De De Se by Pranav Anand Submitted to the Department of Linguistics and Philosophy in partial fulfillment of the requirements for the degree of Doctor of Linguistics at the MASSACHUSETTS INSTITUTE OF TECHNOLOGY September 2006 © Massachusetts Institute of Technology 2006. All rights reserved. AA uutthhoorr . .&. .I.. =.:...... . . -........................................... Department of Linguistics and Philosophy June 26, 2006 Certified by ................................-. ...... Irene Heim Professor Thesis Supervisor A ccepted by ........................... ..... Irene Heim Chair of the Linguistics Program MASSACHUSETTS W&MMTE OF TECHNOLOGY ARCHIVES SEP 2 6 2006 LIBRARIES De De Se by Pranav Anand Submitted to the Department of Linguistics and Philosophy on June 26, 2006, in partial fulfillment of the requirements for the degree of Doctor of Linguistics Abstract In this dissertation, I argue against a unitary treatment of individual de se ascription. Based on con- sideration of Yoruba logophors and English dream-report pronouns, I show that one mechanism is best analyzed as binding by an operator, which is sensitive to binding locality requirements. In contrast, I argue that cases of indexical shift (whereby token-reflexive elements such as I and tomorrow may be dependent on the context of an attitude predicate), which do not show local bind- ing effects, are instances of overwriting of elements of the sequence of evaluation. As pronouns that are not obligatorily read de se show neither of the conditions for shifted indexicals nor West- African logophors, I argue that de se readings of these items must arise as special cases of de re ascription. Cross-linguistic instances of anti-logophoricity (i.e., the obligatory non-de se ascription of pronouns in certain contexts) are correspondingly treated as environments imposing a non-de se demand on de re ascription. Finally, I demonstrate that binding and overwriting mechanisms may both be found within the territory of de se long-distance anaphora, based largely on a systematic split in interpretation amongst Mandarin speakers on licensing and interpretative constraints on long-distance ziji. Thesis Supervisor: Irene Heim Title: Professor Acknowledgments The work contained in this dissertation is the result of many intellectual debts I hope may be one day recompensed; till then, these thanks will have to serve as a promissory note. First, I must thank both Andrew Nevins, my collaborator on portions of Chapter 2, and Feng-fan Hsieh, my collaborator on portions of Chapter 3. The systems constructed in Anand and Nevins (2004) and Anand and Hsieh (2005) survive relatively intact in the present work, demonstrating how pivotal my discussions with them have been in my thinking on these topic. I additionally owe much to my advisors, Irene Heim, Danny Fox, and Sabine Iatridou, all of whom have been wonderfully patient over the past five years in letting me present all manner of inchoate ideas, many of which - amended for the better by their comments - have found their way into this dissertation. My deepest thanks in this regard go to Irene, whose ready skepticism has shaped many of the ideas in this work. However, the readiness of faculty to meet with students is a general property at MIT, and I have profited greatly from meetings with many, in particular, Kai von Fintel, David Pesetsky, and Norvin Richards. Kai has the rare sight for where to probe analyses further, both to break them and to push them further; indeed, much of chapters 1 and 3 can be traced to his urging not to give up the de se character of indexical shift as settled. An hour of David's creativity always challenged me to reconsider my most cherished axioms; I only hope I can do the same myself without him there to do the questioning. Norvin, ever-willing to puzzle out a complex set of data with me, inevitably broadened the scope of the question I was asking, or the data I had to consider. In addition, I learned much about being a linguist from my other teachers at MIT: Noam Chomsky, Andrea Gualmini, Morris Halle, Michael Kenstowicz, Alec Marantz, Shigeru Miyagawa, Wayne O'Neil, Donca Steriade, and Ken Wexler. What I will miss most from my graduate student days is, however, the fluidity of discussion that arises between students, how in one breath the conversation can pass from technical logic to car repair. I have enjoyed such multifarious hallway impromptus with a wide cast of characters, including: Marta Abrusan, Asaf Bachrach, Marcelo Ferreira, Jon Gajewski, Sarah Hulsey, Roni Katzir, Nathan Klinedinst, Ivona Kucerova, Eric McCready, Sarah Moss, Bernard Nickel, Connor Quinn, Raj Singh, Tamina Stephenson, Eric Swanson, Michael Wagner, Linnaea Stockall, and Seth Yalcin. Special thanks to Barry Schein, hallway-discussant extraordinaire in many ways. A special pride of place will always be reserved in my heart for the rest of Ling-01: Allison Adler, Justin Fitzpatrick, Valentine Hacquard, Andres Salanova, Shoichi Takahashi, and Maryann Walter, six individuals who have helped me in incalculable ways intellectually and emotionally. It will be difficult to adjust to not seeing each other's faces everyday. I will miss you all tremendously. Graduate life engenders constant self-evaluation and much concomitant guilt; my non-linguist friends were always there to ensure proper perspective. Thanks to the denizens of Rufus - Jake Flemming, Katie Murphy, Eric Rosenbaum, Alex Scammon, and Alec Speigelman - for living with an academic of the obscure. Thanks as well to Mwanga Mtengule, Amy Offner, and Danni Tang for always being willing to listen. And thanks to my fellow Rambax dummers for ensuring that I found space for music. Finally, to my family, who have had to deal with a person who increasingly proclaimed his increasing busyness: thank you for being incredibly supportive these past five years, though, sadly, I fear the trait has stuck. And to my family to be, Nikki and Hobbes: my love and promise to stay longer. Contents 1 The LFs of dream reports and logophoric environments 1.1 Introduction ...................... 1.2 On de se belief .................... 1.2.1 On de se ascription ......... . .. . . 1.2.2 Dere belief .................. 1.3 The Argument From Only . . . . . . . . . . . . . . 1.3.1 The Argument from Only . . . . . . . . . . 1.3.2 The problem: available with ORC violations 1.3.3 Putting the blame on the quantifier ...... 1.3.4 Specifying SELF ............... 1.4 The De Re Blocking Effect .............. 1.4.1 The Oneiric Reference Constraint . . . . . . 1.4.2 A Centering Approach . ........... 1.4.3 A Binding Economy Approach ........ 1.4.4 Blocking in Yoruba .............. 1.4.5 On Anti-logophoricity . ........... 1.4.6 Summing Up ................. 1.5 Reflections on the theory . ............ . 2 On the shifting of indexicals 65 2.1 Introduction . .................. . . . . . . . . . . . . . . . . . . . 65 2.2 The Kaplanian view of indexicality ...... . ................... 68 2.2.1 Context and Content .......... ................... 68 2.2.2 Direct Reference ............ ................... 72 2.3 Indexical Shifting ................ ................... 74 2.3.1 Some Initial Facts ........... ................... 75 2.3.2 Quotational Theories of Shifting . . . . . . . . . . . . . . . . . . . . . . . 80 2.4 Two ways to shifting .............. ................... 89 2.4.1 A pronoun-centered theory ...... . ................... 89 2.4.2 Binding by Attitude Verbs ...... . ................... 94 2.5 Constraints on Multiple Shifting Indexicals . . . . . . . . . . . . . . . . . . . . . 98 2.5.1 Monoclausal Cases: SHIFT TOGETHER . . . . . . . . . . . . . . . . . . . 99 2.5.2 The case of multiple embedding: No INTERVENING BINDER . ...... 103 2.6 Analysis ............ ....... ................... 107 2.6.1 Fixing the verbal-quantifiers approach ................... . 107 2.6.2 The proposal ..................... ............. .108 2.6.3 On ambiguity theories .............. ............. 110 2.6.4 Recasting this with multi-indexing ................... ... 111 2.6.5 Obviation effects ................ .............. .. 112 2.7 Some Problematic Data ................................ 115 2.7.1 Slave violates No INTERVENING BINDER . ................ 115 2.7.2 Catalan Sign Language does not obey SHIFT TOGETHER . ......... 116 2.8 Conclusion .. .... ... . .. .. .. . .. . .... .. . .. .. ....... 118 3 On the landscape of long-distance anaphora 119 3.1 Introduction ................... ................... 119 3.2 The case for ziji as a shifting indexical ........................ 120 3.2.1 A ziji Primer .................... ............. 120 3.2.2 Assimilating zii ................... ............ 127 3.3 Two M andarin dialects ................................ 134 3.3.1 The Initial Puzzle ............... . .............. 134 3.3.2 Constraints on "logophoric interpretations" ............. . . . . 137 3.4 LDR, indexicals, and rationale clauses ........................ 140 3.4.1 Some properties of rationale clauses ................... .. 141 3.4.2 On the interpretive differences between lai- and qu-headed clauses ..... 145 3.4.3 The Analysis: Disjoint reference as binding competition .......... 147 3.4.4 The qu Blocking Effect ............................ 149 3.4.5 Summing Up .................... ............. 153 3.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 Chapter 1 The LFs of dream reports and logophoric environments 1.1 Introduction Clubs can be strange places: with their glittering disco balls, dim mood lighting, and thronging masses it is often not too impossible to see things that are not there, or mistake one thing for another. Indeed, given an artfully disguised mirror at another end of the room and a suitable level of disorientation, you might even mistake your reflection for someone else. Such, borrowing a scenario of Kaplan (1989, p. 533), might be the case of John, who first mistaking his reflection in the mirror at the club for an actual person, thinks, "His pants are on fire!" and then, upon smelling smoke and feeling some unexpected heat around his ankles, thinks more alarmingly, "My pants are on fire!" The latter belief Lewis (1979) rightly dubbed de se, since it is inextricably about the first-person self. While the existence of de se thoughts seems innocent enough pretheoretically, Perry (1979) noted that it was a problem for the propositional doctrine of mental attitude inherited from Frege, since the thoughts are propositionally equivalent. This, in essence, is the de se attitude puzzle: How do we suitably revise our theory of mental attitudes to capture the first-personal nature of de se? The de se puzzle has attracted much attention in philosophy of mind literature (REFS), some of which I will review below. The focus of this dissertation will not, however, be on de se attitudes, but on de se attitude ascriptions - our predications of various mental states to other sentient individuals. As Chierchia (1989) first noted, there are items in natural language that must be interpreted as unambiguously de se in cases of ascription; Chierchia's own example comes from Italian:' (1) S1: Pavarotti is listening to himself singing La donna e mobile and is impressed by his own skill. He thinks, "I have to admit it: I really am a genius!" 'A word on example notation. I provide scenarios in sans-serif font to distinguish them from the target sentence, whose felicity I am trying to establish for the given scenarios. Where I mark felicity judgments (as in the example below), VS means the sentence is felicitous for secnario S, while # S indicates infelicity; when I leave the example unmarked, it indicates that the example is felicitous for the given scenario. S : Pavarotti is listening to a performer singing La donna e mobile, and is impressed by his artistry. 2 He thinks, "This performer is a genius! I could learn a lot from him." Unbeknownst to him, he is the performer he is listening to. a. Pavarotti crede di PRO essere un genio Pavarotti believe-PRES COMP PRO to-be a genius 'Pavarotti believes, "I am a genius." ' [/S , #S ] 1 2 b. Pavarotti crede che gli e un genio Pavarotti believe-PRES that he be-PRES a genius 'Pavarotti believes that he is a genius. ' [/S1, /S21 Thus, the sentence is only acceptable as a report of Pavarotti's de se belief "I am a genius." This is not the case with the finite form, which can ascribe either a de se belief or a non-de se belief to Pavarotti. Thus, PRO is obligatorily interpreted de se, in contrast to normal pronouns, which seem to have optionality. Since Chierchia's discovery of PRO's de se requirement, many other obligatory de se elements have been discovered, including West African logophors, shifted indexicals, and long-distance anaphora. The existence of these items raises three important questions: (2) DEDICATED LFS: Insofar as de se elements contrast with pronouns in optionality of de se ascription, is this evidence for dedicated LFs for de se elements? (3) How MANY de se LFS: If so, exactly how many de se LFs are there? (4) de se PRONOUNS: What ramifications do these items have for pronouns themselves - are the de se compatible pronouns an instance of vagueness or ambiguity of logical form? The aim of this dissertation is to try and answer these questions. I will claim that there are three ways to de se readings. The first way, which I will argue is true universally for pronouns, is that de se is a species of de re; this is what happens for gli in (lb). However, I will argue that there are, in addition to the de re default case, two dedicated de se logical forms. The first, following Chierchia (1989) will be a syntactic representation, where the de se element is bound by an operator within the scope of the attitude verb: (5) att-holder Vat [OPi ... xi] t I will argue that this syntactic condition is diagnosable by an intervention effect that arises when a a de se anaphor is c-commanded by a de re counterpart: (6) De Re BLOCKING EFFECT No (syntactic) de se anaphor can be c-commanded by a de re counterpart. In this chapter, I will show that the De Re Blocking Effect holds both for English dream report pronouns and Yoruba logophors. As an illustration of the phenomenon, consider Lakoff's famous dream-report sentence: (7) I dreamed I was Brigitte Bardot and I kissed me. (Lakoff, 1972)