More variation in island repair: the clausal/non-clausal island distinction Matthew Barros, Patrick Elliott and Gary Thoms1 Chicago Linguistics Society 49 19th April 2013

Introduction

1.

• The issue: apparent island violations under ellipsis. Sluicing, relative clause island (Ross 1969): (1)

They want to hire someone who speaks a Balkan language, but I don’t know which one. a. ... which one they want to hire someone who speaks t b. *Which language do they want to hire someone who speaks t?

• Some approaches: – Repair by ellipsis: Ross (1969), Chomsky (1972), Lasnik (2001), Merchant (2008) – Island evasion: Fukaya (2007), Abels (2011), Barros (to appear) – Mixture of both, some islands repaired, others evaded: Merchant (2001) – No movement, no matter: Chung et al. (1995), Barker (to appear). • Repair: island-escaping leaves “starred traces” which lead to ungrammaticality unless deleted at PF. • Evasion: no island in the ellipsis site. One possibility for (1): a “short source,” i.e. non-isomorphism. (2)

They want to hire someonei who speaks a Balkan language, but I don’t know which one. Merchant (2001) ... which one hei speaks t / hei should speak t

• No movement: whP base-generated in surface position, no movement out of island. • Important result of much work since Merchant (2001): island repair effects are variable. Dimensions of variation reported in literature: I: sluicing (repair) vs fragments (no repair): Merchant (2004) (3)

A: Does Abby speak the same Balkan language that B EN speaks? B: *No, C HARLIE she speaks the same Balkan language that t speaks. (cf. (1))

1 Barros

(Rutgers): [email protected]; Elliot (Edinburgh): [email protected].

[email protected]; Thoms (Edinburgh):

2

Island repair: clausal vs non-clausal islands

II Dutch fragments (repair) vs English fragments: Temmerman (2013). A simple case:2 (4)

a.

b.

Zijn ze weggegaan omdat jij BEN gekust hebt? are they left because you BEN kissed have “Did they leave because you kissed B EN? Nee, K AREL No Charlie “No, C HARLIE they left because you kissed Jeroen van Craenenbroeck p.c.

III: contrastive (no repair) vs. non-contrastive ellipsis (repair): Merchant (2008), Griffiths and Lipt´ak (2012) (5)

*They want to hire someone who speaks F RENCH, but I don’t know what OTHER LANGUAGE they want to hire someone who speaks. (cf. (1)) • First two generalisations explained by Merchant and Temmerman in terms of cartography of CP; third explained by Griffiths and Lipt´ak (2012) in terms of parallelism. Neither approach generalises to cover all the facts as reported. • Our goal today: to identify another (apparent) generalisation about island repair, namely a distinction between different kinds of islands: clausal islands display repair, non-clausal islands don’t. • We’ll explain this in terms of a ragtag bag of different evasion strategies, showing that clauses are more “permeable” because they allow for more such evasion strategies. We will thus argue that evasion is pervasive and that repair is not real.

2.

Contrastive fragments

2.1 Revisiting English contrastive fragments • Merchant (2004): fragment answers derived by movement to left periphery then clausal ellipsis. Lots of good evidence. • RC island: respected. Note the full answer is good. (6)

A: Does Abby speak the same Balkan language that B EN speaks? B: *No, C HARLIE she speaks the same Balkan language that t speaks. (= (3))

(7)

B0 : No, she speaks the same Balkan language that C HARLIE speaks. B00 : No, the same one that C HARLIE speaks.

• Important point: “island pied-piping” is also good here. May be described as an evasion strategy. • Some more from Merchant: adjunct island, left branch island . 2 Temmerman’s

key data all comes from embedded contrastive fragments or two different types. We focus on the case of a matrix fragment here, for simplicity’s sake. Our example is modelled closely after her adjunct island example (35). We come back to the exact status of the data shortly.

Island repair: clausal vs non-clausal islands

(8)

A: Did Ben leave the party because A BBY wouldn’t dance with him?

(9)

A: Did Abby vote for a G REEN party?

3

B: *No, B ETH.

B: *No, R EFORM.

• But a common objection:3 there are many apparent island-escaping fragments that are a lot better or even fully grammatical. For instance, another RC island, modelled on (1): (10)

A: Did they hire someone who works on F RENCH (last year)?

B: No, G ERMAN.

• Similarly because-islands with object extraction (11): only a little degraded, and a lot better than (8). (11)

A: Did they leave because you offended M ARY?

B: ?No, S ARAH.

• Why would there be a subject/object asymmetry? We come back to this soon but note in passing that subj/obj asymmetries wrt locality are common... • We have found that the choice of island matters a lot. Some don’t seem to repair much at all; for instance, definite DP islands, which are not always the strongest islands but which do resist extraction with complex possessives (see Davies and Dubinsky (2003) for much relevant discussion). (12)

A: Did you give Mary’s picture of P RINCE to her?

B: ?*No, E LVIS.

(13)

A: Did you steal Mary’s picture of P RINCE?

B: ??No, E LVIS.

(14)

*?Which popstar did you give her picture of to her?

• To back this up, we conducted a small informal survey of 12 linguists, testing a bunch of different configurations. Rating examples on a 1-7 scale, 1=*, 7= grammatical. – For (10) (rel. clause), almost all speakers reported it as fully grammatical, av. score 6.7. – For (11) (because-clause + obj.), all scored between 4 and 6, av. 5.1. – For (12), (because-clause + subj.) majority scored around 3, av. 3.1. – For a variant of Merchant’s (6) (rel. clause), majority scored between 1 and 3, av. 2.7 – For a variant of Merchant’s (9) (left-branch), all but one scored 1, av. 1.1 • So both our data and Merchant’s data hold up; some islands show repair, others don’t. • Tentative generalisation: islands made up of larger domains – finite complement adjuncts, some relative clauses – show repair more readily. Islands that are identifiable as smaller domains – left branch islands and definite DPs – show no obvious signs of repair. 3 We

know of only one semi-written source that has disputed the English data (an unpublished ms. by Polly Jacobson, the content of which has been conveyed to us), but have bumped into resistance in our own presentation of the facts in related work over the years.

4

Island repair: clausal vs non-clausal islands

2.2 Dutch fragments reconsidered • Temmerman (2013, p.256) reports that Dutch shows repair even with contrastive fragments. (15)

a.

b.

(16)

a.

b.

Is Jack gekomen omdat hij M ARIN wil versieren? is Jack come because he Marin wants seduce “Has Jack come because he wants to seduce M ARIN?” Nee, ik had gedacht LYNN no I had thought Lynn “No, I had thought LYNN” Dutch, because-island Willen ze iemand aannemen die G RIEKS spreekt? want they someone hire that Greek speaks DO they wan to hire someone who speaks G REEK? Nee, ik sou denken A LBANEES. no I would think Albanian “No, I would think A LBANIAN” Dutch, relative clause island

• Temmerman’s data embedded fragment answers, but extends to simpler test case of matrix responses.4 • Temmerman describes this as a syntactic difference between English and Dutch, with explanation terms of different CP layers in different languages. The claim of difference is based upon Merchant’s 2004 observation that English does not show repair. But this is not a straightforward comparison: – Merchant’s because-islands involve subject extraction, while Temmerman’s object extraction. – Our own English object extraction case was only mildly deviant for most speakers, more or less perfect for some. Moreover we found that some Dutch speakers found cases like (15) less than perfect, some ruling them out altogether. – Temmerman’s good RC island involves object extraction from an indefinite relative, unlike Merchant’s bad example (6) but like our perfect case (10). • Importantly, repair fails in Dutch5 where it fails in English, i.e. definite DPs, left branches. (17)

a.

Heb je die nieuwe kerel van F RANKRIJK ontmoet? have you that new guy from France met “Did you meet the new guy from F RANCE? b. *Nee, van Itali¨e no from Italy “No, (from) Italy”

4 Parentheses

Dutch, definite DP island6

in some parts of Temmerman’s examples indicate so and Jeroen van Cranenebroeck confirms this for us. Note that we have found that not all speakers agree about the status of adjunct island cases like (15), much like with our English data. 5 The Dutch here represents the Flemish-inflected variety of Jeroen van Cranenebroeck, but it has been checked with speakers of other dialects as well (who e.g. offer uit in place of van in (17)). 6 Note this is on the relevant “different guys” interpretation

Island repair: clausal vs non-clausal islands

(18)

5

a.

Heeft Marie een aantal LUIE werkers in dienst? has Mary a number lazy workers in service “Does Mary have a number of LAZY workers in service? b. *Nee, HARD no hard “No, HARD” Dutch, left branch island

• Conclusion: the clausal/non-clausal island distinction holds across English and Dutch contrastive fragments; there is no cross-linguistic difference in repair here. 2.3

Interim summary • Contrastive fragments show repair with some islands but not others • Non-clausal islands particularly resistant to repair

3. Non-repair in “repair” contexts • Possible conclusion from last section: non-repair is still only attested with contrastive ellipsis, cf. Griffiths and Lipt´ak (2012). • But we can go further and show that non-repair also occurs in non-contrastive ellipsis. Again, the key cases involve non-clausal islands, i.e. left branch extraction. 3.1 English “left branch extraction” in sluicing • In English, the DP is an island for left-branch modifiers, such as possessors (19a) and attributive APs (19b) (see esp. Corver (1990)). (19)

a. b.

*Whose did Jane kiss [t husband]? *[How tall] did Mary marry a t man?

• Focus on adjectival LBE; Possessors and demonstratives independently licensed nominal ellipsis (20a) - introducing a confounding factor - whereas attributive APs don’t (20b). (20)

a. b.

Jane kissed Mary’s husband, and Susan kissed Helen’s husband *Mary married a tall man, and Jane married a short man

• Merchant (2001) argues based on data like (21) that LBEs are repaired under ellipsis. (21)

*Mary married a tall man, but i’m not sure how tall she married a t man

• But adjectival agreement in German (and other Germanic languages, such as Dutch) suggests that the

6

Island repair: clausal vs non-clausal islands remnant AP is not merged as an NP-modifier, and hence that the e-site is not, in fact, isomorphic7 The adjectival wh-remnant shows up in its bare form (23), patterning with an AP predicate (22).8 (22)

Der Mann ist groß(*en). The man is tall(. ACC).

(23)

Lena hat einen großen Mann geheiratet, aber ich weiß nicht wie groß(*.en) L. has a tall. ACC man married, but I know not how tall(*. ACC). “Lena has married a tall man, but I don’t know how tall”

• Not just ‘starred trace’ approaches to island repair that encounter difficulties here: semantic accounts of sluicing, such as Barker (to appear) and Culicover and Jackendoff (2005) also troubled. • Observation here: difference between predicative and non-predicative adjectives in “left-branch sluices” provides further evidence against isomorphic e-site and for predicative source. • Predicative adjectives (e.g. diligent) and non-predicative adjectives (e.g. hard as in ‘hard-working’) appear as modifiers (24a), but only predicative adjectives appear as predicates (24). Both kinds can be wh-fronted, so long as they’re gradable (25). (24)

a.

Billy hired a diligent/hard worker.

This worker is diligent/*hard. (25)

[How diligent/hard] a worker did Billy hire t?

• A non-predicative AP remnant is markedly degraded compared to a predicative counterpart (with a modifier correlate). Suggests the AP is merged as a predicate, and that an isomorphic e-site not possible, as it would give rise to unrepaired LBE. (26)

*?Billy hired a hard worker, but I don’t know how hard.

(27)

Billy hired a diligent worker, but I don’t know how diligent.

• Same can be shown for “old” as in “old friend/colleague”; for related data from fragments see e.g. Merchant (2004, 688). See also data from earlier. 7 Merchant

acknowledges these facts, but resists the conclusion that this is evidence for a predicational analysis, speculating that “...this lack [of adjectival agreement] may open an interesting window into the nature of the inflection itself” (p233). Merchant provided two reasons for not adopting a predicative analysis: (i) it would apparently require loosening of the identity condition on ellipsis, and (later on) (ii) we see repair by deletion with LBE in the domain of attributive comparative subdeletion as well (see e.g. Kennedy and Merchant 2000). Here we’ll argue that the loosening of the identity condition is inevitable. For alternative analyses of the comparative subdeletion facts, see Izvorski 1995 and Kennedy 2002. 8 Many languages which seem to show matching effects on APs, such as Greek, independently allow post-adjectival nominal ellipsis, making it difficult to draw any conclusion regarding island evasion vs. repair

Island repair: clausal vs non-clausal islands

7

• Importantly, preliminary investigation indicates non-predicative adjectives can be sluicing remnants in languages that allow LBE (Polish, Russian). • So: attributive agreement fails to show up on “LBE sluicing” remnants, and non-predicative adjectives can’t be remnants at all. • Simplest explanation: sluicing never repair LBEs, and ‘repair’ LBEs have predicational structure. • LB sluices are a case of island evasion; when the evasion strategy isn’t available, extraction from an island is necessary, and this can’t be repaired, even in sluicing. Importantly, this fits with the picture from previously: non-clausal island fails to show repair. 3.2 The necessity of predicational sources • The (lack of) case matching effects in (22)-(23) and the degraded status of non-predicative AP remnants suggest a non-isomorphic source. • Predicational source9 with E-type pronoun subject (Merchant 2001) captures the data we have seen. (28)

Mary married a tall man, but I don’t know how tall hee−type was t

• Independent evidence for predicational source: Sluicing in headed unconditionals (Rawlins (2008)) (Discussed briefly by Merchant (2001, p236) in a footnote as ‘concessive’ sluices). (29)

Mary would marry ANYONE, no matter WHO.

• No repair here, so we can see quite clearly than an isomorphic source is ruled out independently: (30)

Mary would marry ANYONE, no matter WHO Mary would marry.10

• (30) sounds redundant, and not an accurate paraphrase of (29). We can explain the weirdness of (29) by appealing to the interpretive mechanisms introduced by Rawlins 2008. • In Rawlins’ system, no matter is treated as complex head which takes an interrogative complement. [no matter [CP ... ]] is an adjunct that behaves semantically and syntactically as the antecedent of a conditional (like an if -clause). The embedded wh-question gives an (exhaustive) set of alternatives, which jointly entail the antecedent (matrix) clause. (31) 9 It’s

The party will be fun, no matter who comes t. = If John comes, the party will be fun & If Mary comes the party will be fun, etc.

more common in the lit. to invoke a left-source to account for otherwise mysterious phenomena (see e.g. Vicente (2008) on apparent P-stranding violations in Spanish.) but they won’t work for left branch sluicing since predicates generally are degraded as pivots, e.g. ?/??It is TALL that John is., and they don’t work in unelided wh-questions, i.e. *but I don’t know how tall it was (as noted by Merchant). 10 Note that changing the modality of the antecedent does not help matters much, i.e. *Mary would marry anyone, no matter who Mary marries.

8

Island repair: clausal vs non-clausal islands

• Applying this interpretive mechanism to the infelicitous (30), we can see what the problem is. Every alternative antecedent vacuously entails the consequent: (32)

Mary would marry anyone, no matter who Mary would marry. = If Mary would marry John, then Mary would marry him & if Mary would marry Bill, then Mary would marry him, etc.11

• A predicational source gets the right interpretation, and also allows a felicitous unelided continuation. (33)

Mary would marry anyone, no matter who they were. = If they were a shoe salesmen, then Mary would marry them & If they were a fan of Slade, then Mary would marry them

• The argument extends to unconditional sluices with a left branch remnant12 , but this is harder to see, since an isomorphic continuation is independently out due to illicit LBE.

4.

(34)

# Mary would marry any man, no matter how tall Mary would marry the t man. = #If Mary would marry the 6ft tall man, then Mary would marry that man & If Mary would marry the 5ft tall man, then Mary would marry that man, etc.

(35)

Mary would marry any man, no matter how tall he was. = If the man was 6ft tall, then Mary would marry him & If the man was 5ft tall, then Mary would marry him

More on short sources • Preceding discussion shows that non-isomorphic sources for clausal ellipsis must be possible. This is not new: Merchant (2001, ch.5) argues this happens in other cases of sluicing. • For (1) he proposes that the correct underlying structure is a non-isomorphic “short source” like (36), with the identity condition being satisfied by modal subordination (for subjunctive relatives) and an E-type pronoun subject. Another case of evasion. (36)

They want to hire someonei who speaks a Balkan language, but I don’t know which one hei should speak

• So evasion may account for many cases of apparent island violation in sluicing.13 Much subsequent work dismisses this and maintains, presumably due to reactions in Lasnik (2001), but see Barros (to appear) for responses to some points made by Lasnik. 11 Note

that we’re abstracting away from how free-choice ’any’ gives rise to a pair-list reading. also that unconditional sluices violate the general ban on sprouting AP modifiers, i.e. Mary married a man, but I don’t know how tall, suggesting that this is a semantic, rather than a syntactic restriction. 13 Merchant (2001) argues that there are some islands that resist an evasion analysis and thus are true “PF islands:” (i) Left Branch islands, (ii) 12 Note

Island repair: clausal vs non-clausal islands

9

• Simple point here: short sources may account for variation in contrastive fragments with relative clause islands observed earlier. Short sources possible for (10) (=(37)). (37)

A: Did they hire someonei who works on F RENCH (last year)? B: No, G ERMAN hei works on.

• What about RC extractions that fail? Don’t seem to have plausible short source parses (Abels 2011). (38)

A: Does Abby speak the same Balkan language that B EN speaks? B: *No, C HARLIE ... = ??? cf. #No, C HARLIE speaks the same Balkan language that Abby speaks. #No, C HARLIE speaks it.

• Full answer here would not answer the Question Under Discussion (cf. Roberts 1996, Barros to appear) so it would not work as a source for the fragment either. • Similarly consider definite DP islands: (39)

A: Did you give Mary’s picture of P RINCE to her?

B: ?*No, E LVIS.

• No obvious short source. Indefinite variant I gave a picture of Prince to her would be incongruent. • Conclusion: RCs often don’t show repair effects, and this seems to correlate with when evasion by use of a short source is possible. 5.

Clausal islands and island pied-piping • The residue: adjunct islands, which speakers often accept as OK or just marginal. (40)

A: Did they leave because you offended M ARY?

B: ?No, S ARAH.

• There are not amenable to a short source evasion: the putative you offended Sarah is quite distinct from the isomorphic version they left because you offended Mary,14 and an answer of the form of the short source would be incongruent. • Why are these CP complements permeable? A relevant observation: another option in cases like (40) is island pied-piping (see e.g. Krifka 2006, also Merchant 2004): (41)

A: Did they leave because you offended M ARY?

B: No, because I offended S ARAH.

This may be described as an evasion strategy: don’t leave the island! 14 (Merchant,

2001, 300-301) proposes a way of deriving a semantically identical short source for sluicing with if -islands. This strategy is not available here; for discussion of the role of the scope of focus in distinguishing short source and isomorphic source readings, see especially Fukaya (2007).

10 Island repair: clausal vs non-clausal islands • Nishigauchi (1990), Richards (2000): this happens in covert syntax too.15 • Another observation: when A0 -operators that need to move occur in islands, often the result is the operator moves to the left periphery of the island. In Finnish, the island is then pied-piped to the scope position of the operator (Huhmarniemi 2012): (42)

a. [CP1

.... [island wh xi

... ti ... ] j ... ]

b. [CP1 [island whi .. ti x.. ] j ....

t j ... ]

In Romanian, the wh-element is content at the left edge of the island, though its interpretation is restricted to a single-pair question (Cheng and Demirdache 2010). Something similar going on with parasitic gaps in English too (Kayne 1983). • Something similar seen with (non-island) clausal pied-piping in Basque (Arregi 2003), Bavarian (Bayer 2001): move to left edge of CP, then move the containing CP. (43)

a. [CP1

.... [CP2 wh xi

... ti ... ] j ... ]

b. [CP1 [CP2 whi .. ti x.. ] j ....

t j ... ]

• Proposal: these sorts of ingredients are involved in deriving (40) from isomorphic source, i.e. 1. move focus to left edge of island, Spec,CP 2. apply clausal ellipsis, eliding the TP and the overt C because 3. pied-pipe whole island to matrix Spec,CP 4. apply clausal ellipsis again, leaving just the fragment Clausal ellipsis thus applies twice, once with movement to the edge, then with movement of the island. We call this Double Clausal Ellipsis (DCE). A schematic for DCE in (40). (44)

a. [CP1 [T P they left [CP2 Sarah x i [C0 because [T P you offended ti ]]]]] move to edge of island b. [CP1 [T P they left [CP2 Sarahi [C0 because [T P you offended ti ]]]]] c. [CP1 [CP2 Sarahi [C0 because [T P you offended ti ]]]x [T P they left t j ]]

clausal ellipsis pied-pipe island

d. [CP1 [CP2 Sarahi [C0 because [T P you offended ti ]]] [T P they left t j ]]

clausal ellipsis

• All ingredients independently available: – first stage: focus movement to left periphery of island CP, possible in absence of ellipsis in Finnish, Romanian, but only in presence of ellipsis in English. – second stage: requires ellipsis licensing in Spec,CP, eliding overt C. Unproblematic given that with sluicing, stripping, fragments, ellipsis is licensed in this position with focus movement to Spec; see Thoms (2010) for a treatment of why movement to Spec,CP licenses deletion of C (the sluicing-comp generalisation) 15 Note

the tension between these sorts of theories of islands and the PF-theory of islands: covert syntax shouldn’t be constrained by islands if they’re PF phenomena!

Island repair: clausal vs non-clausal islands

11

– third stage: island pied-piping, seen in English (see above) – final stage: clausal ellipsis of usual kind in fragments • Only real leap of imagination is allowing fronting to left periphery of island (what we see overtly in Finnish, Romanian) only when ellipsis applies. This seems inevitable in other contexts too, e.g. pseudogapping, most cases of stripping, multiple fragments etc (see Thoms to appear). • Crucially, DCE will only be available with clausal islands, where clausal ellipsis can be licensed. No analogous position where ellipsis can apply to leave just a contained remnant at the edge in DPs. • May also give us an account of why object extraction cases like (40) differ from subject extraction cases like Merchant’s (8), repeated here (recall they consistently rated lower than (40)). (45)

A: Did Ben leave the party because A BBY wouldn’t dance with him?

B: *No, B ETH.

• Deriving this by DCE would involve an (unrepaired!) that-trace violation: (46)

[CP1 [T P Ben left [CP2 Bethx i [C0 because [T P ti wouldn’t dance with him ]]]]] *stage 1 ×

• Note: DCE won’t work for RC islands, since Spec,CP isn’t an ellipsis licensing site in English RCs (and presumably movement to Spec,CP is blocked by filled spec); moreover it would not license ellipsis of the NNP, so would derive the wrong string. So we don’t (incorrectly) predict DCE to cure all islands (cf. Bl¨umel 2013).16 • DCE is taking advantage of fact that clauses are larger domains in which ellipsis can readily apply. 6.

Conclusion • In general, non-clausal islands show “repair” more than clausal islands. Clausal/non-clausal distinction cuts across categories previously identified in literature. • But clausal islands show non-repair too, so this isn’t a syntactic difference. • Best explanation: various evasion strategies, all disguised by ellipsis, all available independently. When you’re cornered and don’t have any of these evasion strategies, you get island sensitivity in sluicing and fragments. No repair!

References Abels, Klaus. 2011. Don’t fix that island! it ain’t broke, paper presented at the Islands in Contemporary Linguistic Theory conference, Vitoria-Gasteiz. Arregi, Karlos. 2003. Clausal pied-piping. Natural Language Semantics 11:115–143. Barker, Chris. to appear. Scopability and sluicing. Linguistics and Philosophy . 16 All

we have for this is an abstract (for a talk at iGG39), and we happened upon it very late in the day, so we may be missing subtleties in Bl¨umel’s full treatment.

12 Island repair: clausal vs non-clausal islands

Barros, Matthew. to appear. A non-repair approach to island-sensitivity in contrastive TP ellipsis. In Proceedings of Chicago Linguistics Society 48. Chomsky, Noam. 1972. Spheres of influence in the age of imperialism; papers submitted to the Bertrand Russell Centenary Symposium, Linz, Austria, September 11th to 15th, 1972. Nottingham: Bertrand Russell Peace Foundation for The Spokesman. Chung, Sandra, Ladusaw, William A., and McCloskey, James. 1995. Sluicing and Logical Form. Natural Language Semantics 3:239–282. Corver, Norbert. 1990. The syntax of left branch extractions. Ph.D. thesis, Katholieke Universiteit Brabant. Culicover, Peter W. and Jackendoff, Ray. 2005. Simpler Syntax. Oxford University Press. Davies, William D. and Dubinsky, Stanley. 2003. On extraction from NPs. Natural Language and Linguistic Theory 21:1–37. Fukaya, Teruhiko. 2007. Sluicing and stripping in Japanese and some implications. Ph.D. thesis, University of Southern California. Griffiths, James and Lipt´ak, Anik´o. 2012. Contrast and island-sensitivity in clausal ellipsis. Syntax . Izvorski, Roumyana. 1995. A solution to the subdeletion paradox. In Jos´e Camacho, Lina Choueiri, and Maki Watanabe, eds., Proceedings of WCCFL 14, 203–219, Stanford, CA: CSLI Publications. Kayne, Richard S. 1983. Connectedness. Linguistic Inquiry 14:223–249. Kennedy, Chris. 2002. Comparative deletion and optimality in syntax. Natural Language and Linguistic Theory 20:553–621. Kennedy, Chris and Merchant, Jason. 2000. Attributive comparative deletion. Natural Language and Linguistic Theory 18:89– 146. Krifka, Manfred. 2006. Association with focus phrases. In V. Moln´ar and Susanne Winkler, eds., The architecture of focus., number 82 in Studies in Generative Grammar, 105–135, Berlin: Mouton de Gruyter. Lasnik, Howard. 2001. When can you save a structure by destroying it? In Minjoo Kim and Uri Strauss, eds., Proceedings of the North East Linguistic Society 31, 301–320, Georgetown University: GLSA. Merchant, Jason. 2001. The syntax of silence: sluicing, islands, and the theory of ellipsis. Oxford: Oxford University Press. Merchant, Jason. 2004. Fragments and ellipsis. Linguistics and philosophy 27:661–738. Merchant, Jason. 2008. Variable island repair under ellipsis. In Kyle Johnson, ed., Topics in Ellipsis, 132–153, Cambridge, United Kingdom: Cambridge University Press. Nishigauchi, Taisuke. 1990. Quantification in the Theory of Grammar, volume 37. Dordrecht, The Netherlands: Kluwer Academic Publishers. Rawlins, Kyle. 2008. (un)conditionals: an investigation in the syntax and semantics of conditional structures. Ph.D. thesis, UCSC, Santa Cruz. Richards, Norvin. 2000. An island effect in Japanese. Journal of East Asian Linguistics 9:187–205. Roberts, Craige. 1996. Information Structure: Towards an integrated theory of formal pragmatics. Ohio State University: OSU Working Papers in Linguistics. Ross, John Robert. 1969. Guess who? In Robert I. Binnick, Alice Davison, Georgia M. Green, and Jerry L. Morgan, eds., Chicago Linguistics Society, 252–286, Chicago, Illinois. Temmerman, Tanja. 2013. The syntax of Dutch embedded fragment answers: on the PF-theory of islands and the WH/sluicing correlation. Natural Language and Linguistic Theory 31:235–285. Thoms, Gary. 2010. Syntactic reconstruction and Scope Economy. Paper presented at GLOW33, Wroclaw, Poland. Thoms, Gary. to appear. Constraints on exceptional ellipsis are only parallelism effects. In Proceedings of NELS 43. Vicente, Luis. 2008. Syntactic isomorphism and non-isomorphism under ellipsis, ms., UCSC.