can be illustrated by examples (1), (2), and (3) below.
Unac-interrupted.3FS the-electricity-NOM 'Electricity was interrupted.' Obviously in (1b), the unaccusative COS verb ʔinfataħa 'was opened' is straightforwardly derived from the causative verb fataħa 'open' in (1a) by the addition of the prefix n-at the beginning of the verb and the removal of germination in the second.
Similarly, in (2b) the unaccusative COS verb ʔinkasara 'broke' is derived from the causative verb kasara 'break' in (2a) by the addition of the prefix n-at the beginning of the latter verb.
In (3b), the unaccusative COS verb ʔinqaTaʕa-ti 'was interrupted' is equally derived from the causative COS verb qaTaʕa 'interrupt' in (3a) by adding the prefix n-at the beginning of the latter verb.The suffix -ti attached to the verb ʔinqaTaʕa in (3b) is a portmanteau morpheme encoding both gender and number.
On the basis of what has been said so far with regard to the derivation of unaccusative COS verbs beginning with the prefix n-, it can be concluded that all verbs belonging to the subclass of unaccusative COS verbs beginning with the prefix n-, such as verbs listed in   Notice that the prefix ta-is added at the beginning of the causative COS verb haddama 'demolish' in (4a) to derive the unaccusative COS verb tahaddama 'was demolished' in (4b).
Likewise, in (5b) the unaccusative COS verb taGajjara 'was changed' is derived by adding the prefix ta-at the beginning of the causative COS verb Gajjara 'change' in (5a).
In (6b), the unaccusative COS verb tamaddada-ti 'was extended' is equally derived by adding the prefix ta-at the beginning of the causative COS verb maddada 'extend' in (6a).The suffix -ti at the end of the verb tamaddadati is a portmanteau morpheme which marks both gender and number.
In (7b), the unaccusative COS verb taʕaTTala 'broke down' is straightforwardly derived by adding the prefix taat the beginning of the causative COS verb ʕaTTala 'break down' in (7a).
Interestingly, and as has previously been noted with respect to the derivation of unaccusative COS verbs beginning with the prefix n-, it could be inferred from the examples given above that all verbs belonging to the subclass of unaccusative COS verbs beginning with the prefix ta-, such as verbs listed in table 2 below, are likely to be systematically derived by adding the prefix ta-at the beginning of the causative COS verbs from which they are derived.Published by IDEAS SPREAD

Verbs Containing the Infix -ta-
A close scrutiny of the examples ( 8), ( 9), (10), and (11) below reveals that there are two derivational patterns of unaccusative COS verbs which contain the infix -ta-in SA: (i) the pattern of deriving this subclass of unaccusative COS verbs from triliteral causative COS verbal stems, and (ii) the pattern of deriving these verbs from quadriliteral causative COS verbal stems.The first subset is derived by the insertion of the infix -ta-after the first consonant and deleting the vowel following the first consonant of the triliteral verb, and adding the consonant /Ɂ/ and the vowel /i/ at the beginning of the verbal stem.The second subset is derived by the insertion of the infix -ta-after the second consonant of quadriliteral causative COS verbal stems and the replacement of the vowel /a/ after the first consonant /Ɂ/ by the vowel /i/.So, it may be noted here that the derivation of the subsets of unaccusative COS verbs containing the infix -ta-is somewhat complex, and might cause some difficulty or confusion for learners of SA.
(8) a. Ɂaħraqa l-ħari:q-u l-maʕmal-a.As can clearly be noticed in (8b), the infix -ta-is inserted between the second and the third consonants of the quadriliteral COS verb Ɂaħraqa 'burn' in (8a) with the replacement of the vowel /a/ after the first consonant /Ɂ/ by the vowel /i/, yielding the unaccusative COS verb Ɂiħtaraqa 'burned.' The unaccusative COS verb Ɂištaʕalati 'was lit' in (9b) is equally derived by the insertion of the infix -ta-between the second and the third consonants of the quadriliteral COS verb Ɂašʕala 'light' in (9a) with the replacement of the vowel /a/ after the first consonant /Ɂ/ by the vowel /i/.
Notice that in (10b), the unaccusative COS verb Ɂixtanaqa 'was suffocated' is similarly derived by the insertion of the infix -ta-after the first consonant of the verbal stem and deleting the vowel following the latter consonant of the triliteral verb xanaqa 'suffocate' in (10a), and adding the consonant /Ɂ/ and the vowel /i/ at the beginning of the verbal stem.
Just as in (10b), the unaccusative COS verb Ɂiltaħamati 'were welded' in (11b) is derived by the insertion of the infix -ta-between the first consonant of the triliteral verb laħama 'weld' in (11a), and adding the consonant /Ɂ/ Published by IDEAS SPREAD and the vowel /i/ at the beginning of the verbal stem.Tables 3 and 4 below show some other unaccusative COS verbs containing the infix -ta-in SA.To become rich Ɂaɣna: To make rich Ɂiftaqara To become poor Ɂafqara To make poor It follows from the above discussion that I have identified three derivational patterns yielding three categories of unaccusative COS verbs in SA: (i) the subclass of verbs formed by the addition of the prefix n-at the beginning of the causative COS verbal stem, (ii) the category of verbs formed by the addition of the prefix ta-at the beginning of the causative COS verbal stem, and (iii) the subset of verbs formed by the insertion of the infix -ta-in the causative COS verbal stem accompanied by a vocalic change.

Some Counterevidence
As mentioned in the introduction of this paper, an important point that should be clear at this juncture is that not all SA unaccusative COS verbs are derived from causative COS verbs.In other words, there is a subset of unaccusative COS verbs which does not alternate with causative COS verbs.Indeed, it has been noticed that some verbs which mysteriously resemble derived unaccusative COS verbs in their forms, but which are lexicalized as such.These verbs can be exemplified by the following: ɁinqaraDa 'become extinct,' Ɂindaɵara 'perish,' Ɂindalaʕa 'break out,' taɁa:kala 'lose some parts' or 'corrode,' and taDa:Ɂala 'become smaller.' I would claim here that unaccusative COS verbs which are derived from causative COS verbs in SA may be characterized as core unaccusative COS verbs; whereas unaccusative COS verbs which are not derived from causative COS verbs may be described as peripheral or lexicalized unaccusative COS verbs.
It should be capitalized here that these allegedly underived or lexicalized unaccusative COS verbs need more research and deeper analysis to unravel and account for their characteristic properties and behavior in their crosslinguistic dimension.Published by IDEAS SPREAD Having provided a brief overview of the derivation of some unaccusative COS verbs in SA, I now turn to address the interaction of some affixes with the argument structure and argument realization of some COS verbs in the language under study.

Affixes and COS Verbs' Argument Structure in SA
In this section, I will try to address the issue raised in the introduction of whether affixes can affect the argument structure and argument realization of COS verbs in SA.
A crucial point worth making at this juncture is that SA is a Semitic language which has rich and complex morphology.In other words, it has a variety of affixes, be they inflectional or derivational, which have crucial semantic and syntactic repercussions, among others.For present purposes, I focus the discussion here on the morphosyntactic effects of the derivational prefixes n-and ta-and the derivational infix -ta-.
Concretely, I hereafter address the issue of how the previous affixes interact with and affect the argument structure and argument realization of the subclass of causative COS verbs in SA.

Causative COS Verbs' Argument Structure and Argument Realization in SA
As is well-known in the relevant literature, causative COS verbs tend to be associated with two arguments crosslinguistically: An internal argument bearing the Patient/Theme theta role and an external one bearing the Agent theta role.In English, for instance, we find verbs such as: break, melt, crack, freeze, evaporate, etc.In French, we find verbs like casser 'break,' fondre 'melt,' liquéfier 'liquefy,' congeler 'freeze,' exploser 'explode,' démolir 'demolish,' fragmenter 'split up,' etc.
Interestingly, SA constitutes no exception as far as causative COS verbs are concerned.For expository clarity and convenience, a characteristic property of these verbs is that they tend to select two arguments, viz.At first glance, we clearly notice that in (12) the causative COS verb Ɂaħraqa 'burn' selects two arguments, namely an internal argument l-maʕmal-a 'the factory' and an external one l-ħari:q-u 'the fire'.The internal argument thus bears the Patient theta role (i.e., it is undoubtedly the undergoer of the action of burning denoted by the verbal predicate) and the external one bears the Agent theta role (i.e., it is the actor or doer of the action of burning that brings about the change in the state of the internal argument, viz. the burned factory).
In ( 13), the causative COS verb ɣajjara 'change' equally selects two arguments, namely an internal argument lqa:nu:n-a 'the law' and an external one l-ħuku:mat-u 'the government.'The internal argument obviously bears the Theme theta role (i.e., it is the entity undergoing the action of changing denoted by the verbal predicate) and the external one bears the Agent theta role (i.e., it is the actor or doer of the action of changing that brings about the change in the state of the internal argument, namely the changed law).
The same pattern is noticed in ( 14) where the causative COS verb xanaqa 'suffocate' is associated with two arguments, namely an internal argument r-ražul-a 'the man' and an external one l-ɣa:z-u 'the gas.'The internal argument obviously bears the Patient theta role (i.e., it is the undergoer of the action of suffocating denoted by the verbal predicate) and the external one bears the Agent theta role (i.e., it is the actor or doer of the action of suffocating that causes the change in the state of the internal argument, namely the suffocated man).Published by IDEAS SPREAD Viewed from a syntactic perspective, the internal arguments in ( 12), (13), and ( 14), viz.l-maʕmal-a, l-qa:nu:n-a, and r-ražul-a are projected as direct objects in the respective constructions in which they appear.In contrast, the external arguments in the same constructions, namely l-ħari:q-u, l-ħuku:mat-u, and l-ɣa:z-u are realized as subjects.
Having examined the argument structure and argument realization of some causative COS verbs in SA, I now turn to examine the argument structure and argument realization of some unaccusative COS verbs in the language being studied.

Unaccusative COS Verbs' Argument Structure and Argument Realization in SA
As is well established cross-linguistically, the subclass of unaccusative verbs uniformly selects one argument, specifically an internal argument and lacks an external argument.On the basis of evidence culled from SA, I assume that unaccusative COS verbs in SA constitute no exception.This subclass of verbs, more specifically those which are derived from causative COS verbs, tends to be associated with one and only one internal argument.
For the sake of expository clarity and concreteness, consider the following illustrative examples: In ( 16), one internal argument is selected by the unaccusative COS verb ʔinkasara 'broke', namely l-žisr-u 'the bridge.'This internal argument bears the Patient theta role (i.e., it is the undergoer of the action of breaking denoted by the verbal predicate).
Similarly, the unaccusative COS verb ʔinqaTaʕa 'was interrupted' in ( 17) selects Ɂal-kahraba:Ɂ-u 'the electricity' as its unique internal argument which bears the Patient theta role.
In examples ( 18) through ( 24), the unaccusative COS verbs ta-haddama, ta-ɣajjara, ta-maddada, ta-ʕaTTala, ʔiħtaraqa, ʔištaʕala, and ʔiltaħama uniformly take one and only one internal argument, respectively l- manzil-u, lqa:nu:n-u, l-minTaqat-u, l-qiTa:r-u, l-maʕmal-u, n-na:r-u, Ɂažza:Ɂ-u.Concretely, in the examples given above (viz.from ( 15) to ( 24)), the internal arguments l-maTʕam, l-žisr, Ɂalkahraba:Ɂ, l-manzil, l-qa:nu:n, l-minTaqat, l-qiTa:r, l-maʕmal, n-na:r, and  resulting verbal form moves up to the T node and ends up under the AgrS node, to fulfil Case and agreement requirements, as a full-fledged unaccusative COS verb Ɂiħtaraqa 'was burned.' As far as the DP l-maʕmal 'the factory' in the above tree diagram is concerned, it might be proposed that it occupies the direct object position within VP at D-Structure where it is assigned the Patient theta-role, and it moves cyclically to [Spec, InAspP], then to [Spec, TP] where it surfaces as the subject l-maʕmal-u of the sentence, as is required by the EPP, and is consequently assigned Nominative case.

Argument Structure or Valence-Decreasing Affixes in SA
As is well established in the literature, the way(s) in which the argument structure of a verbal predicate is syntactically realized may be affected by certain identified and attested operations or processes, among them the following have been examined and tested in a number of languages: (i) passivization, (ii) middle formation, (iii) reflexivization, (iv) the causative-inchoative alternation, (v) applicative, and (vi) antipassive.(Note 3) Other processes have also been identified in the literature which can alter the argument structure of verbal predicates, among these processes figure out some morphological operations such as word formation.(Note 4) For present purposes, I will discuss two processes subsumed under word formation, namely prefixation and infixation in SA, which I claim to be argument-reducing operations.

The prefix n-
As has been noted above, the prefix n-plays a crucial role in deriving a subset of unaccusative COS verbs in SA.What interests us more here is not the derivational role of this bound morpheme, but rather its ability to affect the argument structure or valence of causative COS verbs to which it is attached.I hence claim that this affix is morphosyntactically active.
Concretely, when this prefix is added at the beginning of a causative COS verb, it tends to systematically alter its argument structure by decreasing it by one.The external argument is consequently suppressed from the construction and the internal argument which bears the Patient theta role is promoted to subject.
One clear-cut result of this morphosyntactic process is that the causative verb is straightforwardly unaccusativized, yielding an unaccusative COS verb with one-place argument structure.
It should also be noted here that this affix is productive in SA, to the extent that it can derive a significant number of unaccusative COS verbs.I have listed some 25 verbs, as is illustrated in table 1.

The prefix ta-
Another prefix which equally plays a significant role in deriving unaccusative COS verbs in SA is the prefix ta-.
As has already been claimed above, when this bound morpheme is added at the beginning of some causative COS verbs it tends to change their argument structure or valence by reducing it by one.The external argument is consequently erased from the construction and the internal argument which bears the Patient theta role is promoted to subject.The straightforward and obvious result of this morphosyntactic process, then, is that the causative verb is unaccusativized or decausativized, yielding an unaccusative COS verb with one-place argument structure.
Like the prefix n-examined above, the prefix ta-is therefore very active in SA morphosyntactically speaking.Concretely, it allows the derivation of an important number of unaccusative verbs from causative COS verb stems.
Furthermore, this affix is highly productive in SA.This is corroborated by the evidence listed in table 2.Not less than 37 verbs have been listed.

The Process of Infixation
The infix -ta-As has already been put forward, the infix -ta-plays a not insignificant role in deriving some unaccusative COS verbs in SA.Concretely, when this infix is inserted or incorporated in some causative COS verbal stems, it changes their argument structure or valence by reducing it by one.The external argument is consequently cut off from the construction and the internal argument which bears the Patient theta role is promoted to subject.As a result of this morphosyntactic process, the causative COS verb is unaccusativized, yielding an unaccusative COS verb with oneplace argument structure.
From what has been said above, it clearly follows that the infix -ta-is also active in SA.However, its productivity is somewhat limited, as far as I know.I have listed some 15 verbs, as tables 3 and 4 show.Published by IDEAS SPREAD constructions to alter the argument structure of certain verbs by reducing it.This observation may have some implications for Universal Grammar and may be worth investigating.
This conclusion might lead one to raise the issue of the universality of valence-decreasing operations and mechanisms and its possible implications for linguistic typology, in correlation with what Babby (2009) insightfully puts forward: Languages typically have a closed class of productive, diathesis-altering, paradigm-creating affixes (-af), which have their own diathesis.Since these affixes both alter V's initial diathesis and head their own projections in the syntax (afP), diathesis theory provides a natural setting in which an important lexalist dictum can be formalized: in addition to parameter-setting, the morphosyntactic differences we observe among languages can in large part be attributed to the language-specific properties of their diathesis-bearing affixes.(p.13)

Conclusion
To sum up, it has been argued in this paper that unlike Hallman's (2006) account which claims that causative verbs are derived from unaccusative verbs in Arabic by the processes on ablaut and gemination, it has been argued that there still is a subclass of unaccusative COS verbs in SA which is systematically derived from causative COS verbs by the addition of some affixes, be they prefixes such as n-and ta-or infixes such as -ta-.When these affixes are attached to or inserted in causative COS verbal stems, unaccusative COS verbs are straightforwardly and systematically derived yielding three verbal paradigms or patterns depending on the type of the affix (namely, the prefixes n-and ta-, and the infix -ta-), the subject is consequently removed from the syntactic structure, and the direct object is promoted to subject position.
Importantly, I have shown that some verbal affixes (more specifically the prefixes n-and ta-and the infix -ta-) tend to mark the subclass of unaccusative COS verbs in SA.
As has equally been examined and illustrated earlier in this paper, unaccusative COS verbs tend to show the same argument realization patterns in SA.In other words, the Patient/Theme or undergoer of the action is systematically realized in the syntactic structure as a subject, which is generally assumed to occupy the object position at the level of D-Structure.
Of equal importance, it has been found out that the above mentioned affixes function as unaccusativizers or decausativizers in this language.These affixes are also able to alter the argument structure or valence of a subset of causative COS verbs by decreasing it by one.
Another important finding is that morphology cross-linguistically interacts with syntax, so much so that some morphemes (most notably affixes, be they prefixes, infixes or suffixes) can change the argument structure or valence of certain verbs by decreasing or reducing it.This is corroborated with evidence culled from different languages.This promising and interesting linguistic phenomenon deserves deeper and cross-linguistic analysis and investigation to account for the variation in the types of morphemes which have syntactic effects or consequences (most notably argument structure or valence-changing effects) and to reveal the possible implications for what is called 'Universal Grammar' and 'parameterization.'It seems fairly reasonable to assert that addressing this issue is beyond the scope of the present study.

Notes
Note 1.Before embarking on the study of these phenomena, it should be noted at this stage that SA (called in Arabic « Al FuSħa » is distinguished from other spoken regional dialects in the Arabic world (known as Al ʕa:mmija ).It is widely assumed that SA is nowadays the native language of no speakers anywhere in the world.
In the present study, SA is assumed to be the language of written Arabic media, e.g., newspapers, books, journals, etc. -all forms of the printed word.It is also the language of public formal speaking and news broadcasts on radio and television.As far as the clause structure of this language is concerned, the unmarked word order is VSO.
Morphologically, SA has a rich and diversified inflectional system (Number, Gender, Person, Case, Tense/Aspect are morphologically marked; i.e., there are specific affixes which encode these inflections).
The Furthermore, I will specifically examine unaccusative verbs which denote an externally caused change of state; i.e., a change of state coming about because of something external to the entity that undergoes the change of state.
For further details and an insightful discussion about the distinction between internally and externally caused change-of-state verbs with examples, see McKoon and Macfarland (2000:833-837) and Wittek (2002:5ff.).
Note 4. For more detail and insightful discussion with examples drawn from various languages, see Haspelmath and Sims (2010: 234-263).Payne (1997: 55) also points out that "[M]ost languages employ various derivational operations that adjust the argument structure of a verb." Note 5. Haspelmath and Sims (2010: 240) define an 'antipassive' as "the term for a morphological operation whose effect is to background the patient in much the same way as the agent is backgrounded in the passive." the-council-NOM the-municipal-NOM l-minTaqat-a S-Sina:ʕijjat-a the-zone-ACC the-industrial-ACC 'The municipal council extended the industrial zone.' down.3MSthe-train-NOM 'The train broke down.' .3MS the-factory-NOM 'The factory was burned.' (9) a. Ɂašʕala l-ɁaTfa:l-u n-na:r-a fi: l-ɣa:bat-i.lit.3MPL the-boys-NOM the-fire-ACC in the-forest-OBL 'The boys lit the fire in the forest.' b.Ɂištaʕalati n-na:r-u fi: l-ɣa:bat-i.lit.Unac.3FSthe-fire-NOM in the-forest-OBL 'The fire was lit in the forest.' (10) a. xanaqa l-ɣa:z-u r-ražul-a.suffocated.3MSthe-gas-NOM the-man-ACC 'The gas suffocated the man.' b.Ɂixtanaqa r-ražul-u.suffocated.Unac.3MSthe-man-NOM 'The man was suffocated.' (11) a. laħama S-Sa:niʕ-u Ɂažza:Ɂ-a s-sajja:rat-i.welded.3MSthe-manufacturer-NOM parts-ACC the-car-GEN 'The manufacturer welded the parts of the car.' b.Ɂiltaħamati Ɂažza:Ɂ-u s-sajja:rat-i.welded.Unac.3MPLparts-NOM the-car-GEN 'The parts of the car were welded.' an internal argument and an external one.The internal argument bears the Patient/Theme theta role (i.e., the entity undergoing the action denoted by the verbal predicate) and the external one bears the Agent theta role (i.e., the actor or doer of the action that brings about the change in the state of the internal argument).Consider the following examples which illustrate what has just been noted: suffocated the man.' welded.Unac.3MPLparts-NOM the-car-GEN 'The parts of the car were welded.'Notice that in (15), the unaccusative COS verb ʔinfataħa 'was opened' selects one argument, namely the internal argument l-ba:b-u 'the restaurant.'This internal argument obviously bears the Theme theta role (i.e., it is the undergoer of the action of opening denoted by the verbal predicate).Published by IDEAS SPREAD Ɂažza:Ɂ originate in the object position at D-structure (i.e., the position normally occupied by objects of transitive verbs) where they are assigned the relevant theta role.Essentially adoptingTravis's (2010) andNossalik's (2010) phrase structure model, and as far as the syntactic projection of unaccusative COS verbs' argument structure is concerned, it might be proposed here that the internal arguments in the examples above are likely to move from their D-Structure position within VP to the Spec(ifier) position of InAspP (Inner Aspect Phrase) directly above VP, and end up in [Spec, TP] (i.e., in the Specifier position of Tense Phrase), where they are promoted to subject position of the clause and are, consequently, likely to fulfil Agreement requirements and be assigned Nominative Case, as illustrated by the structures (25), (26), and (27) below.(25) It might be proposed that in (25) the causative COS verb stem fataħa 'open' is generated under the V node at D-Structure; it first adjoins to the InAsp node under which the unaccusativizer head n-is generated, then the resulting verbal complex moves on to the T node and ends up under the AgrS node, to fulfil Case and agreement requirements, as a full-fledged unaccusative COS verb Ɂi-n-fataħa 'was opened.'As far as the DP l-ba:b 'the restaurant' in the above tree diagram is concerned, it might be proposed that it is generated in the direct object position within VP at D-Structure where it is assigned the Theme theta-role.It moves cyclically to [Spec, InAspP], then to [Spec, TP] where it surfaces as l-ba:b-u assigned Nominative case and, consequently, becoming the subject of the sentence in satisfaction of the Extended Projection Principle (EPP for short), which roughly stipulates that sentences need subjects.(26)As we have seen in (25), it might equally be proposed that in (26) the causative COS verb stem haddama 'demolish' is generated under the V node at D-Structure, it first adjoins to the InAsp node under which the unaccusativizer head ta-is generated, then the resulting verbal form moves up to the T node and ends up under the AgrS node, to fulfil Case and agreement requirements, as a full-fledged unaccusative COS verb tahaddama 'was demolished.' Concerning the DP l-manzil 'the house' in the above tree diagram, it might be proposed that it occupies the direct object position within VP at D-Structure where it is assigned the Patient theta-role, and it moves cyclically to [Spec, InAspP], then to [Spec, TP] where it surfaces as l-manzil-u assigned Nominative Case, and becoming the subject of the sentence, in fulfilment of the EPP.(27) It might equally be claimed that in (27) the causative COS verb stem Ɂaħraqa 'burn' is generated under the V node at D-Structure, it first adjoins to the InAsp node under which the unaccusativizer head -ta-is generated, then the Published by IDEAS SPREAD following abbreviations are used throughout this paper: NOM=Nominative.ACC=Accusative.GEN.= Genitive.OBL.= Oblique.M= Masculine.F=Feminine.S=Singular.PL. = Plural.PREF= Prefix.INF=Infix.REFL.=Reflexive.Expl.=Expletive.ERG.= Ergative.PART.=Particle.PERF.=Perfective.Unac= Unaccusativizer.Moreover, the following symbols are used to refer to IPA symbols: Ɂ= Glottal stop.ʕ= Voiced pharyngeal fricative.Ɣ= Voiced velar fricative.z= Voiced alveolar fricative.x= Voiceless velar fricative.ž= Voiced postalveolar fricative.š= Voiced postalveolar fricative.S= Voiceless retroflex fricative.D= Voiced retroflex fricative.T= Retrofex plosive.In addition, morphemes are used between two slashes.Note 2. By 'unaccusative change-of-state verbs,' I mean, in this study, verbs which denote the result state of the act of changing something into something different.On the other hand, change-of-state verbs are standardly defined as those verbs which denote the action of the bringing about of a result state.
table 1 below, are systematically derived by adding the prefix n-at the beginning of the causative COS verbal stems from which they are derived.Published by IDEAS SPREAD

Table 1 .
List of unaccusative COS verbs beginning with the prefix n-in SA 3.2 Verbs Beginning with the Prefix ta-I also claim that unaccusative COS verbs which begin with the prefix ta-in SA are derived by the addition of this prefix at the beginning of causative COS verbs.This claim is supported by the following pieces of data: Published by IDEAS SPREAD

Table 2 .
A list of unaccusative COS verbs beginning with the prefix ta-in SA

Arabic derived unaccusative COS verb Approximate meaning in English Arabic causative COS verb Approximate meaning in English tamazzaqa
wassaʕaTo widen, to extend, to expand Published by IDEAS SPREAD

Table 3 .
A list of unaccusative COS verbs containing the infix -ta-derived from triliteral verbal stems in SA

Table 4 .
A list of unaccusative COS verbs containing the infix -ta-derived from quadriliteral verbal stems in SA