Language Forum 29 (No.2), July-Dec. 2003, pp. 79-98.
Raising to Subject in Standard
Arabic and Modern Standard English*
Ahmed-Sokarno Abdel-Hafiz, South Valley University
This paper argues against Mohammed (1990)'s claim that Arabic does not have raising operations. Having established raising as a reality in Arabic syntax, I set out to discuss raising to subject in English. I have argued against Wardhaugh (1995)'s claim that verbs like want in English are raising verbs. Such verbs are control rather than raising verbs. This study then deals with the similarities and differences between Arabic (standard variety) and English (Modern Standard) as far as raising constructions are concerned.
Raising to subject is a type of argument movement (A-movement) which involves raising up a complement-clause subject or object into the position of matrix-clause subject. There are many languages (e.g. Greek (Soams and Perlmutter 1979); Ilokano (Gerdts 1980); Niuean (Seiter 1983); Romance languages (e.g. French, and Portuguese (Perlmutter and Postal 1983)) that have such constructions. Both Modern Standard English (henceforth English) and Standard Arabic (SA) have such a movement operation in which the subject of a complement clause can end up as subject of the matrix clause. In SA even a direct object or an indirect object can be raised up to become the matrix- clause subject (cf. Salih 1986:325). Note, however, that neither a direct object nor an indirect object can be directly raised up in English.
Although syntactic comparative and contrastive studies between English and Arabic are enormous (cf. Anani 1983; Bakir 1996; Alkhuli 1997; Bakir 1999; Homeidi 2000; Atawneh 2001). none of them has compared raising constructions in the two languages. Mukattash (2001) has compiled a preliminary bibliography of Arabic-English contrastive studies. This bibliography does not contain any work done on raising to subject in the two languages. Nor any reference is made to the topic of this study in the bibliography of Arabic Linguistics 1979-1994, which can be obtained from the internet. This paper attempts to fill this gap.
In this paper, I start by giving arguments against Mohammed (1990)'s claim that Arabic does not have raising with seem-type verbs. He analyses the constructions involving 'yabdu' as "examples of left dislocation" (Mohammed 1990:101). This paper demonstrates that there are differences between raising constructions and left-dislocation constructions in SA. Evidence is also presented concerning the distinction between raising and control verbs in SA. I have also argued against Wardhaugh (1995)'s claim that English does not distinguish between raising and control verbs. Having established raising in the two languages and using Principles and Parameters (cf. Chomsky 1995) as a frame of reference, I compare raising to subject in English and SA. The comparative approach used here focuses on points of similarity and points of difference between the two languages. The paper is organized as follows: in section 1 I discuss raising-to-subject in SA. In section 2 I tackle raising to subject in English. In section 3 I point out the similarities and differences between the two languages.
1. Raising to Subject in Standard Arabic
Mohammed (1990) claims that SA does not have raising with "seem-type" verbs.2 I claim that SA has such verbs as yebduu/ yadharuu 'seem' which trigger raising the subject (cf.1b), direct object (cf.2b). or indirect object (3c) of the complement clause) into the matrix clause (cf. Salih 1986:333). as in:3
(1) a. yabduu [?anna l-mu9allim-a Šaraha 1-qasiidat-a]
seem that the-teacher-Acc explained the-poem-Acc
'It seems that the teacher explained the lesson'
b.yabduu l-mu9allim-u[?anna-hu Šaraha 1-qasiidat -a]
seem the-teacher-Nom that-he explained the poem-Acc
'The teacher seems to have explained the poem'.
(2) a. yabduu ?anna [as-sayyaara-ta daraba-t al-walad-a]
seem that the-car-Acc hit-F the-boy-Acc
'It seems that the car hit the boy.'
b. yabduu 1-walad-u [?anna as-sayyaara-ta daraba-t-hu]
seem the-boy-Nom that the-car-Acc hit-F-him
The boy seems to have been hit by the car.
Lit. 'The boy seems that the car hit him.'
(3) a. yabduu [?anna 1-walada-a ?a9taa kitaab-an li-1-bint-i ]
seem that the-boy-Acc gave the-book-Acc to-the-girl
'It seems that the boy gave the book to the girl'
b.tabduu 1-bint-u [?anna al-walad-a ?a9taa-ha kitaab-an]
seem the-girl-Nom that the-boy-Acc gave-her hook-Acc
Lit. 'The girl seems the boy to give her a book.'
c. tabduu 1-bint-u [?anna l-walad-a?a9taa kitaab-an la-haa]
seem the-girl-Nom that the-boy-Acc gave book- Acc to-her
'The girl seems to have been given a book'
Lit. 'The girl seems the boy gave a book to her.'
d.*tabduu 1-bint-u[?anna[1-walad-a ?a9taa kitaab-an li]
seem the girl-Nom that the-boy-Acc gave book-Acc to
Lit. 'The girl seems the boy to give a book to.'
We know from sentences like (1-3) that a subject, a direct object or an indirect object can be raised up to become the subject of the matrix clause.4 For instance, the structure of (lb) can be given as in:
[TP [VP yabdu al-mu9allimu [CP ?anna COPY [ TP t [VP Šaraha t ad- darsa]]]]
(4) shows that a DP (determiner phrase) originating in the specifier position within the VP of the complement clause can be raised to Spec-TP (i.e. the subject of the matrix clause). Note that the raised element is not abruptly raised to the specifier position of the yabdu-clause. Rather, it is raised up in a successive cyclic fashion: it is first raised to the Spec-TP position in the complement clause. It is then raised to Spec-VP position before it is finally raised to the higher Spec-'I P position.
Evidence that the DP (al-mu9allimu) originates in Spec-VP (the fahima- clause) in (1b) can be drawn from different sources: first, notice that the raised DP leaves behind a trace (t) in all positions to which it is raised except for the spec-TP position in the complementiser phrase (CP) where a pronominal copy is left. 'The pronominal copy reflects the status occupied by the raised element prior to raising to a higher position: If the raised DP originates in the Spec-VP position in the CP, the pronominal copy gets cliticized onto the complementizer (?anna) as in (1b). If it is the complement of the verb in the CP, the pronominal copy gets cliticized onto the verb, as in (2b). If, however, the raised DP originates as the projection of the preposition li (complement of a preposition like li), the pronominal copy is cliticized onto the preposition as in (3c).
Two facts about this type of citicization is in order: (i) note that the pronominal copy is obligatorily attached to the appropriate element alter raising has applied; if it does not, the sentence is ill-formed:
(5) a. *yabduu 1-walad-u [?anna fahima d-dars-a]
seem the-boy-Nom that understood the-lesson-Acc
lit. 'The boy seems that understood the lesson.'
b .*yabduu 1-walad-u [?anna s-sayyaara-ta daraba-t]
seem the-boy-Nom that the-car-Acc hit-fem
lit. 'The boy seem that the car hit.'
(ii) the pronominal cop y of the raised element is cliticized onto the complementizer (but not onto the verb) if the raised clement occupies the Spec-VP position in the complement clause prior to raising. Conversely, the copy is cliticized onto the verb (but not onto the complementizer) if the raised element occupies the position of complement to the V:
(6) a. *yabduu 1-walad-u [?anna fahima-hu d-dars-a]
b. *yabduu 1-walad-u [?anna-hu s-sayyaara-ta darabat]
Another argument for the subjecthood of the raised entity in the VP of the complement clause comes from verb agreement: In SA, if the subject precedes the verb (as in complement clauses), the verb must agree with its local subject in number, person and gender (cf. Mohammed 1990:95)5
(7) a. qaal-uu [?inna 1-walad-a fahima d-dars-a]
said-3pl that the-boy-Acc understood-3Msg the lesson-Acc
'They said that the boy understood the lesson.'
b. qaal-uu [?inna 1-bint-a fahima-t d-dars-a]
said-3pl that the-boy-Acc understood-3Fsg the-lesson-Acc
'They said that the girl understood the lesson.'
c. qaal-uu [?inna l-?awlaad-a fahim-uu d-dars-a]
said-3pl that the-boys-Acc understood-3Mpl the-lesson-Acc
'They said that the boys understood the lesson.'
d. qaal-uu [?inna 1-binaat-a fahim-na d-dars-a]
said-3pl that the girls-Acc understood-3Fpl the-lesson.
'They said that the girls understood the lesson.'
The sentences in (7) clearly show that the complement verb obligatorily agrees with its own subject: for example, in (7b) tine DP subject (?al-bint) cues agreement on the verb which carries the ending -t (indicating a third person singular feminine subject). In (7c), the ending -uu indicates that the agreement is with a third person masculine plural subject. If. however, the third person plural subject is feminine, the marker -na appears on the verb. as in (7d).
Returning to our raising sentences like (1b), we note that the verb carries markers that indicate the status of the raised DP in the VP. Thus the use of a plural nominal instead of the raised element in the SA sentence (1b) entails a different agreement marker:
seem the-boys-Nom that-them understood-they the-lesson
In (8) the complement verb happens to agree with the DP subject of the matrix clause. Since agreement is clause internal, we assume that the DP subject l-?awlaad leaves a trace in the Spec-VP position before it is raised up. The complement verb now agrees with the trace of the raised DP. Thus verb agreement provides us with evidence that the raised element is subject in the complement clause.
We have pointed out that in SA the DP subject, direct object or indirect object of the complement clause is raised up to become the subject of matrix-clause. But we have not provided any evidence for this claim. I will provide two arguments to substantiate this claim: first, note that in SA, the subject of a clause follows the verb in a matrix clause and precedes it in a complement clause:
think- the-man-Nom that the-thief stole the-money-Acc
'The man thinks that the thief stole the money.'
It should be noted that sentences like (9) contain two clauses: a matrix clause and a complement clause (introduced by a complementizer like ?anna). The subject occupies different positions in the two parts of the sentence: "the subject of the main clause (rajulu) follows the matrix clause verb (ya9taqidu): the subject of the complement clause (?allissu) precedes the verb of the complement clause (saraqa). Given this, if (in a raising construction) the subject of the complement clause is raised up into the subject position in the matrix clause, we expect it to follow the matrix verb. That this is the case can be seen in (1b) where the raised element occupies the post-verb position.
Another argument can be drawn from verb agreement. Note that in SA we have two sets of agreement endings: one set is used with complement verbs and another set with matrix verbs. The former is already attested when we marshaled evidence for the status of the raised element in the complement clause. We have noted that a complement verb agrees with its subject in person, number, and gender. In contrast, a matrix verb agrees with its subject in gender only (cf. Holes 1995:213):
says the boy-Nom that the-sky-Acc clear
'The boy says that the sky is cloudless.'
M-say the boys-Nom that the-sky-Acc clear
'The boy says that the sky is cloudless.'
c. ta-quulu 1-bint-u [?inna s-samaa?-a saafiyat]
F-says the girl-Nom that the-sky-Acc clear
'The girl says that the sky is cloudless.'
F-say the-girls-Nom that the sky-Acc clear
'The girls say that the sky is cloudless.'
In (l0a-d) the matrix verb precedes the matrix subject; the verb is singular irrespective of the number of the subject. We also note that the verb takes the prefix ta- in agreement with a feminine subject as in (l0c-d). Given this, if the raised element ends up as the matrix subject in sentences like (1b) and (2b), it should manifest gender agreement. That this is the case can be seen if we substitute a feminine subject for the masculine subject in (1b):
F-seem the-girl-Nom that-he understood-F the-lesson-Acc
'The girl seems to have understood the lesson.'
Like (1b) in which the matrix verb is both masculine and singular, the matrix verb in (11) agrees with the subject in gender: this agreement is manifested by the prefix on the verb. Note that the sentence is ill-formed if there is no (matrix) verb-
subject agreement. Thus verb agreement provides further evidence for the post-raising status of the raised element.
Note that the above analysis argues against Mohammed (1990)'s claim that "Arabic does not allow raising with 'seem'-type verbs, whether or not a trace or a pronominal is in the 'extraction' site."(p. 101). He argues that in sentences like (12b) and (12c), which are considered as examples of left dislocation with the matrix verb yabduu 'seem', "the verb yabduu 'seem' never changes its features of number, gender, and person since no NP or DP can move into its subject position" (Ibid.).
(12)a. yabduu [?anna I-'?awlaad-a/l-banaat-i saafaruu/saafarna]
seem that the-boy-Acc/the-girls departed 3Mpl/3Fpl
'It seems that the boys departed.'
b. al-?awlaad-u pro yabduu [?anna-hum saafaruu)
the-boys-Nom PRO seem that-they departed 3Mpl
'The boys, it seems that they departed.'
c. al-binaat-u pro yabduu [?anna-hunna saafarna]
the-girls-Nom PRO seem that-they departed 3Fpl
'The girls, pro seems that they departed.'
Mohammed (1990:101) goes on to claim that "the expletive pronoun is the only candidate that meets the requirements of subjecthood for such a verb". I have argued that in SA raising to subject constructions, a complement clause subject or DO or IO is placed after the matrix clause verb with which it agrees in gender:
(13) yabduu l-?awlaad-u [?anna-hum saafaruu)
(14) tabduu 1-binaat-u [?anna-hunna saafarna]
cf. *yabduu 1-binaat-u [?anna-hunna saafarna]
If the expletive pronoun is the "only candidate that meets the requirements of subjecthood" for such a verb as yabduu, why does the verb change its gender feature so that it agrees with al-?awlaadu/?albinaatu in (13)-(14)? The fact that these nominals cue agreement on the verb is evidence that they now occupy the Spec-VP position in the matrix-clause in sentences like (13)-(14). Thus there is evidence that SA allows raising with 'seem'-type verbs.
Another problem with Mohammed's analysis is that it simply denies the existence of the raising phenomenon in SA: Mohammed (1990) deals with sentences involving a verb like yabdu as examples involving left-dislocation rather than raising. It might be argued that such sentences involve a left-dislocated DP which leaves behind a pronominal copy in the position out of which it moved.
?almu9allim-u sa?ala-hu t-tilmiiz-u
the teacher-Nom ask-him the-student-Nom
'The teacher, the student asked him'
Haoaan ?ar-rajulaan ?al-9ajiibaan la yabduu ?ana
These the-two men then strange don't seem that
humaa ya- tagyyaraani ma9 ?az-zaman
-they change with the-time
Lit. 'These two strange men don't seem that-they
change with the time'. (Qasr all-shooq (Palace of
Note that the raising sentences we have discussed are different from sentences like (15) or (16) in that raising sentences begin with the verb yabdu. which is followed by the raised DP. There is a weak agreement between the raising verb and the following DP: agreement only in person and gender. In contrast, the sentences with left dislocation start with the dislocated DP and they involve agreement in person, gender and number.
The fact that this type of verb can permit left dislocation does not necessarily rule out the possibility of raising in such constructions. If Mohammed's arguments were viable, we would also deny the existence of passive sentences in SA on the basis of sentences like (17)
the-boy-Nom hit-him the-teacher-Nom
'The boy, the teacher hit him.'
Thus the fact that sentence (17) involves left-dislocation does not rule out the validity of passive sentences in SA:
hit-pass the-boy-Nom by the-teacher-Ob!
'The boy was hit by the teacher.'
The above discussion has shown that one difference between a structure involving left dislocation and structures such as passive and raising is that a left dislocated DP precedes verbs, whereas a raised or passivized DP follows the verb.
1.1The difference between raising verbs and control verbs in SA
It might be claimed that Arabic does not distinguish between verbs like yabduu 'seem' and verbs like haawal 'try'. This section sets out to demonstrate that there are differences between these verbs: verbs like yabduu are raising, verbs, whereas verbs like haawal are control verbs in SA. Note that SA has such control verbs as haawal and ?araad 'want', as in (19a) of which the structure is represented as in (19b):
a. haawal ar-rajul-u [?an(*-hu) pro yadriba l-bint-a]
b. [TP [VP haawal ar-rajulu [CP ?an [TP [VP
In (19), the PRO has a controller (=ar-rajulu) in the haawal-clause. An important question to ask is how we can distinguish between a control verb like haawal and a raising verb like yebduu in (20):
seem the-man-Nom that-he hit the-girl
'The man seemed to have hit the girl.'
One difference between the two is that control verbs like haawal take a complement with a PRO as subject. Such verbs subcategorize for the complementizer ?an 'that' which requires a verb in the subjunctive mood to immediately follow it; it cannot be followed by the complement subject:
(21) * haawalat al-bint-u [?an ar-rajul-a daraba-haa]
tried-she the-girl-Nom that the-girl-Acc hit-her
lit. 'The girl tried that the man hit her.'
In contrast, a raising verb like yabduu subcategorizes for the complementizer ?anna (cf.20) which is not necessarily followed by the complement verb:
seem-she the-girl-Nom that the man-Acc hit-her
lit. 'The girl seems that the man hit her.'
(22) shows that the complementizer is immediately followed by the complement subject and the complement verb is not in the subjunctive mood.
Another difference is that no pronominal copy can he cliticized onto the complementizer ?an in sentences with a control verb (cf.19). whereas raising verbs allow a pronominal copy to be cliticized onto (the complementizer (cf. 20).
2. Subject-to-subject raising in English
English has predicates such as seem and appear that trigger Subject-to- Subject raising, as in (23b) and (24b):
b. The boy does [seem to understand the lesson].
(24) a. It does seem [that the car hit the boy].
b. The boy does [seem to have been hit by the car].
In (23a) the boy is the subject of the complement clause (understand the lesson): in (23b) the boy is raised up to become the subject of the matrix clause. In (24a) the boy is the direct object of the complement clause—the hit-clause; in (24b) the boy first becomes the subject of the complement clause and is subsequently raised up to become the subject of the matrix clause. Thus the structure of sentences like (23b) can be represented as in (25):
[TP the boy does [VP t seems ['IP t to [VP t understand the lesson]]]]
As shown in (25), the DP, the boy, is raised in a successive cyclic fashion: it originates in the Spec-VP position (understand the lesson). It is then raised up to the Spec-TP position (to understand the lesson), then to the Spec-VP position (seems to understand (he lesson). Finally, it is raised up to the Spec-TP position (does seem to understand the lesson). Each raising operation results in a trace.
Similarly, the DP direct object in (24b) cannot be directly raised up to become the matrix-clause subject; it originates as the complement of the verb hit. It has to first be raised to the Spec-VP position, then raised to the Spec-TP, then to the Spec-VP position (seem) before it is raised to its final destination, namely, the Spec-TP position. Any attempt to directly raise it will result in an ill-formed sentence:
(26) *The boy seems the car hit.
Similarly, an indirect object cannot directly raise to become the matrix- clause subject: only when it becomes the subject of the complement clause can it raise to become the matrix-clause subject, as in (27)
(27c) shows that any attempt to directly raise the indirect object the girl as the subject of the matrix results in an ill-formed sentence. In (27b) the indirect object the girl first becomes a direct object and is then made the complement subject which is subsequently raised up to become the subject of the matrix clause. Evidence that the indirect object first becomes a direct object can be seen in (27b) where the preposition, which marks an indirect object (cf.27a), disappears from the sentence. Note also that the presence of the passive auxiliary (been) which is followed by the past participle (given) provides evidence for the fact that passivization has applied in the complement clause.
It has been claimed that the DP the boy in sentences like (23b) originates as the subject of the complement clause (understand the lesson). Is there any evidence for such a claim? The first argument comes from reflexivization. Note that in English reflexivization can only apply if the reflexive and its antecedent are located within the same clause:
(28) I believe that the boy has injured himself/*myself
In (28) the reflexive nominal can only refer to the boy but not to I. Why so? The DP the boy shares the same clause (injured-clause) with the reflexive. According to the Binding theory (Chomsky 1981; 1982), the reflexive is bound (not free) by the c-commanding DP the boy: the reflexive is not c-commanded by I in (28); hence, the reflexive is not construed with I.6 This characterization seems to be violated by the following sentence where the reflexive is construed with the DP the boy which is located in a different clause; hence it does not c-command the reflexive pronoun:
(29) The boy seems to me to have injured himself.
One might argue that (29) is a single clause; therefore, the reflexive can be construed with the boy. If this were true, the reflexive should be able to refer to any nominal in sentences like (29):
(30) The boy seems to me to have injured himself/*myself.
The fact that the reflexive cannot be construed with me is evidence that we have more than one clause in raising sentences like (29)-(30). The only way to solve this problem is to claim that the DP the boy in sentences like (30) has originated in the complement clause before it is raised up to a higher position, leaving behind an empty category trace as the subject of the complement clause with which the reflexive can be construed:
(31) The boy seems to me [to have t injured himself]
According to the Binding Conditions, himself (as an anaphor) must be c- commanded by an antecedent in its governing category. The reflexive himself in sentences like (31) is c-commanded by the trace t since the first branching node above t is the VP node which dominates himself; hence the trace is the only antecedent of the reflexive anaphor (himself). Thus reflexivization provides us with evidence for the claim that the boy in sentences like (23b) originates as the subject of the complement clause.
Another argument comes from predicate agreement (cf. Radford 1981:186). In English a predicate nominal agrees with the subject of its own clause. Thus in the following sentences, it agrees with the subject of the complement clause but not with the subject of the matrix clause:
(32) a. The women think [the man is a fool/* fools]
b. The woman thinks [the men are fools/*a fool]
In (32a) the predicate nominal a fool is in agreement with the man but not with the women. The reason is that both the predicated nominal a fool and the DP the man share the same clause. Conversely, in (32b) the predicate nominal fools agrees with the men but not with the woman. It is clear from (32a-b) that the predicate nominal must agree with the subject of its own clause. In other words. the predicate nominal agrees with the DP occupying the subject position in the complement clause.
Although this principle is observed in most English sentences, the following sentences seem to pose problems for our agreement facts (cf. Radford 1981):
(33) a. The man seems [to be a fool].
b. The men seem [to be fools].
Here the predicate nominal agrees with the subject of the matrix clause: note that in (33) the complement clause does not have an overt subject in the VP-position: the predicate nominal seems to agree with the main-clause subject. How come the predicate nominal agrees with the subject in the matrix clause in (33) but not in (32b)? How can we resolve this problem? This problem can be overcome if we assume that the DP (the man) originates in the Spec-VP position of the complement clause (be-clause) and is subsequently raised up to become the subject of matrix clause (the seems-clause), leaving behind a separate trace in the vacated position:
(34) a. The man seems [t to be a fool]
b. The men seem [t to be fools]
Now, if we further assume (following Chomsky 1995) that "a trace is a silent copy of the relevant moved constituent, it follows that traces will have the same syntactic and semantic properties as their antecedents" (Radford 1997:320), we can say that the predicate nominal in (34) agrees with the trace which is singular in (34a) and plural in (34b). Thus agreement provides further evidence for raising in sentences like (23b).
There are numerous arguments that show that the boy in sentences like (23b) ends up as the matrix-clause subject: One of these arguments comes from word order. In English the subject precedes the verb of its own clause: thus in (23a) the DP the boy does not precede the verb seems but the verb understand. In contrast, the DP the boy ends up as preceding the verb seems in (23b) indicating that it is the subject of matrix clause.
Another argument comes from verb agreement. In English, the verb agrees with its subject:
b. The men like/*likes the dogs.
In (35a-b) the verb agrees with its subject (the man/the men). If we now go back to (23b), the verb seems obligatorily agrees with the matrix-clause subject (the boy):
(36) The boy seems/*seem to understand the lesson,
Thus verb agreement is a further argument that the boy in sentences like (23b) must be the subject of the matrix clause).
2.1 The difference between raising verbs and control verbs in English
Wardhaugh (1995:169) claims that verbs like want are raising verbs in English. In the words of Wardhaugh, "in sentences like I want him to go, a constituent (him) has been taken from the embedded clause and raised to become a constituent in the matrix clause" (Wardhaugh 1995:169). In this section, I argue that verbs like try and want are control verbs in English. Constructions containing such verbs are called subject/object-controlled equi which deletes a complement-clause subject under coreference with a matrix-clause subject or direct object. Consider:
(37) She will try [to PRO help you]
where PRO has a controller (=she) in the will try clause. There are several differences between a controller verb like try and a raising verb like seem (cf. Radford 1997a). Consider the following sentence:
(38) She seems to have helped you.
First, note that controller verbs like try do not allow a dummy subject like expletive there (cf. 39a), whereas raising verbs like seem allow such a dummy subject (cf. 39b); also raising verbs like seem can take the expletive it as a subject (cf.l6c) whereas non-raising verbs like want do not take the expletive it as a subject (39d):
(39a) is ill-formed because PRO, which is a referential pronoun, needs a referential expression to be its controller. Since there is a nonreferential pronoun. it cannot serve as the controller of PRO. In (39b) there, which is a nonreferential element, originates as the subject of the complement clause and is then raised up to become the subject of the matrix clause.
Also, as Radford (1997:338) puts it, (39a) is a violation of the č-criterion which requires that "each argument bears one and only one č-role, and each č-role is assigned to one and only one argument" (Chomsky 1981:36). Thus (39a) is ill- formed because the verb try assigns the č-role AGENT to its subject which must be an expression denoting a rational being. Since there is a nonreferential dummy pronoun, the requirement for try to have an AGENT subject is not met in sentences like (39a). Conversely, the verb seem does not assign a role to its subject (i.e. it does not č-mark its subject); hence, it is appropriate to have expletive there as matrix-clause subject in sentences like (39b)
3. Similarities and Differences between SA raising and
Note that both SA and English have raising to subject: the subject of the complement clause is raised up to become the subject of the matrix verb as evidenced by sentences like (1b) and (23b). In both languages the raised DP cues agreement on the matrix verb as seen in (1b) and (23b). In both SA and English the complement clause—prior to raising-is introduced by a complementizer (?anna in SA and that in English). Also, both languages distinguish between raising verbs and control verbs.
However, SA and English show the following differences: First, the subject that ends up as the subject of the matrix verb happens to occur in preverbal position in English and in post-verbal position in SA as in:
(34) a. The man seems [to have understood the lesson].
b. yabduu r-rajul-u [?anna-hu fahima d-dars-a]
Second, English does not allow other than a subject to directly raise (cf.35c). Thus (35c) is ill-formed because a direct object is permitted to directly raise up as subject of the matrix clause: the direct object should become subject before raising as in (35b):
(35) a. It seems [that the car hit the boy]
b. The boy seems [to have been hit by the car]
c. *The boy seems [the car hit]
Note that in (35b) passivization has applied in the complement clause such that the direct object is turned into subject and is then raised into the subject of the matrix clause.
Conversely, in SA a direct object (36b) or an indirect object (36c) can he directly raised up to become the subject of the matrix verb. Also, the SA indirect object can be a direct object which is subsequently raised into the matrix clause, as in(36d):
(36) a. yabduu [?anna 1-walad-a ?a9taa risaalat-an li- 1-bint-i]
seem that the-boy-Acc gave letter-Acc to-the-girl-Obl
'It seems that the boy gave a letter to the girl.'
b. tabduu r-risaalat-u [?anna 1-walad-a-a ?a9taa-ha li- 1-bint-i]
seem the-letter-Nom that the-boy-Acc gave-it to the-girl-Obl
The letter seems to have been given to the girl.
Lit. 'The letter seems the boy gave it to the girl.'
c. tabduu 1-bint-u [?anna 1-walad-a ?a9taa risaalat-an la-haa]
seem the-girl-Nom that the-boy-Acc gave letter-Acc to-her
'The girl seems to have been given a letter to.'
Lit. 'The girl seems the boy gave a latter to her.'
d. tabduu 1-bint-u ?anna 1-walad-a ?a9taa-ha risaalat-an
seem-she the-girl-Nom that the-boy-Acc gave-her letter-indef
Lit. 'The girl seems that the boy gave her a letter.'
(36b) involves raising a direct object. If the DP ?rrisaalatii in (36b) is directly raised up to become the subject of the matrix clause, the pronominal copy should be cliticized onto the complement-clause verb. This prediction is borne out by sentences like (36b) where the copy appears on the verb. indicating that the raised DP must be a direct object in the complement clause. In addition, the verb maintains its active form; if passivization is applied in the complement clause to turn the direct object into subject, the verb should take the passive form (duriba). In contrast (36c) involves the direct raising of an indirect object into the subject position in the matrix clause. Evidence that the raised NP is directly raised comes from the fact that the pronominal copy is cliticized onto the preposition in the complement clause; if it were made direct object first, the copy would be cliticized onto the verb as in (36b). Thus English has subject-to-subject raising only, whereas SA has subject-to-subject raising, direct object-to-subject raising, and indirect object-to-subject raising..
Fourth, we have already noted that prior to raising the raised DP in both SA and English originates in the complement clause which is introduced by a complementizer. After raising has applied, the complementizer disappears from the clause in English (cf. 35b) and the complement verb, being an infinitive, no longer agrees with the subject. In contrast, the complementizer remains intact after raising has applied in SA (cf. 36b). Also raising has no effect on subject-verb agreement in the complement clause: the verb maintains agreement with its subject.
Fifth, the raised DP in English leaves behind an empty category trace as the subject of the complement clause (cf. Radford 1997:335). Evidence (e.g. reflexivization) has been given for the status of the trace in English. In contrast. the raised DP in SA leaves behind a pronominal copy of itself in the complement clause; the pronominal copy is cliticized onto the complementizer if the subject is raised (1b) and is cliticized onto the verb if the direct object is raised (2b). However, if the indirect object is directly raised, the pronominal copy is cliticized onto the preposition in the complement clause (3c). Thus the Raising rules in SA differ from those of English in an important way: the SA Raising rules are copying rules.7
Sixth, both SA and English distinguish between raising verbs and control verbs. In SA control verbs like haawal take a complement with PRO as subject and subcategorize for the complementizer ?an which is immediately followed by a verb in the subjunctive mood. Conversely, raising verbs like yabduu in SA subcategorize for the complementizer ?anna which is not necessarily followed by a complement verb. Moreover, in SA no pronominal copy is cliticized onto the complementizer ?an in sentences with control verbs, whereas subject-to-subject raising in SA allows a pronominal copy to be cliticized onto the complementizer. In English control verbs like try do not subcategorize for the complementizer (that) and do not allow a dummy subject (e.g. there) to serve as the controller of PRO.
This paper has presented arguments against Mohammed (1990)'s claim that Arabic does not have raising with seem-type verbs. It has been shown that SA has raising to subject. I have also argued that there is a difference between raising verbs and control verbs in English and SA. Thus the claim presented by Wardhaugh (1995) is arguably inaccurate. The paper lias also pointed out certain similarities and differences between English and SA concerning raising constructions. Both languages have subject-to-subject raising, but only SA has object-to-subject and indirect object-to-subject raising. In both languages the complement clause is introduced by a complementizer. In both languages the raised DP agrees with the matrix-clause verb. However, the two languages have their differences: The raised DP is placed before the matrix-clause verb in English and after the matrix-clause verb in SA. In English the raised DP leaves behind an empty category trace in the complement clause, whereas in SA the raised NP leaves behind a pronominal copy in the complement clause. Furthermore, in English the complementizer disappears (from the complement clause) after raising and the complement clause becomes non-tensed. In SA the complemcntizer remains intact after raising and the clause is tensed.
Some of the SA raising sentences are drawn from Salih 1986. I asked some native speakers who teach at the Department of Arabic, Faculty of Arts at Qena whether the raising sentences are acceptable or not. Most of them claimed that such sentences can be used in writing but warned that they would sound nonnative. However, Arabic is replete with nonnative constructions that came into the language as the result of translation. I would like to thank Dr. Aatif. Who is a lecturer at the Department of Arabic language (Qena). for his help and advice on the adequacy of sentences used in this section. Any errors, stylistic or otherwise, are mine. The following abbreviations are used throughout the paper:
DO Direct Object
MV Matrix Verb
1 First person
2 Second person
3 Third person
Gary and Gamal-Eldin (1980:55-56) report that Cairene Colloquial Arabic has object to subject raising as in (ii):
(i) sahl ?innu jihdim ?ilbitillu
easy that he digest veal
'It is easy that he digests veals.
(ii) ?ilbitillu sahl ilhadm
veal (is) easy digesting
'Veal is easy to digest.'
?ilbitillu, which is direct object in (i), is raised up to become subject in (ii).
Mohammed(1990) argues that "(i) in Arabic, when the verb precedes the subject there is no agreement between the two; (ii) in such constructions, the verb agrees with the expletive pronominal whose properties were established to be third person masculine singular" (p. 119)
Chomsky (1981:188; 1982:21) formulates the conditions of the Binding Theory as follows:
(A) An anaphor is bound in its governing category.
(B) A pronominal is free in its governing category.
(C) An R-expression is free.
These conditions are known as Condition A, Condition B, and Condition C, respectively.
The Greek Raising rules are also copying rules: when the complement subject is raised into the matrix clause by Raising rules, a pronominal copy of the raised NP (agreeing with it in person, gender, and number) is left behind in the complement (Soams and Perlmutter 1979:161) But the pronominal copy, if not emphasized or contrasted, can be deleted by Subject Pronoun Drop (Ibid., p. 162)
Alkhuli, Muhammad A. 1997. Comparative Linguistics:
English and Arabic. Jordan: Alfalah House.
Anani, M. 1983. Some Differences between Arabic and English
Comparative Structures. Papers and Studies in Contrastive
Linguistics 17. pp. 79-84.
Atawneh, A. 2001. A Contrastive Study of Aspect in English and
Arabic. I.IAES 2.1&2. pp. 127-140.
Murtadha. 1996. Notes on Passive and Pseudo
Intransitive Constructions in English and Arabic. Paper
and Studies on Contrastive Linguistics 31. pp. 39-50.
Bakir, Murtadha. 1999. Verb Movement, Subject Movement, and
word Order in English and Arabic. In Proceedings of the
First International Conference on Arabic-English
Contrastive and Comparative Studies, pp. 173-181.
Chomsky, Noam. 1981. Lectures on Government and Binding.
Chomsky, Noam. 1982. Some Concepts and Consequences
Chomsky, Noam. 1995. The Minimalist Program. Cambridge.
Mass.: MI I press.
Gary, Olmsted and Saad Gamal-Eldin. 1980. Cairene Egyptian
Colloquial Arabic. Amsterdam: North-Holland Publishing
Gerdts, Donna. 1980. "Towards an Ergative Analysis of
Ilokano", Paper presented at the conference of the Canadian
Linguistic Association held in Montreal, June 2, 1980.
Holes, Clives 1995. Modern Arabic: Structures, Functions
and Varieties. London: Longman.
Homeidi, M. 2000. A Syntactic Contrastive Analysis of the
Relative Clause in Arabic and English in the GB with
Reference to Translation. Papers and Studies in Contrastive
Linguistics 36. pp. 95-110.
Hussein, Ahmed-Sokarno. 1999b. Negation in Cairene
Colloquial Arabic, English and French. In Proceedings of
the First International Conference on Arabic-English
Contrastivc and Comparative Studies. Pp. 183-191.
Mohammed, M. A. 1990. "The Problem of Subject-Verb
Agreement in Arabic: Towards a Solution", in Mushira Eid
(ed.). Perspectives on Arabic Linguistics I: Papers from the
First Annual Symposium on Arabic Linguistics.
Amsterdam: John Benjamins Publishing Company. pp. 95-
Mukattash, Lewis. 2001. A Preliminary Bibliography of Arabic-
English Contrastive Studies. IJAES 2:1&2. pp. 229-240.
Ouhalla, Jamal. 1994. Transformational Grammar: from
Rules to Principles and Parameters. London: Edward
Perlmutter, David and Paul Postal. 1983. "The Relational
Succession Law". in David Perlmutter (ed.). Studies in
Relational Grammar I, Chicago: University of Chicago Press.
Postal, Paul. On Raising. Cambridge, Mass.: MIT Press.
Radford, Andrew. 1981. Transformational Syntax.
Cambridge: Cambridge University Press.
Radford, Andrew. 1997a. Syntactic Theory and the Structure
of English: A Minimalist Approach. Cambridge: Cambridge
Radford, Andrew. 1997b. Syntax: A Minimalist Introduction.
Cambridge: Cambridge University Press.
Salih, Mahmud Hussein. 1986. "Raising in Universal
Grammar", Eastern States Conference on Linguistics 2:
Wardhaugh, Ronald. 1995. Understanding English
Grammar: A Linguistic Approach. Oxford: Blackwell.