2. SUBJECTS
2.1. Preliminaries
2.1.4. The Issue of Terminology
2.1.4.4. Olarrea (1996)
Rhematization is, by definition, emphatic, since it presupposes that the rhematized element carries contrastive focus.
LD structures, such as (48), are equivalent to what Hernanz and Brucart term discursive anaphor. Thematization parallels CLLD structures in (49), and rhematization phenomena are defined in Olarrea's terminology as focus fronting constructions, as in (50).150
When focusing on the position of the subject as occupying this absolute initial position, Olarrea (1996) defends that Spanish sentences with preverbal subjects have all the characteristic properties of what Cinque (1990) calls clitic left- dislocation constructions (that have neither a pause nor a special intonation), as in (51), where the empty element is the equivalent to lo in (52b) below:
(51) Juani [ e ]i fue a la fiesta [John [e] went to the party]
He adds that these properties differ in certain important aspects from those of ordinary left dislocation found in English.151
Olarrea (1996) also points out the terminological disagreement in the literature with respect to the classification and description of the properties of left- dislocated structures in Spanish. In any case, and under his terminology (LD,
150 Liceras (personal communication) points out that in examples like that in (50), we may be dealing with two different structures: one that makes use of a special intonation and a pause between the focus-element/rheme and the rest of the sentence, as in (i); and a second one in which no pause appears, as in (ii):
(i) esas flores, compré ayer [these flowers, buy-1stps-past yesterday]
(ii) ESAS FLORES compré ayer [THESE FLOWERS buy-1stps-past yesterday]
In the first case, no movement operation takes place and no emphasis is used. The second is perhaps the result of movement.
151 Cinque (1981) already pointed out the different treatment that structures such as those in (i) and (ii) should receive:
(i) Maríai , ésai sí que está como una cabra [María, she is nuts indeed]
(ii) ¿Maríai ? lo sabe todo [ e ]i [María? it know-3rdps all] [María? she knows it all]
CLLD) or Hernanz and Brucart's (1987) (discursive anaphor, thematization), these structures are clearly opposed to topicalization/rhematization processes (dislocation, in the case of Hernanz and Brucart 1987).152
There exist important differences between LD and CLLD constructions.
What follows summarizes the differences between the two constructions taking into consideration pragmatic, stylistic and syntactic issues:
1) both LD and CLLD constructions involve a left-dislocated element, as in (52):
(52) a. Juani , no me acuerdo de éli
[John, not me remember-1stps of him]
[John, I don’t remember him]
b. a Juani loi vimos en la fiesta
[to John him see-1stpp-past in the party]
[John we saw him at the party]
The difference between them is that while LD can only have NPs as dislocated elements, CLLD allow for any phrasal type to be dislocated, such as the PP a Juan in (52b).153
152 On focus, see Rochemont (1986), Campos (1986) and Campus and Zampini (1990).
153 When dealing with subjects, the fact that phrases other than NPs can be clitic left dislocated does not apply. It may be necessary to point out, though, that CPs and some PPs can be preverbal subjects in Spanish:
(i) [entre Juan y Pedro] arreglarán el coche (PP) [[between Juan and Pedro] mend-3rdpp-future the car]
(ii) [que tú no vayas a terminar la tesis] es increíble (CP) [[that you are not going to finish your thesis] is unbelievable]
Nevertheless, we may argue as well that in the cases of (i) and (ii) there is an empty subject and that the structures between square brackets are adjuncts that are located in a position higher than [Spec IP].
2) the dislocated element is co-referent with another element in the sentence as in (53), but while LD requires a phrase or a pronoun (either a clitic or a tonic pronoun), CLLD can only have an empty pronominal (licensed by agreement or by a clitic; since Spanish has no subject clitics, the presence of strong agreement is what licenses the empty resumptive pronoun, as in (51) above):154
(53) a. John Coltranei , ese saxofonistai me encanta [John Coltrane, this saxophonist me love-3rdps]
[John Coltrane, I love this saxophonist]
b. en Juani no es posible confiar [ e ]i
[in John not is possible trust-infinitive [e] ] [it is not possible to trust in John]
So LDs may or may not be constructed with a clitic, while CLLDs require the obligatory presence of a clitic or the corresponding strong agreement (in the case of subjets) in order to license the gap.
3) Topicalizing expressions are possible in LD, as in (54a), but cannot be used in CLLD, as the ungrammaticality in (54b) shows, since there is no pause between the dislocated constituent and the rest of the structure:
(54) a. en cuanto a Ismaeli , éli no va a suspender el examen [as for Ismael, he not go-3rdps to fail the exam]
[as for Ismael, he is not going to fail the exam]
154 The parallelisms between strong Agr and clitics in the Romance Languages have been pointed out in several works since Rizzi (1982).
b. *en cuanto a Juani loi vi ayer155
[as for John him see-1stps-past yesterday]
[as for John I saw him yesterday]
4) LD constituents take absolute first position, while CLLD ones can be freely embedded as the ungrammaticality in (55a) versus the grammaticality in (55b) shows:156
(55) a. *todos piensan que Ismaeli , éli no va a suspender el examen [all think that Ismael, he not go-3rdps to fail the exam]
[everybody thinks that Ismael, he is not going to fail the exam]
b. todos piensan que de Juani no deberíamos hablar [e]
[all think that about John not should-1stpp talk [e] ] [everybody thinks that we should not talk about John]
5) The examples in (52a) and (54a) show that identity of Case and subcategorization between the dislocated element and the co-referential element is not necessary in LD. The example in (56) shows a mismatch between the Case
155 The ungrammaticality of (54b) contrasts with the grammaticality of (i) where a pause turns the structures into an LD ones:
(i) en cuanto a Juani , loi vi ayer [as of Juan, him see-1stps-past yesterday]
hablando de Juani , loi vi ayer [talking about Juan, him see-1stps-past yesterday]
en cuanto al vinoi , te loi prohibieron con razón [as far as the wine, no wonder you mustn’t take it]
156 Absolute first position makes reference to the fact that LD constituents cannot be embedded although they can be preceeded by topicalizing expressions, as in (54a). Notice also that LD constituents will appear in embedded position in those constructions in Spanish in which a complementizer can be followed by a wh-word:
(i) me pregunto que quién trajo el vino [I wonder that who brought the wine]
[I wonder who brought the wine]
(ii) me pregunto que a Juani quién loi llamó [I wonder that to John, who him called]
[I wonder who called John]
features of the co-referential element (en él), in which the preposition en marks accusative Case, and those of the dislocated element (Juan) with nominative Case (as opposed to the oblique accusative Case of en Juan):157
(56) Juani , estaba pensando en éli en este momento [Juan, be-past-1stps thinking in him in this moment]
[Juan, I was thinking of him at this moment]
On the contrary, in CLLD, there is an obligatory connectivity (Cinque 1990) between the two items, as in (57) where they both show accusative Case features:
(57) a nosotrosi no nosi han dicho nada [to us not us have-3rdpp said nothing]
[they have not said anything to us]
7) In both constructions more than one constituent can be dislocated, with the difference that LD structures require a conjoined phrase:
(58) a. en cuanto a María y a su maridoi + j , es evidente que élj lei pega a ellai
[as for Mary and her husband, is evident that he her hits to her]
b. a Maríai esa película no lei interesa [to Mary this movie not her interest-3rdps]
157 The lack of connectivity in LD structures can be shown by the possibility of having a left- dislocated element that does not agree in Gender and Number with the co-referential element, as in (i):
(i) el ordenadori , yo odio esas máquinas infernalesi
[Mary she is not interested in this movie]
8) The types of constituents involved in each construction vary as well. In LD these are adjuncts to CP and in CLLD they are adjuncts to AgrSP, as in (59a) and (59b):
Both LD and CLLD are base-generated (they are adjuncts), so no movement applies.
9) In both cases there is the possibility of a preceding wh-word as in (60):
(60) a. en cuanto a Juani , ¿qué quiere comer hoy ese chicoi? [as far as John, what want-3rdps eat today?]
b. ¿a Juani qué lei pasa?
[to John what him happen-3rdps?]
[what happens to John?]
[the computer, I hate these hideous machines]
AgrSP (59b)
AgrSP
AgrS VP ...
CLLD
constituent
J
CP (59a)
CP
C IP ...
LD
constituent
Juan
But in the case of LD, as shown in (60a), a long pause is required between the constituent and the wh-word, the LD structure being external to the interrogative sentence.158
10) Constituents from which extraction is not possible are referred to as islands.
Regarding island constraints, LD is insensitive both to strong and weak islands, as in (61):
(61) a. en cuanto a ese trabajoi, [no puedo [aceptar la idea [de que ya loi ha conseguido]]]
[as for that job, [not can-1stps [accept the idea [of that already it have-3rdps got]]]
b. Maríai , [me pregunto [que quién lai ha visto]]
[María, me wonder-1stps that who her have-3rdps seen]
LDs in (61) allow the constituents ese trabajo and María to be coreferential with the clitics lo and la respectively, both located in the lowest clause. In the case of ese trabajo, the corresponding clitic is separated by three barriers from its coreferential LD structure and, therefore, the term strong island is used. Since the NP María in (61b) is separated from its clitic by two barries, we are dealing with a weak island.
As opposed to LD, CLLD is only insensitive to weak islands, as in (62):
(62) a. ?ya te he dicho que [el dineroi [no puedo [aceptar la idea [de que ya loi han conseguido]]]]
[already you have-1stps told that [the money [not can-1stps [accept the idea [of
158 The assumption in the case of CLLD constituents in interrogatives is that there is covert I-to-C movement in Spanish and that this movement is blocked after Spell-Out if any preverbal adjunct is
that already it have-3rdpp got]]]]
[I have already told you that [the money [I can’t [accept the idea [that they have already got it]]]]
(62) b. a esos espíasi [no sé [cómo se puede saber [quién losi traicionó]]]
[to these spies not know-1stps how can-3rdps-impersonal know who them betray-3rdps-past]
In (62a), the CLLD constituent el dinero is more strongly felt to be apart from its corresponding clitic (strong island), while the weak island in (62b) poses no problem for the relationship betwen the constituent a esos espías and its corresponding clitic los.