The position of se

Embedded yes/no questions are introduced by se (“if”) in Italian. This element has certain positional properties in common with declarative che. For instance both are compatible with a focussed phrase which must follow them (in Italian the left peripheral focus position is restricted to contrastive focus, so that in the relevant examples we will always overtly express a contrast; the focussed phrase is capitalized):

(6)a   Credo che QUESTO avreste dovuto dirgli (non qualcos’altro)

          ‘I believe that THIS you should have said to him, not something else’

     b * Credo QUESTO che avreste dovuto dirgli (non qualcos’altro)

           ‘I believe THIS that you should have said to him, not something else’

          

(7)a   Mi domando se QUESTO gli volessero dire (non qualcos’altro)

          ‘I wonder if THIS they wanted to say to him, not something else’

         

    b  * Mi domando QUESTO se gli volessero dire (non qualcos’altro)

           ‘I wonder THIS if they wanted to say to him, not something else’

On the other hand, se can be preceded and followed by a topic, while che can only be followed by a topic, as we have seen:

(8)a Credo che a Gianni, avrebbero dovuto dirgli la verità

       ‘I believe that to Gianni, they should have said the truth to him’

     b * Credo, a Gianni, che avrebbero dovuto dirgli la verità

           ‘I believe, to Gianni, that they should have said the truth to him’

(9)a Non so se, a Gianni, avrebbero potuto dirgli la verità

        ‘I don’t know if to Gianni, they could have said the truth’

     b Non so, a Gianni, se avrebbero potuto dirgli la verità

       ‘I don’t know, to Gianni, if they could have said the truth’

     c Mi domando se questi problemi, potremo mai affrontarli

        ‘I wonder if these problems, we will ever be able to address them’

     d Mi domando, questi problemi, se potremo mai affrontarli

        ‘I wonder, these problems, if we will ever be able to address them’

This suggests that se occupies a position distinct from, and lower than, the one occupied by che, a position which is necessarily higher than FOC, but can be preceded by a topic. So, if che expresses FORCE, we should postulate a distinct position, call it  INT(errogative) for se, to be inserted in the sequence of positions (7) in the following way (see Aboh (1998) for independent evidence in favor of an interrogative position distinct from the landing site of Wh movement in Gungbe):

(10)    FORCE    (TOP*)   INT   (TOP*)   FOC   (TOP*)   FIN  IP

A question that arises at this point, given the distinct positional properties of che and se, is whether we should now understand Force in (10) as specialized for declarative force, with interrogative force expressed by the lower INT position. There are reasons to reject this option. First of all, the presence of a topic to the left of INT in examples like (9)b shows that additional clausal structure is involved above INT, minimally a Topic Phrase hosting the topic in its specifier (see Rizzi (1997) for detailed evidence against an adjunction approach to topic-comment configurations); moreover, it is implausible that the proposition may be closed upward by a Topic Phrase because of  selectional reasons: the main verb selects for an indirect question, not for a clause with a topic. So, there must be a (phonetically null) Force head whose  projection closes the structure upward in  (9)b, thus locally meeting the selectional  requirement of the higher verb. A second reason for  postulating a Force position higher than INT in yes/no questions is that in a closely related language, Spanish, some embedded questions overtly express the force head in cooccurrence with INT by  allowing  the que si (“that if”) sequence in this fixed order (examples from Plann (1982:300), Suner (1994:349)):

(11)a   Maria decia / preguntaba que si no debiéramos dejarlas en paz

            ‘Maria was saying /asking that if we shouldn’t leave them in peace’

       b    Me preguntaron (que) si  tus amigos ya te visitaron en Granada

              ‘They asked me that if your friends had already visited you in Granada’

In such cases, it is transparent that two distinct positions are involved. See McCloskey (1992) for an analysis of the fine semantic properties of the subclass of indirect questions allowing the cooccurence of que and si or a wh element. (The other subclass of indirect questions selected by such verbs as “find out”, “discover”, etc., not allowing initial que, nor an initial topic, may perhaps be analyzed as not involving the Force layer at all, a direct translation into our system of McCloskey’s CP recursion analysis). Languages like Dutch permit the sequence in the opposite order of dat (“if that”). Presumably here “that” expresses a position different from and lower than Force, as in the many Romance and Germanic varieties allowing for the sequence Wh that. 

Going back to the expression of (10) in Italian, the three topic positions following FORCE, INT and FOC, respectively, can be simultaneously realized in examples like the following, which sounds somewhat cumbersome but not grammatically degraded:

(12)  Mi domando, a Gianni, se, ieri, QUESTO, alla fine della riunione, avremmo

          potuto dirgli (non qualcos’altro)

          ‘I wonder, to Gianni, if, yesterday, THIS, at the end of the meeting, we could     

            have said to him (not something else)’

The structural layer whose head is se is clearly distinct from the position occupied by the wh elements in main question: we have seen that se can cooccur with a lower focus (as in (7)a), whereas wh elements in main questions cannot cooccur with a focus, in either order:

(13)a  * A chi QUESTO hanno detto (non qualcos’altro)?

              ‘To whom THIS they said (not something else)?

       b * QUESTO a chi hanno detto  (non qualcos’altro)?

              ‘THIS to whom  they said (not something else)?’

       c * A GIANNI che cosa hanno detto (non a Piero)?

              ‘TO GIANNI what they said (not to Piero)?

       d * Che cosa A GIANNI hanno detto (non a Piero)?

              ‘What TO GIANNI they have said (not to Piero)?

This incompatibility is interpreted in Rizzi (1997) as showing that wh elements in main questions move to the specifier of the FOC head, therefore they compete with focussed constituents for this position. Clearly, (7) shows that se occupies a position higher than FOC.

What about Wh elements in embedded clauses? The absolute incompatibility with focus by and large disappears, even though the judgments become somewhat murky and graded:

(14)a   *? Mi domando a chi QUESTO abbiano detto (non qualcos’altro)

                ‘I wonder to whom THIS they have said (not somethin else)’

      b    *? Mi domando QUESTO a chi abbiano detto (non qualcos’altro)

                 ‘I wonder THIS to whom they have said (not something else)

      c         Mi domando A GIANNI che cosa abbiano detto (non a Piero)

                 ‘I wonder TO GIANNI what they have said (not to Piero)

      d    *? Mi domando che cosa A GIANNI abbiano detto (non a Piero)

                 ‘I wonder what TO GIANNI they have said (not to Piero)

When the focalized constituent is the direct object and the wh element is a PP, both orders appear to be degraded (as in (14)ab), whereas when the wh element is the direct object and the focalized  constituent a PP, the order FOC Wh is significantly more acceptable (as in (14)c). The possible cooccurrence shown by (14)c clearly suggests that wh elements in embedded questions are not forced to move to the Spec of FOC, contrary to main questions; therefore, there must be a position lower than FOC available to wh elements in embedded questions. If this position is necessarily lower than FOC, we expect both (14)a and (14)d to be excluded, as they express the wrong order. What remains to be explained is why (14)b, but not (14)c, is well-formed.  The contrast between (14)b-c is reminiscent of the crossing constraint (Pesetsky (1982) and references cited there): the two A’ dependencies are crossed in b and nested in c. Whether or not the crossing explanation is correct, and whatever the exact nature of the position filled by the wh element in (14)c, the well-formedness of this example with this particular order shows that the position occupied by se is distinct from, and higher than, the position occupied by wh elements in embedded questions like (14)c (compare this example with (7)b).

As the position occupied by se is higher than the position occupied by FOC (see (7)), and the FOC position is higher than the position occupied by Wh in embedded questions (14)c, we conclude, by transitivity, that the position of se is higher than the position of embedded Wh elements. So, whatever the exact position of the latter (noted here as Wh), we have the following ordering in embedded clauses:

(15)     … Force… INT … FOC … Wh …      (embedded clauses)