Linguists have learned a great deal about (internal) language by looking at complex constructions in a number of different (external) languages, but sometimes constructions that seem simple on the surface can also shed light on the workings of the grammar and fundamental theoretical issues. Take for example the English highest subject question in (1).
Despite its straightforward appearance, there is very little consensus on what the proper derivation for (1) should be. The most intuitive and standard analysis would have the wh-phrase moving to the specifier of CP through the TP specifier (e.g., Pesetsky & Torrego 2001), as shown in (2).1
Still, many authors have argued that unlike other types of wh-questions, highest subject wh-phrases do not move to the specifier of CP and instead stay in the specifier of TP (e.g., George 1980; Chung & McCloskey 1983; Chomsky 1986; Agbayani 2000; Carstens et al. 2016; Gallego 2017).
Yet other authors assume (explicitly or implicitly) that the wh-phrase moves directly from its argument position inside the vP to the specifier of CP, eschewing intermediate movement to the specifier of TP (e.g., Rizzi 1982; Brandi & Cordin 1989; Campos 1997; McCloskey 2000; Holmberg & Hróarsdóttir 2003; Fitzpatrick 2006; Bošković 2016; Erlewine 2016), as shown in (4).
As I will detail below, there is evidence that (4) is indeed the correct derivation, but the derivation in (4) raises an immediate question about the status of the EPP on T in languages like English where there appears to be an otherwise obligatory constraint that the specifier of TP be filled. In fact, the authors arguing for the derivation in (4) have not provided a satisfying answer to this question (though there have been analyses that account for a derivation like (4) with languages that allow null subjects such as Romance languages (Rizzi & Shlonsky 2007) and Bantu languages (Diercks 2010)).2
In the literature, there have been two general approaches to EPP effects. On the standard account, T has a formal requirement that its specifier be filled. In the feature checking framework, this was accomplished by T having an uninterpretable feature that could only be checked by movement to its specifier (Chomsky 1995; 2000; 2001). In the labeling framework, Chomsky (2015), argues that T is too “weak” to label on its own, and hence requires a specifier in order to project. The other view of EPP effects takes them to be epiphenomenal. T does not have a specifier requirement. Movement to the specifier of TP is independently motivated either to check the features of the moving element (Epstein & Seely 2006; Bošković 2007) or to repair an otherwise illicit interface representation (Alexiadou & Anagnostopolou 2001; 2007; Landau 2007; Richards 2010; 2016; Bayer & Salzmann 2013; Chomsky 2013; McFadden & Sundaresan 2018). Chomsky (2013) provides such an account within this view in the labeling framework. Under this account, merging of the subject and vP causes a problem for the labeling algorithm that is remedied by the subject moving away. The latter view of EPP effects quite naturally allow for a derivation like (4), while the standard analysis fails to account for it. The paper first reviews the arguments from the literature that there is no movement to the specifier of TP in highest subject questions. Then it provides an analysis of the lack of movement to the specifier of TP in the labeling framework of Chomsky (2013).
In this section, I will present arguments that the correct derivation for highest subject questions involves direct movement from the wh-phrase’s base position to the specifier of CP. I will first present three arguments against the wh-staying-in-SpecTP approach and then two arguments against the intermediate movement to the specifier of TP approach.3
Let us first examine the arguments against the subject wh-phrase staying in SpecTP. The first argument comes from agressively non-d-linked wh-phrases (Pesetsky 1987; Ginzburg & Sag 2000). Ginzburg & Sag (2000: 229–337) and Pesetsky & Torrego (2001: fn.9) note that the distribution of aggressively non-d-linked wh-phrases like what/who the hell is a subset of the distribution of plain wh-phrases (see also den Dikken & Giannakidou (2002) for an analysis of these phrases in matrix and embedded environments). As (5) shows, while aggressively non-d-linked phrases can occur in the specifier of CP, they cannot occur in situ.
|(5)||a.||What the hell did you buy?|
|b.||*Who bought what the hell?|
The authors take the acceptability of constructions like (6), involving an aggressively non-d-linked subject wh-phrase, to provide evidence that the wh-phrase here moves to the specifier of CP,4 because if it stayed in SpecTP we would expect (6) to be ungrammatical on a par with (5b).
|(6)||Who the hell bought a car?|
Another argument for movement to the specifier of CP was noted in Agbayani (2000: fn.2). Agbayani notes that the typical analysis of the elliptical construction known as sluicing involves movement to the specifier of CP followed by deletion of the TP, as shown in (7).
|(7)||John bought something, but I don’t know what
If this analysis is correct, then the existence of sluices where the wh-phrase is a highest subject indicates that the subject moves out of TP into the specifier of CP.
|(8)||A:||Someone bought a car.|
In order for this argument to go through, a few possible alternative explanations must be ruled out that Agbayani did not consider. First, the example in (8) cannot be analyzed as an instance of what is called pseudosluicing or deletion of an underlying cleft, as illustrated in (9).
|(9)||A:||Someone bought a car.|
Merchant (2001) presents a number of tests that can control for a derivation like the one in (9). One test is the use of else-modification. As Merchant shows, else-modification is incompatible with clefts as shown in (10a), but is acceptable in sluicing (10b) (capitalization indicates focal stress).
|(10)||a.||#John bought A CAR, but I don’t know what else it was.|
|b.||John bought A CAR, but I don’t know what else.|
If the matrix sluice in (8) is an underlying cleft, then we would expect an else modified matrix subject sluice to pattern with (10a) in being unacceptable, but if the source of such sluices is indeed full syntactic structure, then we would expect such an example to pattern with (10b). As (11) shows, a matrix subject sluice is acceptable with else modification, indicating that the source of such sluices is not underlying cleft.
|(11)||A:||A TELEVISION was stolen.|
Second, one may wonder whether the correct generalization is that any moved wh-phrase can license the deletion of its sister. This would allow for the example in (8) to not involve ellipsis of TP as is standardly assumed, but rather T’. Stjepanović (1999), however, provides evidence against such an analysis. Stjepanović shows that wh-phrases in Serbo-Croatian may front below CP, but in cases where sluicing applies, the wh-phrase must reside in Spec CP.
The final argument I will discuss involves apparent coordinate structure constraint (CSC) violations such as (12) discussed in Bošković (2018).
|(12)||*Who left and John went to the store?|
If subject wh-phrases were able to stay in the specifier of TP, as schematized in (13a), the structure would not violate the CSC and examples like (12) would be predicted to be grammatical, contrary to fact. If there is movement to the specifier of CP as in (13b), we have a CSC violating structure and correctly predict the ungrammaticality of (12).
|(13)||a.||[&P [IP who left] & [IP John went to the store] ]|
|b.||[CP whoi [&P [IP ti left] & [IP John went to the store] ] ]|
Evidence that this is indeed a CSC violation comes from the fact that the structure can be salvaged by the across-the-board exception to the CSC shown in (14).
|(14)||a.||Who left and went to the store?|
|b.||[CP whoi [&P [IP ti left] & [IP ti went to the store] ] ]|
All these data suggest that subject wh-phrases do not stay in the specifier of TP and instead, like all other wh-phrases, move to the specifier of CP. The next section will further provide evidence that not only do subject wh-phrases not stay in the specifier of TP, they never move to that position in the first place.
I will present two arguments that movement of highest subject questions does not stop off in the specifier of TP. The first argument I will present comes from the distribution of quantifier float in West Ulster English, as described in McCloskey (2000) and further elaborated in Fitzpatrick (2006) and Henry (2012).
|(15)||West Ulster English (McCloskey 2000: 58)|
|a.||What did you get all for Christmas?|
|b.||Who did you meet all when you were in Derry?|
|c.||Where did they go all for their holidays?|
Importantly, highest subject questions also allow for quantifier float (16a), although, as in standard English, quantifier float is impossible in (16b). The possibility of quantifier float in (16a) is unexpected if there is first A-movement to the specifier of TP as that movement step would be indistinguishable from the movement in (16b), where quantifier float is impossible.
|(16)||West Ulster English (McCloskey 2000: 77)|
|a.||Who was throwing stones all around Butchers’ Gate?|
|b.||*They were throwing stones all around Butchers’ Gate.|
These data suggest that the proper treatment of highest subject sentences is one in which the wh-phrase moves directly from its base position to the specifier of CP without intermediate movement to the specifier of TP.5
The second argument comes from extraction out of British do ellipsis as discussed in den Dikken & Griffiths (2018). It has been noted that A-movement is allowed out of such ellipsis while wh-extraction is not allowed. This is shown in (17). In (17a), we have an example of NP-raising out of the elided VP and the example is grammatical even with do suggesting that A-movement is possible out of such constructions. In (17b), on the other hand, we have wh-movement of an object out of the ellipsis site and the result is ungrammatical with do, suggesting that this type of extraction is disallowed.
|(17)||British English (den Dikken & Griffiths 2018: (21)–(22))|
|a.||John might seem to enjoy that, and Petei might (do)
|b.||I know whoi John will kiss and who Pete will (*do)
Crucially, wh-movement in highest subject questions is also disallowed out of do ellipsis, as shown in (18). This again suggests that movement of highest wh-subject questions does not involve a step of A-movement to the specifier of TP that feeds movement to the specifier of CP, as if this were the case we would expect such a movement to be possible in (18) just as it possible in (17a). The fact that (18) and (17b) pattern together suggests that movement of highest subject questions instead move directly to the specifier of CP without an intermediate step of A-movement.
|(18)||British English (den Dikken & Griffiths 2018: (23))|
|A:||Sue wouldn’t kiss Peter last night|
|B:||Well, whoi would (*do)
These data point to a conclusion that has been reached by a number of authors: that movement from SpecTP to SpecCP is generally not allowed. This has been confirmed for a number of languages such as dialects of Italian (Rizzi 1982; Brandi & Cordin 1989; Campos 1997; Rizzi & Shlonsky 2007), Lubukusu (Diercks 2010), Kaqchikel (Erlewine 2016) and Kinande (Bošković 2016).6
The languages cited above, however, do not have a strong EPP requirement on T, but the data presented here suggest that this also the case in languages like English where there is otherwise an obligatory requirement that the specifier of TP be filled. This raises the question, if the subject wh-phrase can move directly to the specifier of CP, how can the EPP be satisfied? In the next section, I will provide an answer to this question building off of the labeling theory in Chomsky (2013).
There have been two approaches to the movement of the subject to the specifier of TP in the literature. The traditional approach posits that the target (i.e., T) is deficient or needy in some way: it has an EPP/D/Edge feature that needs to be checked by having an element in its specifier. The second approach argues that the movement in question is triggered by the moving element or by the need to repair an illicit interface representation. In recent work, Chomsky has provided two approaches to the EPP in the labeling framework (Chomsky 2013; 2015). Chomsky (2015) is the traditional account where T is deficient; Chomsky (2013) on the other hand, is in the spirit of the second approach. As detailed below, the lack of EPP effects in highest subject questions can be easily accounted for in the system presented in Chomsky (2013), but it is difficult to capture under the traditional approaches to EPP, including Chomsky (2015).
Chomsky (2013), following work by Moro (2000; 2008), gives an algorithm for providing labels to newly created syntactic objects. When a head and a phrase undergo merge, the head provides the label for the output. When two non-minimal projections undergo merge, there are two possible ways for this output to be labeled: (i) prominent feature sharing, where the shared feature provides the label for the object; or (ii) one of the two phrases undergoes movement, allowing for the remaining phrase to project the label (traces are ignored for labeling). With this in mind, let us examine the configuration where the subject is first merged into the derivation.
|(19)||[? DP [vP v [ … ]]]|
When the subject, a DP, merges with the vP, we have the case of merge with two phrases. There is then no head that can project the label so the label needs to be provided by either feature sharing or movement. Since the vP and DP do not share features here, the only possibility is movement. The derivation continues with the T being merged into the derivation; as a head, T will project as the label of the object in question. The DP then moves to the specifier of TP, allowing for the v to project a label in the lower part of the structure. Since the the DP and T share φ-features, the new object is labeled with those features.
|(20)||[φ DP [TP T [vP t [ v [ … ]]]]]|
Unlike traditional approaches to movement of the subject to the specifier of TP, this analysis does not rely on T being deficient in any way: there is in fact, no requirement for T to have a specifier. It is the “local instability” between the DP and vP, discussed above, that drives the movement of the subject to the specifier of TP.
Returning to highest subject questions, again when the subject is first merged, the resulting structure cannot be labeled.
|(21)||[? DP[+wh,+Q] [vP v [ … ]]]|
The T is merged into the derivation, and it again projects a label. C is then merged and projects a label. Since T is not deficient in any way and hence does not require a specifier, the subject can move directly to the specifier of CP. The labeling problem at the vP is again resolved as there is only one phrase left to project, and the output of subject DP and CP undergoing merge can be labeled via sharing of the [+Q] feature.
|(22)||[Q DP[+wh,+Q] [ C+Q [TP T [vP t [vP v [ … ]]]]]]|
The “local instability” approach hence allows us to give a principled explanation for why the otherwise obligatory movement to the specifier of TP does not occur in highest subject questions.7
Let us now investigate how the standard approach would deal with such data using Chomsky (2015) as a model though this holds of any variant of the standard approach. Chomsky proposes that T is “too weak” to provide a label on its own, which means that the result of merge in (23a) cannot be labeled. T requires merger with another element that it agrees with in its specifier in order for labeling in (23b) to occur.
|(23)||a.||[? T [? DP [vP v [ … ]]]]|
|b.||[φ DP [TP T [vP t [ v [ … ]]]]]|
Now in highest subject questions, if the subject wh-phrase were to move directly to the specifier of CP, a problem arises under this approach. Because T is too weak to label on its own, the output of its merge with vP in this hypothetical representation is never labeled and the derivation does not converge as a result.
|(24)||[Q DP[+wh,+Q] [ C+Q [? T [vP t [vP v [ … ]]]]]]|
Hence, the only convergent representation under this approach would have the subject wh-phrase move to the specifier of TP before moving to the specifier of CP, but the data from all float and do-ellipsis extraction presented in the previous section demonstrate that such a movement step does not in fact take place.
In this section, I have shown that the lack of movement to the specifier of TP in highest subject questions can be naturally captured by approaches where movement to the specifier of TP is not tied to a deficiency of T. Approaches that tie the movement in question to such a deficiency face a problem with the derivation of subject questions.
The previous section demonstrated that the Chomsky (2013) approach to EPP effects allows for a natural explanation for the lack of movement to the specifier of TP in highest subject questions. The question arises whether such an approach can further extend to account for other types of EPP and subject movement effects found cross-linguistically. As noted in the previous section, it has been argued that languages from Romance, Bantu and Mayan families also have direct movement to SpecCP in highest subject questions. This pattern seems robust cross-linguistically, and hence the analysis presented here has at least some cross-linguistic generality. There are other EPP related constructions that need to be looked at in more detail within this framework (e.g., stylistic fronting and quirky subjects in Icelandic (Holmberg 2000; Sigurðsson 2010; 2017), transitive expletive constructions also in Icelandic (Bobaljik & Jonas 1996), VP-fronting with the subject in German (Wurmbrand 2006), and the subject/object asymmetry in the island effects in relative clauses (Chung & McCloskey 1983) to name a few). I leave much of this investigating to future work, however, this section briefly investigates one additional area where the EPP is relevant: the presence of overt expletives in existential constructions.
|(25)||a.||There arrived a train in the station.|
|b.||*arrived a train in the station.|
A number of authors (Richards & Biberauer 2005; Deal 2009; Alexiadou & Schaefer 2011; Wu 2018), however, provide evidence that there must first be merged at the edge of vP phase.8 If this is truly the case, then the local instability of the labeling algorithm of v and there would again cause there to move to the specifier of TP. T and there would agree in φ-features and label the resulting structure (there shares φ-features with the associate DP via an independent agreement relation see Deal 2009). It again appears that the local instability approach can account for the data without the need of an EPP.9
This paper investigated the nature of the EPP through the lens of highest subject questions. The EPP has been a contentious issue in syntactic theory, as the typical analysis of it involves positing features on T that merely encode the need of a specifier. Because of this, many authors have attempted to eliminate the EPP, arguing instead that the effects often attributed to it in fact follow from other independent aspects of the grammar. These works typically show that EPP-less theories can have the same empirical coverage as theories that do employ the EPP. This paper argues further that not only can the EPP-less theories account for the same data as the EPP theories; they, in fact, have superior empirical coverage because they correctly predict that the requirement of T to have a specifier be lifted in highest wh-subject questions. This was demonstrated via a comparison of Chomsky’s recent analyses within the labeling framework.
1Note that the lack of do-support in such constructions follows from the morphological merger approach to do-support (Halle & Marantz 1993; Bobaljik 1995; Lasnik 1995b). Since the affix in C (assuming T-to-C) is PF adjacent to the verb, morphological merger is not blocked here (traces do not block it), hence do-support does not take place. See also Pesetsky & Torrego (2001) for an alternative analysis. Do does appear in emphatic uses in highest subject questions (e.g., Who DID leave?). I suspect that whatever mechanism that conditions emphatic do in non-questions (e.g., John DID (so) leave) is also responsible here (see Bruening 2010; Biberauer & Roberts 2012 for recent approaches and discussion to emphatic do).
2It should be noted that the disagreement about the placement of the wh-phrase in highest subject questions is especially acute in English because the lack of do-support and also the lack of other obvious effects on word order. In other languages, such as Mainland Scandinavian languages, there is much more robust evidence that favors a representation that places the wh-phrase in the specifier of CP. Take for example the Swedish cases in (i). In (ia), the order of the finite verb har and negation inte indicate that the verb has raised to the V2 position in C. Hence, as the wh-phrase vem precedes the verb, it indicates that vem occupies SpecCP. Likewise in embedded clauses, vem precedes the complementizer som, again indicating that vem resides in SpecCP (see Holmberg 1986 and much subsequent work).
While this is not as obvious from the word order in English, the evidence documented in the next section shows that English is like Swedish in that highest subject wh-phrases also reside in SpecCP.
3Agbayani (2000: 704) argues that the data below in (i) provide evidence that highest wh-subjects do not move to SpecCP. The data in (i) show that highest subjects cannot be moved to the left periphery via topicalization (original observation is from Lasnik & Saito 1992).
|(i)||a.||John, I like t.|
|b.||*John, t left.|
|c.||John thinks that Bill, Mary likes t.|
|d.||*John thinks that Bill, t likes Mary.|
Agbayani takes this data to indicate that short movement of subjects (including wh-movement) to the left periphery is generally disallowed. In order for this argument to go through, however, we must assume that topicalization is a perfect proxy for wh-movement, but evidence suggests that the two movement types diverge in non-trivial ways. For instance, Postal (1994) and Poole (2017) show there a number of environments that disallow topicalization, but do allow for wh-movement. So although it may be the case that highest subjects cannot be topicalized, it does not, however, indicate that they also cannot undergo wh-movement to SpecCP.
4In den Dikken & Giannakidou’s (2002) analysis, the specifier of a Focus projection.
5One may wonder about the possibility of examples like the following (taken from McCloskey 2000: 78 fn. 32)
|(i)||Who was all throwing stones in Guildhall Square?|
The possibility of such examples would suggest that wh-movement does not proceed directly to the specifier of CP because it appears all can be stranded in an intermediate position. As reported by McCloskey, most informants do find such examples unacceptable, but there is inter speaker variation. This was further backed up by dialectal work described in Henry (2012). For dialects that do allow for such examples, Henry puts forth an analysis where the quantifier is stranded at the vP phase edge and not the edge of TP (see also Davis 2018 for discussion).
6For example in Trentino and Florentino preverbal but not post verbal subjects are able to control full agreement on the verb, as shown by (i). In the cases where full agreement is not controlled, we find what is called the neutral clitic gl(i) (the third person masculine clitic form) as in (ib) (Campos 1997: 93).
Interestingly, highest subject questions pattern with (ib) in not showing full agreement and instead showing the neutral clitic form (ii).
This follows if the wh-phrase cannot even pass through SpecTP on its way to SpecCP, assuming that full agreement can only be controlled by a subject in SpecTP.
7As a reviewer notes, under this analysis, a derivation of matrix subject question with movement to the specifier of TP followed by movement to SpecCP would also converge, but as discussed in section 1.2, the data point to such movement never taking place. One way of blocking such a derivation is to rely on an Anti-locality condition (Erlewine 2016; Bošković 2016) which blocks movement that is too short. Erlewine and Bošković both argue that movement from SpecTP to SpecCP runs afoul of such a condition. One other possibility, hinted at by the reviewer, is to block the derivation on the grounds of economy conditions. A converging derivation with two steps of movement is blocked by a converging derivation that only requires one step of movement.
8See also Richards (2016) for another approach to expletives in existential constructions that does not require an EPP feature on T.
9Note that filling the specifier of TP with an expletive can also occur with an all stranding wh-phrase (McCloskey 2000: 79 fn. 33).
|(i)||Who was there throwing stones all down the town?|
In such cases, both the expletive and the wh-phrase would begin the derivation within the vP phase. Subsequently the expletive moves to the specifier of TP and the label for the phrase is created via φ-feature sharing with the T head. The wh-phrase likewise moves to the specifier of CP where the phrase is labeled via Q-feature sharing.
3 = third person, M = masculine, SG = singular, PL = plural, CL = clitic.
For helpful comments, discussion and guidance, I thank the Glossa editors and three anonymous reviewers as well as Željko Bošković. All errors are of course my own.
The author has no competing interests to declare.
Agbayani, Brian. 2000. Wh-subjects in English and the vacuous movement hypothesis. Linguistic Inquiry 31. 703–713. DOI: https://doi.org/10.1162/002438900554523
Alexiadou, Artemis & Elena Anagnostopolou. 2001. The subject-in-situ generalization and the role of case in driving computations. Linguistic Inquiry 32. 193–231. DOI: https://doi.org/10.1162/00243890152001753
Bayer, Josef & Martin Salzmann. 2013. That-trace effects and resumption: How improper movement can be repaired. In Patrick Brandt & Eric Fuß (eds.), Repairs: The added value of being wrong, 275–333. Berlin: Mouton de Gruyter.
Bošković, Željko. 2007. On the locality and motivation of Move and Agree: An even more minimal theory. Linguistic Inquiry 38. 589–644. DOI: https://doi.org/10.1162/ling.2007.38.4.589
Bošković, Željko. 2016. On the timing of labeling: Deducing comp-trace effects, the subject condition, the adjunct condition and tucking in from labeling. The Linguistic Review: Special issue on Labeling 33. 17–66. DOI: https://doi.org/10.1515/tlr-2015-0013
Brandi, Luciana & Patricia Cordin. 1989. Two Italian dialects and the null subject parameter. In Osvaldo Jaggli & Kenneth Safir (eds.), The null subject parameter, 111–142. Dordrecht: Kluwer. DOI: https://doi.org/10.1007/978-94-009-2540-3_4
Bruening, Benjamin. 2010. Language-particular syntactic rules and constraints: English locative inversion and do-support. Language 86. 43–84. DOI: https://doi.org/10.1353/lan.0.0201
Carstens, Vicki, Norbert Hornstein & T. Daniel Seely. 2016. Head movement in Problems of Projection. The Linguistic Review: Special issue on Labeling 33(1). 67–86. DOI: https://doi.org/10.1515/tlr-2015-0014
Chomsky, Noam. 2000. Minimalist inquiries: The framework. In Roger Martin, David Michaels & Juan Urigareka (eds.), Step by step; Essays on minimalist syntax in honor of Howard Lasnik, 89–115. Cambridge, MA: MIT Press.
Chomsky, Noam. 2013. Problems of projection. Lingua 130. 33–49. DOI: https://doi.org/10.1016/j.lingua.2012.12.003
Chomsky, Noam. 2015. Problems of projection: Extensions. In Elisa Di Domecinco, Cornelia Hamann & Simona Matteini (eds.), Structures, Strategies and Beyond, 3–16. Amsterdam/Philadelphia: John Benjamins. DOI: https://doi.org/10.1075/la.223.01cho
Deal, Amy Rose. 2009. The origin and content of expletives: Evidence from selection. Syntax 12(4). 285–323. DOI: https://doi.org/10.1111/j.1467-9612.2009.00127.x
den Dikken, Marcel & Anastasia Giannakidou. 2002. From Hell to polarity: “Aggressively non-d-linked” wh-phrases as polarity items. Linguistic Inquiry 33(1). 31–61. DOI: https://doi.org/10.1162/002438902317382170
Erlewine, Michael. 2016. Anti-locality and optimality in Kaqchikel Agent Focus. Natural Language and Linguistic Theory 34(2). 429–479. DOI: https://doi.org/10.1007/s11049-015-9310-z
Gallego, Ángel. 2017. The EPP in Labeling Theory: Evidence from Romance. Syntax 20(4). 384–399. DOI: https://doi.org/10.1111/synt.12139
Halle, Morris & Alec Marantz. 1993. Distibuted morphology and the pieces of inflection. In Kenneth Hale & Samuel Jay Keyser (eds.), The view from Building 20: Essays in linguistics in honor of Sylvian Bromberger, 111–176. Cambridge, MA: MIT Press.
Holmberg, Anders. 2000. Scandinavian stylistic fronting: How any category can become and expletive. Linguistic Inquiry 31(3). 445–483. DOI: https://doi.org/10.1162/002438900554406
Holmberg, Anders & Thorbjörg Hróarsdóttir. 2003. Agreement and movement in Icelandic raising constructions. Lingua 113. 997–1019. DOI: https://doi.org/10.1016/S0024-3841(02)00162-6
Landau, Idan. 2007. EPP extensions. Linguistic Inquiry 38. 485–523. DOI: https://doi.org/10.1162/ling.2007.38.3.485
Lasnik, Howard. 1995b. Verbal morphology: Syntactic Structures meets the Minimalist Program. In Héctor Campos & Paula Kempchinsky (eds.), Evolution and revolution in linguistic theory: Essays in honor of Carlos Otero, 251–275. Washington, DC: Georgetown University Press.
McCloskey, James. 2000. Quantifier float and Wh-movement in an Irish English. Linguistic Inquiry 31(1). 57–84. DOI: https://doi.org/10.1162/002438900554299
McFadden, Thomas & Sandhya Sundaresan. 2018. What the EPP and COMP-trace effects have in common: Constraining silent elements at the edge. Glossa 3(1). 1–34. DOI: https://doi.org/10.5334/gjgl.419
Postal, Paul. 1994. Contrasting movement types. Journal of Linguistics 30. 159–186. DOI: https://doi.org/10.1017/S0022226700016212
Richards, Marc & Theresa Biberauer. 2005. Explaining Expl. In Marcel den Dikken & Christina Tortora (eds.), The function of function words and functional categories, 115–153. Amsterdam/Philadelphia: John Benjamins. DOI: https://doi.org/10.1075/la.78.06ric
Richards, Norvin. 2010. Uttering trees. Cambridge, MA: MIT Press. DOI: https://doi.org/10.7551/mitpress/9780262013765.001.0001
Richards, Norvin. 2016. Contiguity theory. Cambridge, MA: MIT Press. DOI: https://doi.org/10.7551/mitpress/9780262034425.001.0001
Rizzi, Luigi. 1982. Issues in Italian syntax. Dordrecht: Foris. DOI: https://doi.org/10.1515/9783110883718
Sigurðsson, Halldór. 2010. On EPP effects. Studia Linguistica 64(2). 159–189. DOI: https://doi.org/10.1111/j.1467-9582.2010.01171.x
Sigurðsson, Halldór. 2017. Stylisitc fronting in corpora. In Höskuludur Thràinsson, Caroline Heycock, Hjalmar P. Peterson & Zakaris Svabo Hansen (eds.), Syntactic variation in insular scandinavian, 307–338. Amsterdam/Philadelphia: John Benjamins.
Wurmbrand, Susi. 2006. Licensing case. Journal of Germanic Linguistics 18(3). 175–236. DOI: https://doi.org/10.1017/S1470542706000079