Perché Rizzi is right * - PDF

Perché Rizzi is right * CARLO CECCHETTO University of Milan-Bicocca CATERINA DONATI Sapienza University of Rome In this squib we elaborate on the proposal

Please download to get full document.

View again

of 9
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.


Publish on:

Views: 6 | Pages: 9

Extension: PDF | Download: 0

Perché Rizzi is right * CARLO CECCHETTO University of Milan-Bicocca CATERINA DONATI Sapienza University of Rome In this squib we elaborate on the proposal that perché( why ) does not leave a trace inside the IP and argue that this peculiarity explains the absence of free relatives introduced by why as well as some order restrictions concerning reason clauses. We also discuss why many languages, but not Italian, distinguish between why and because. 1. Introduction. Starting from a seminal observation by Rizzi (1990), several authors have argued that the adjunct why (and its equivalents in other languages) is externally merged in the left periphery of the clause (e.g., Hornstein 1995, Ko 2005, Rizzi 2001, Stepanov and Tsai 2008, Thornton 2008), or that it moves locally within the left periphery, crucially without leaving a trace within IP (Shlonsky and Soare 2011). The aim of this squib is to discuss a possible consequence of this observation concerning reason clauses. Starting from a theory of labeling and movement (section 2) and a free relative analysis of adverbial clauses (section 3), we will briefly review the original facts pointed out by Rizzi (section 4) and derive from them a) the non intersective reading of perché-clauses as opposed to other WHadverbial clauses in Italian (section 5); b) the non-existence of free relatives introduced by why and its equivalent in many languages (section 6); some order restrictions concerning perché and other reason clauses (section 7). 2. Labels and probes: free relatives The starting point is the general idea, developed in Cecchetto and Donati (2010) and Donati and Cecchetto (2011), that there is a crucial relation between syntactic probes and syntactic labels. They propose that the core results of X-bar theory can be derived from the definition of label in (1) and from the labeling algorithm in (2). (1) When two objects α and β are merged, a subset of the features of either α or β becomes the label of the syntactic object {α, β}. A label: (i) can trigger further computation (ii) is visible from outside the syntactic object {α, β} for selection. * Although in this paper we argue that Rizzi is right, we do not think that this is always the case. Rizzi can be wrong in many different ways, but always insightfully. We wish we could always be wrong the way he is! Buon compleanno, Luigi! 2012CarloCecchetto&CaterinaDonati InternetcelebrationforLuigiRizzi s60 th birthday Perché RizziisrightCarloCecchetto&CaterinaDonati (2) The label of a syntactic object {α, β} is the feature(s) which act(s) as a Probe of the merging operation creating {α, β} The algorithm in (2) says that the label of any merge output is always the feature asymmetrically triggering the Merging operation. This simple algorithm immediately derives one empirical generalization that any version of phrase structure theory must account for: namely, that the target of movement (a Probe) typically projects. However, (2) can also capture the other fact that any version of phrase structure theory must derive, namely that a lexical item (a head ) projects when it is merged with an XP. The algorithm in (2) can do so if it is assumed, following Chomsky (2008), that every lexical item is endowed with a feature, the edge feature, which forces it to merge with other material (this feature is ultimately responsible for the fact that we speak by uttering phrases, clauses and sentences rather than isolated words). If this is assumed, any time a lexical item is merged, it qualifies as a Probe by virtue of its edge feature. This means that a lexical item, being a Probe by definition, always activates the algorithm in (2) and its categorial feature can provide the label. We argued elsewhere (see also Donati 2006) that the double life of free relatives immediately derives from the labeling algorithm in (2). (3) a. I wonder what you read b. I read what you read In (3), a WH- lexical item, what, is internally merged to a Probing C. The Probing Algorithm in (2) correctly predicts that there should be a labeling conflict here. If the LI provides the label, the structure ends up being a DP, i.e. a free relative; if the probing C provides the label, the structure is an (interrogative) clause: as a result, the structure is systematically ambiguous, as shown by its compatibility both with verbs selecting for nominal complements (e.g. read in 3b) and with verbs selecting for clauses, as in (3a). No ambiguity arises when a phrase is WH-moved: in (4) what book is a phrase, not a lexical item, so it does not qualify as a Probe, and only the target C is bound to project. (4) can only be an (indirect) interrogative clause. (4) What book you read a. I wonder what book you read b. *I read what book you read. 3. Adverbial clauses Interestingly, many adverbial clauses can be analyzed as free relatives. This is clearly the case for when, where and how clauses: they are introduced by a bare WH-word and display a systematic ambiguity between an interrogative reading (the probe C provides the label: 5a; 6a; 7a) and an adverbial use. In the latter case they display the typical reading characteristic of relativization: to illustrate, the when clause (7b) has the same interpretation as the explicit relative clause in the moment in which you will leave and the sentence talks about a set of moments which is the intersection of the moments when you are sad and the moments when you leave. For concreteness, we will say that the when clause in (7b), as well as the how clause in (5b) and the where clause in (6b), have an intersective meaning. 2 InternetcelebrationforLuigiRizzi s60 th birthdayciscl,siena This is reminiscent of what happens in a full relative like (a) boy who likes you, which identifies the intersection of the set of boys and of set of people who like you. Under our analysis, the free relative reading is due to the combination of two ingredients: the raising of the WH-element, leaving a gap (or a copy) in the embedded clause, and the labeling of the structure by the WH-element itself. Here the category will be of course prepositional or adverbial, not nominal as in standard free relatives like (3b) 1. (5) How you prepared your job talk a. I wonder how you prepared your job talk b. I will prepare my interview how you prepared your job talk = I will prepare my interview in the same ways as you prepared your job talk (6) Where he will tell me to go a. I wonder where he will tell me to go b. I will go where he will tell me to go = I will go in the places to which you will tell me to go (7) When you will leave a. I wonder when you will leave b. I will be sad when you will leave = I will be sad in the moment(s) in which you will leave It has been proposed that also if-clauses can be analyzed as free relatives. While if may not be a plain WH-word in English, it does have an interrogative use, as shown by the fact that in some varieties you can say things like I wonder if (in fact, Kayne 1991 has argued that the conditional if and the interrogative if are one and the same element). Furthermore, as discussed by Bhatt and Pancheva (2006), that the complementizer introducing the protasis is a WH-word is pretty clear in many Romance varieties (where the equivalent of if is the canonical complementizer of embedded yes/no question), in German (where the equivalent of if is wenn, which also appears in when clauses) and in Bulgarian (which also uses an interrogative complementizer to form a conditional clauses). Finally, if-clauses have an intersective meaning, too. From an interpretive point of view, (8) is not fundamentally different from the correspondent when-clause, namely sentence (7b) above. After all, (8), like (7b), can be roughly paraphrased by using a nominal + relative clause: (8) I will be sad if you will leave = I will be sad in in the situations/possible worlds in which you will leave Starting from this type of observation, Bhatt and Pancheva (2006) propose that ifclauses are just another case of free relative, where a WH-word (or a null operator) is a binder of a possible world variable. So, while a canonical free relative as what John bought is interpreted as the plural definite description ıx 1 Caponigro and Pearl (2009) discuss some evidence that when, where and how are indeed nominal, but merged as a complements of a silent preposition. This is not directly relevant here. 3 Perché RizziisrightCarloCecchetto&CaterinaDonati [John bought x], the if-clause if he talks the president is interpreted as the plural definite description ıw [he talks to the president in w]. Haegeman (2010) supports the analysis that posits an analogy between temporal clauses and if-clauses by offering several arguments in a cartographic framework. An interesting exception to this pattern of adverbial clauses is why-clauses, or their Italian counterpart perché-clauses illustrated in (9). (9) Perché parti why 2 you leave a. Mi chiedo perché parti (I) wonder why (you) leave b. Sono triste perché parti ( I) am sad why (you) leave Perché-adverbial clauses crucially differ in interpretation from when-clauses, where-clauses, how-clauses and if-clauses (Caponigro 2003). In particular they do not display the intersective reading which is the defining feature of relativization and which we have just seen displayed by the other WH-adverbial clauses. To clarify, (9) does not mean (10b), but rather (10a). (10) a. I am sad because you are leaving b. * I am sad for the same reason why you are leaving Why so? Notice that in English (and French and many other languages) the anomaly of why-clauses is even more severe: essentially, why-clauses can only have an interrogative reading: (11) 3. (11) Why you are leaving a. I wonder why you are leaving b. *I am sad why you are leaving In the remaining part of the paper we will address these two anomalies one at a time: we will first focus on why perché in Italian can introduce an adverbial clause, but not with the intersective reading. Then, we will ask why why can only be interrogative in English and other languages. 4. The case of Italian perché: no intersective reading Remember that under the approach briefly sketched above, relativization is the result of two components: a raising operation, leaving a gap in the embedded structure related to a position in the root clause; and a relabeling outcome, by which the moved element labels the structure turning it from a simple CP into a different type of object (a DP in argumental free relatives, a PP or an AdvP in 2 As is obvious, perché translates both why and because. We uniformly use the gloss why to stress the fact that concerns us in this paper, namely that perché is a WH-word. 3 Caponigro (2003) discusses sentences like (i) as the only cases in which why does not have an interrogative use. Arguably these cases should be analyzed as full relative clauses headed by a null N, as suggested by the periphrasis in (ii). (i) This is [why he never laughs] (ii) This is the reason why he never laughs 4 InternetcelebrationforLuigiRizzi s60 th birthdayciscl,siena adverbial ones). Now, starting from an observation by Rizzi (1990), many authors have argued that why (and its equivalents in other languages) is externally merged in the left periphery of the clause (e.g., Hornstein 1995, Ko 2005, Rizzi 2001, Stepanov and Tsai 2008, Thornton 2008), or that it moves locally within the left periphery, crucially without leaving a trace within the IP (Shlonsky and Soare 2011). 4 Let us briefly review here the main arguments put forward by Rizzi. Rizzi (1990) notices that, unlike other adjuncts, why is not sensitive to negation in the clause with which it is construed. This is illustrated in the contrast between how and why in (12) taken from Shlonsky and Soare (2011). (12) a. Why didn t Geraldine fix her bike? b. *How didn t Geraldine fix her bike? If the ungrammaticality of (12b) follows from a Relativized Minimality violation induced by the intervention of negation (Rizzi 1990), then it follows that there is no chain crossing negation in (12a). Another piece of evidence comes from WHin situ. In French, where WH-is situ is a possible interrogative strategy (13), pourquoi is never allowed to sit in a postverbal position (14). (13) a. Tu as rencontré qui you have met who b. Tu as fais quoi? you have done what c. Tu es parti où/quand? you have left where/when (14) *Tu es venu pourquoi 5? you are come why Another argument concerns direct WH-questions in Italian. While subject inversion is obligatory with the other adverbial WH-elements, such as dove (where), come (how), quando (when), it is not with perché (why). The following examples are adapted from Rizzi (15) a. * Dove Gianni è andato? Where Gianni went? a Dove è andato Gianni? Where went Gianni? b * Come Gianni è partito? How Gianni left? b Come è partito Gianni? How left Gianni? 4 One can ask what makes why special. Maybe our ontology contains entities like objects, places, times/situations, manners/ways but not reasons. So, one can quantify over the former entities but not over reasons. We do not address the general question of what makes why special in this squib, though. 5 We are slightly simplifying things since purpose pour quoi ( what for ) is acceptable in situ. (i) Tu es parti pour quoi? you have left for what? This is however irrelevant for us, since we are concerned only with reason adverbials. 5 Perché RizziisrightCarloCecchetto&CaterinaDonati (15) c. *Quando Gianni è partito? When Gianni left? c Quando è partito Gianni? when left Gianni? (16) a Perché Gianni è venuto? Why Gianni came? b. Perché è venuto Gianni? Why came Gianni? This contrast is explained by Rizzi assuming that WH-arguments and WHadverbials other than why can only meet the WH-criterion by being moved to the left periphery, hence triggering inversion, while why does not trigger any inversion because it is directly merged in the left periphery. Now comes the first main remark we want to introduce in this squib: if no movement is involved when perché is merged (or a very short one, crucially not creating a gap within the IP, as Shlonsky and Soare 2011 argue) the Italian anomaly of reason adverbial clauses is straightforwardly derived. As we have briefly shown in the preceding section, the intersective reading proper to relativization requires a raising operation transformationally relating a position in the IP with the head of the relative construction: perché has no access to this kind of derivation, if it does not move (or does not move long enough). So, no intersective reading arises. Summarizing, the absence of the intersective reading with perché adverbial clauses is just what one expects under the hypothesis that perché leaves no trace in the IP area. 5. The case of English why and the like: adverbial reading Remember however that in our approach relatives are made of two components, not necessarily related to one another: a raising component and a labeling component. Putting aside the raising component, which is not available to perché and its kin, since they do not move, we would nevertheless expect the labeling component to be accessible: perché is a head and the same for its kin in other languages, like why. Even if the probe of the merging operation is C there should be a labeling conflict, and the adverbial should be able to provide the label. Now, Italian is well behaved in this respect: perché does indeed provide the label in (10), which is an adverbial adjunct as predicted. The problem is that in other languages the equivalents of perché cannot introduce reason clauses. Clearly, this must have to do with the fact that there is an alternative (say, because) which is lexically specified for this adverbial function. One can imagine why such lexically specified alternative to why may evolve in a certain language. For example, this avoids ambiguity (see below for a case in which ambiguity impacts on the distribution of adverbial perché in Italian). Still, the question is: why does only why have a non-interrogative alternative among all WH-elements? In other words, why is ambiguity avoided only with why and not, say, with what, who, when, where, how, if? We have a good answer to this question. In order to convey the intersective meaning, adverbial when clauses, where clauses and other free relatives need to form a chain, but a chain is typically triggered by the WH-feature. This is why you do not have the non-wh 6 InternetcelebrationforLuigiRizzi s60 th birthdayciscl,siena counterpart of when, how etc: without the WH-feature there would be no movement, hence no relativization. So, the intersective reading these WH-words convey might not arise. Notice that if we are on the right track, we expect that the Italian strategy may not be completely out even in languages that do not normally use the WH-word corresponding to why in reason clauses. For example, there might be overuses of why in reason clauses up to a certain stage of acquisition, or maybe slips of the tongue or other performance errors where why replaces because. This is something that needs to be checked. Some weak evidence going in this direction is that in text messages, French youngsters use an underspecified sign ( pk ) to mark both pourquoi and parce que. 6. Some indirect evidence: ordering restrictions We expect other differences to follow from the fact that only in Italian a WHword can introduce a reason clause. This is indeed the case. In Italian, perché-clauses must follow the main clause (cf. 17a and 17b), while reason clauses introduced by a specialized non-wh word like poiché ( since ) can precede the main clause (cf. 17c): (17) a. *Perché parti domani sono triste why (you) leave tomorrow (I) am sad b. Sono triste perché parti domani (I) am sad why (you) leave tomorrow c. Poiché parti domani, sono triste Since (you) leave tomorrow (I) am sad Similarly, in French and English the reason clause, which is not introduced by a WH-expression, can precede the matrix clause 6. (18) a. Because I worked fast, I finished early b. Parce qu elle est riche, Anne a été épousée par Jean It is tempting to relate the impossibility of fronting perché-clauses to the fact that perché is a WH-word. We conclude this squib by offering some speculations in this direction. As a matter of fact, the sequence perché parti domani is temporary ambiguous between an interrogative interpretation ( why interpretation) and a reason clause interpretation ( because interpretation). Suppose that the interrogative interpretation is preferred, say because interrogative perché is more 6 This does not mean that the positioning of reason clauses is entirely free in English or French. Various factors play a role at the discourse level. One example is reason clauses introducing a premise to the main assertion rather than introducing the cause of the event in the matrix clause. Premise reason clauses cannot naturally precede the matrix clause, as shown by the awkwardness of (i) and (ii). See Kanetani (2007) and Hancock (2000) for some discussion in a non-generative framework. (i) * Because the ground is wet, it has rained. (ii) *Parce que sa voiture est devant la porte, Paul est là Because his car is in front of the door Paul is there 7 Perché RizziisrightCarloCecchetto&CaterinaDonati frequent. If so, (17a) would be weird because of a sort of garden path effect. The wrong structural analysis is chosen and this comes with a cost. Although this explanation may be on the right track, it is definitely too simpleminded. An obvious objection is that the same reasoning should apply to other reasons clauses introduced by a WH-word. For example, the sequence quando parti introduces a temporary ambiguity, but (19b) is fully natural. (19) a. Quando parti? When (you) leave? b. Quando parti sono triste When (you) leave (I) am sad What would be the difference here? Clearly, it is desirable to reduce any diff
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks