Dropping the F-bomb: An argument for valued features as derivational time-bombs

*

Laura Kalin

Princeton University

1. Introduction

A long tradition in generative syntax holds that nominals have two fundamental licensing
needs (Vergnaud 1977/2008, Chomsky 1980 et seq): (i) every nominal needs a semantic
role (theta role), and (ii) every nominal needs abstract syntactic licensing, capital C Case.
This paper joins a rich body of work that challenges some aspect of the latter component of
this picture (Zaenen et al. 1985, Marantz 1991, Carstens 2001, Pesetsky & Torrego 2007,
Bobaljik 2008, Danon 2011, Preminger 2011, Arregi & Nevins 2012, Bhatt & Walkow
2013, i.a.). The argument made in this paper is that nominal licensing is not driven by
abstract Case, but rather by the valued, so-called “interpretable” features nominals bear.
The motivation for this argument comes from crosslinguistic restrictions on 1st/2nd
person and on nominals high in definiteness/animacy. Jewish Zakho (Neo-Aramaic) pro-
vides a clear illustration of such restrictions. In the simple perfective, a 3rd person object
must be accompanied by agreement if specific, (1a), but must not if nonspecific, (1b), and
a 1st/2nd person object is entirely disallowed, accompanied by agreement or not, (1c).

(1) a. ’ĀnA zwı̄n-Ā-li tlımsA. (specific object agrees)
I buy.PFV -O BJ.3 ˙ FS -S BJ .1 MS flatbread.F
‘I bought the flatbread.’
b. ’ĀnA zwın-ni tlımsA. (nonspecific object does not agree)
I ˙
buy.PFV-S BJ .1 MS flatbread. F
‘I bought some flatbread.’
c. *’ĀnA zwı̄n(-ıt)-ti (’ĀttA). (1st/2nd person obj. disallowed)
I buy.PFV (-O BJ .2 MS )-S BJ .1 MS you
Intended: ‘I bought you.’
* Thankyou to András Bárány, Jonathan Bobaljik, Chris Collins, Amy Rose Deal, Anoop Mahajan, Troy
Messick, Carson Schütze, Adrian Stegovec, Coppe van Urk, Martin Walkow, Lauren Winans, and Susi Wurm-
brand for extremely helpful discussions of and feedback on this work. Thanks also to audiences at Leipzig
University, NYU, Rutgers, UConn, UMass, USC, University of Toronto, SLE 48, and GLOW 40.

DAT have. Theoretical accounts of the PCC and DOM typically treat the two phenomena as being completely independent of each other.’ b. and often appeal to licensing conditions that exist alongside (on top of) traditional abstract Case licensing. in §4.DO (5) a. NOM 3FS . (5) (Béjar & Rezac 2003). The PCC and DOM are both extremely common crosslinguistically. NOM 2SG . (4) Canonical Strong PCC: *IO > 1/2.1 SG .a. §3 composes the core of the paper. DAT have. *Je te lui ai présent-é. (3). the DO must be 3rd person. ACC 3SG . Je la lui ai présent-é. NOM 1SG . as seen in French. Comrie 1979. ACC 3SG . The canonical strong PCC holds between an indirect and direct object (IO.). comparing the empirical profiles of the PCC and DOM and making a number of novel observations about the ways in which they are parallel.’ . surfacing in numerous unrelated languages. 1994. *Elle me lui a présent-é. DO) (Bonet 1991. Laura Kalin The pattern in (1a–b) plausibly falls under the umbrella of Differential Object Marking (Coghill 2014). (PCC violated) 1SG . ACC 3SG . Finally. I conclude by proposing that we can understand these commonalities if the burden of nomi- nal licensing is shifted to valued features. Anagnostopoulou 2003. i. clitics. canonical strong version. Béjar & Rezac 2003. the licensing needs of a nominal depend on its “interpretable” features—such features can be derivational time bombs.1 SG . I begin in §2 by exemplifying the PCC and DOM and noting some recurring features of generative syntactic accounts of these phenomena. (3) The Person Case Constraint (PCC. 2. agreement) occupy some domain. while (1c) can be seen as an extended instance of the Person Case Constraint (as argued by Doron & Khan 2012. PRES introduce-PART ‘I introduced her (DO) to him (IO). (4). in other words.3 SG .1 The PCC The PCC refers to the ungrammaticality of certain person combinations when two “weak” arguments (pronouns. DAT have.a.’ c. (2) Differential Object Marking (DOM. stated in (2). PRES introduce-PART Intended: ‘She introduced me (DO) to him (IO).): Objects high in definiteness/animacy are more likely to need to be overtly marked. Bonet 1991): In a combination of a weak DO and a weak IO. PRES introduce-PART Intended: ‘I introduced you (DO) to him (IO). Bossong 1991. The main goal of this paper is to show that the PCC and DOM are related phenomena that call for a unified account. (PCC respected) 1SG . Kalin & van Urk 2015). Brief overview of the PCC and DOM 2. i. (PCC violated) 3FS .

Anagnostopoulou 2003. (6) (Anagnostopoulou 2003. in a PCC configura- tion. Béjar & Rezac 2003. > Name > Definite > Specific Indef. allowing # to reach the direct object. Enç 1991.2 DOM DOM is a widespread phenomenon that differentiates objects into two groups—a group that gets or triggers overt marking and a group that does not (Comrie 1979. Dalrymple & Nikolaeva 2011. de Swart 2007. 2. Aissen 2003. the PCC arises because the direct object in a PCC con- figuration will never be reached by a π -probe.. it is not able to license 1st/2nd person. via an operation of clitic-doubling. while those on the right are unmarked. (7). The π -probe removes the indirect object as an intervener (e. i. An argument for valued features as derivational time-bombs Abstracting across generative syntactic accounts. We can see these scales at work in object-marking in Senaya and Palatinate German: . Woolford 1999. Walkow 2013. Bossong 1991. Torrego 1998. then. since the indirect object is itself already licensed (by Appl) and so does not actually need the licensing that is on offer from the π -probe.). (7) Person Licensing Condition (PLC. > Nonspecific Objects on the left side of the scale (often considered the “more prominent” or “less canon- ical” objects) are overtly marked. Nevins 2007. it has been argued that licensers (modeled as ϕ -probes) decompose into separate person (π ) and number (#) licensing components. de Hoop 1996. because of an additional licensing condition.. first the π -probe probes and agrees with indirect object—this is referred to as “defec- tive intervention”. (6). While the agreement configuration in (6) is able to license a 3rd person direct object. Objects are mainly differentiated along two scales. i. 2011. (8)–(9) (Silverstein 1976. This ordered probing makes it possible.. Croft 1988. (8) Animacy / person 1/2 > 3 Pronoun > Name > Human > Animate > Inanimate (9) Specificity / definiteness Pron. i. In particular. Adger & Harbour 2007. PCC effects are typically taken to arise when two arguments compete for the attention of one licenser. (6) π .a. Béjar & Rezac 2003).# IO .). and so is limited by the PLC to 3rd person.). DO In (6). Croft 1988.g. Moravcsik 1978.a.a. Comrie 1979. with π search- ing for a goal first. According to these accounts. Rezac 2008. for the two licensing components of the probe to reach different nominals. # second. Béjar and Rezac 2003) Interpretable 1st/2nd-person features must be licensed by entering into an Agree relation with an appropriate functional category.

(12) a. gets ACC) you have. Baker 2014b. (12) and (13) can derive DOM: DPs are structurally high and bear Case/ϕ - features. .a. or perhaps they in fact do get Case. agrees) we book.’ (11) Palatinate German: DOM-case. no agreement) we book. perhaps they need licensing by being adjacent to the verb (Baker 1988. while unmarked objects do not (Bhatt & Anag- nostopoulou 1996. Rodríguez-Mondoñedo 2007. Laura Kalin (10) Senaya (Neo-Aramaic): DOM-agreement. Ormazabal & Romero 2013. (non-human object. Lyutikova & Pereltsvaig 2015. Du hAS deA æmA gseje.ACC man see. and information structure (Woolford 1999. NP (unmarked) D NP N [ϕ :VAL] [Case: ] N (13) [TP T . DP (marked) b. Bhatt 2007. Baker & Vinokurova 2010. [VP V t ] ] ] Taken together. Bhatt 2007. (specific object. (13). it’s just morphologically zero (Laka 1993. i.’ b. [ marked object ..). Axnii ksuuta kasw-ox. F write. marked objects are taken (i) to have more internal structure (or features) than unmarked objects do (Danon 2006. Lidz 2006. Dalrymple & Nikolaeva 2011). F write. specificity-based (Kalin 2014) a. Abstracting across generative syntactic accounts. NOM bucket see.’ Other factors influencing DOM are affectedness (Næss 2004). Richards 2008. NPs are structurally low and do not bear Case/ϕ -features.2 SG DEM . Du hAS den mAn gseje.3FS ‘We (will) write a (specific) book. de Hoop 1996. López 2012.IMPF-S BJ .) a. Rodríguez-Mondoñedo 2007). Danon 2006. Axnii ksuuta kasw-ox-laa. (12). animacy-based (Philipp Weisser p. which makes them invisible and/or inaccessible to Case/agreement processes. López 2012.’ b..).1 PL-OBJ. Richards 2010. which makes them require Case-licensing and/or makes them visible/accessible to Case-licensing. i. Woolford 1999. Baker & Vinokurova 2010).PRT ‘You saw that man..PRT ‘You saw that bucket.c. (nonspecific object. Accounts differ as to what they take such low NP objects to need in terms of licensing—perhaps they need no li- censing at all (Massam 2001. Ormazabal & Romero 2013). gets NOM) you have.a.IMPF-S BJ . and/or (ii) to raise out of VP.1 PL ‘We (will) write a/some book.2 SG DEM.. (human object.

the direct object must be 3rd person.g.1 Third. of the two nominals involved in the phenomenon. subjects all behave uniformly—all are marked. Differential marking can appear on subjects of nominalized clauses. both the PCC and DOM crucially involve two arguments. this is explicitly stated in the definition of the phenonenon: in the context of an indirect object. for both the PCC and DOM. 2 Although these latter configurations don’t obviously involve two nominals.2 Intervention The configurational facts discussed above—that the lower nominal of two is restricted in some way—suggest that the PCC and DOM might be due to intervention of some sort by the higher nominal. and in scrambled DO>IO configurations in Slovenian (Stegovec 2015). For the PCC. An argument for valued features as derivational time-bombs 3.. Classic syntactic accounts of the PCC need indirect objects to always have a person feature (such that they always intervene for person). e. PCC effects surface in double unaccusatives in Chinook (Rezac 2011. 5). the PCC and DOM are typically taken to be independent phenomena. Classic syntactic accounts of DOM need to say that subjects always have ϕ -features/a Case feature (and so are always visible to Case/agreement). or 3rd person. Ch. the subject. both DOM and the PCC target the lower nominal. while direct objects only have a person feature (only need person licensing) when they are 1st or 2nd person. subjects are (typically) not differentiated from each other. First. in Turkish (Kornfilt 2008).1 Configuration A preliminary set of PCC/DOM commonalities has to do with the number and relative height of nominals targeted (and not targeted) by the phenomenon. The PCC restricts the lower nominal to 3rd person. 3. the higher nominal in a PCC or DOM configuration is (canonically) immune from the effect. but nonetheless. For DOM.2 3. making reference to an “object” typically implies the presence of another nominal. Second. or all are unmarked. We can test whether intervention is involved by removing the higher nominal from the triggering configuration. In DOM configurations. they do plausibly involve intervention by something like a phase boundary. the higher nominal (typically the IO) can freely be 1st. which is canonically the direct object. and on arguments of certain adjectives in Hebrew (Danon 2006). The PCC and DOM are closely related phenomena As seen in §2. while direct objects only have these features when they are high in definiteness/animacy. it is ultimately the above-described configurations that matter. . this is somewhat less obvious. while DOM requires that the lower nominal be marked if it is high in definiteness/animacy (and not otherwise). and are given very different sorts of analyses. in simple transitives in Jewish Zakho (Kalin & van Urk 2015). not the grammatical roles of the nominals involved. In contrast to the lower nominal. In strong PCC configurations. 2nd. I will make the case for treating the PCC and DOM as closely related and therefore meriting a unified analysis. 1 The fact that the higher argument in the configuration is typically immune from differential marking and person restrictions has led most of the accounts discussed in §2 to posit featural asymmetries across arguments with different grammatical roles. the brief proposal in §4 is compatible with this. In this section.

no ACC) b. (15). it is subject to differential marking. The case is murkier for DOM.3 PL see. NOM tractor see. (10). as in an active transitive clause.2 SG DEM . 3 There are some exceptions to this generalization in Hebrew and Hindi—where the internal argument can remain differentially marked even when it looks like the external argument is absent—which merit a closer look in future work. it need not be syntactically expressed. ACC-have introduce-PART ‘I introduced you (DO).’ In French. Thank you to Hagit Borer and Ethan Poole for bringing these to my attention.3 Nevertheless.F fall.’ We can see the same leveling of marking in passives in Palatinate German. (10a)) is marked (and is a subject). there is uniform marking or lack of marking). . NOM bucket and DEM . NOM bucket and DEM .PFV-SBJ. one book. the internal argument is not marked differentially (i. and the direct object can be then freely be 1st or 2nd person. Take. p. you have. the exact same nominal (boxed below) that would have been unmarked as an object (cf.PRT ‘You saw the bucket and the tractor. Laura Kalin For the PCC. (16a). This strongly suggests that there is indeed intervention by the higher nominal involved in both phenomena. 4 This example features coordinated singular nominals so that (i) case is evident on the demonstratives and (ii) there is non-3rd-singular agreement (i. agreement that is not potentially default) on the verb when the object is promoted to subject. NOM tractor be. When the internal argument stays an internal argument. Along the same lines. PASS ‘The bucket and the tractor were seen. it is clear that the effect goes away when the higher nominal is removed. First consider Senaya. if the IO is supplied from context.. An internal ar- gument that gets nominative case (unmarked) as an object.’ (Subject is non-human. a direct object clitic in simple transitives (no indirect object at all) is not restricted to 3rd person. for example. 1SG ..’ (Object is non-human. (16b) (Philipp Weisser.e. When there is no external argument. we can see that when an internal argument is not under a higher nominal.PRT AUX . removing the higher nominal involved in a PCC or DOM configuration obviates the restriction on the lower nominal.c. then. DEM . (15) Xa ksuuta mpel-aa. DeA æmA un deA trekA sin gseje wOK@. the differentiation of internal arguments disappears. the French example in (14). nevertheless participates in agreement (marking) when it is promoted to subject. but agrees) If the external argument is demoted (passives) or absent (unaccusatives). (14) Je t’-ai présent-é. the ungrammatical version in (5b).).3FS ‘A/some book fell. as removing the higher nominal (the subject) typically leads to promotion of the object. NOM 2SG .4 (16) a. Du hAS deA æmA un deA trekA gseje. cf. In the typical case.e.

(17b). However.. Taking the existence of the relevant nominal features as a starting point. there is strong morphological and semantic evidence that the feature [specific] is needed. (10). Lyons 1999. Crucially. .g. (17) a.IMPF-S BJ . Simpson et al. as specificity is often taken to be emergent. causes a crash if it is unmarked (if it fails to occur with agreement). this is still so: DOM is a generalization about objects with certain features requiring marking. so can (17a) be seen as a DOM “violation”.’ b. An argument for valued features as derivational time-bombs 3. (18). The con- figuration ruled out by the PCC is clearly featural: a weak direct object bearing 1st or 2nd person features is disallowed. Axnii xa ksuuta kasw-ox. and (iii) to ac- count for specific interpretations of in situ indefinites (Fodor & Sag 1982). (DOM “respected”) we that book. i. Wiltschko & Ritter 2014. Just as (5b)–(5c) can be seen as PCC “violations”. It is worth pausing for a moment to consider the plausibility of specificity-related fea- tures. (ii) to delimit the different types of indefinite pronouns found across languages (Haspelmath 1997). Animacy-related features are also robustly attested crosslinguistically. How can we tell that it is the object’s features that are responsible for the ungrammaticality of (17a)? If we minimally change the features of the object. it is valued features that are responsible for determining whether a particular configuration of nominals is grammatical or ungrammatical. There are two ways we can see this. Ionin 2006. then the result is ungram- maticality. *Axnii oo ksuuta kasw-ox.1 PL Intended: ‘We (will) write that book. In Senaya. Cowper & Hall 2014). (DOM “violated”) we that book. and impersonal pronouns crosslinguistically (SigurDsson & Egerland 2009.F write. based on the syntactic height of a nominal (Diesing 1992).F write. First. (i) to characterize the distribution of deter- miners and classifiers crosslinguistically (Haspelmath 1997.’ The valued features of the lower nominal modulate grammaticality in DOM configurations. cited by Mithun 1986). (5). then the sen- tence becomes grammatical. Cowper & Hall 2002.a. nominal marking in Selayarese (Finer 1997).IMPF-S BJ . which rule out different combinations of indirect object/direct object pairings based on their features. in verbal-marking in Al- gonquian languages (Piggott 1989. While it is less obvious for DOM that grammaticality is featurally-determined. like that in (17a). 2011. for example. if an object that bears those features goes unmarked.3 Valued features Both PCC effects and DOM effects are modulated by valued features on nominals. for example. Recall Senaya’s specificity-based DOM. The PCC comes in many flavors. the second way we can see that such features are crucially implicated in the PCC and DOM is that different languages care about different combinations of valued features both for the PCC and for DOM. so grammaticality comes and goes with whether the lower nominal bears those features.1 PL ‘We (will) write a book. e.) and Abkhaz (Hewitt 1979. an unambiguously specific object. Fenger 2015).

Walkow 2013): (i) the repair could target the IO.IO>1/2. (19).g. then. [human] objects marked—Palatinate German (Philipp Weisser. the DO whose features are (potentially) problematic is bolded and the repair is boxed. specific] objects marked—Spanish (Rodríguez-Mondoñedo 2007) e. whereas DOM cares only about the features of the lower nominal. (In the examples below. [definite] objects marked—Hebrew (Danon 2006) c. p.IO>1. 66).g.IO>1/2.) (20) a.IO>3. languages differ in which features.DO)—Kambera (Doliana 2013) b.DO. e.’ 5 An apparent difference here between the PCC and DOM seems to be that certain PCC configurations care about the features of the higher nominal.c.DO. (21) (Bonet 2002.g.. that this is not in fact a meaningful difference between the two phenomena. . 1994. (5)/(17)).. both can also be seen as involving “repairs” for these illicit configurations. e. are allowed and/or require marking. *3. Weak (*3. e.DO)—Romanian (Nevins 2007) DOM also comes in different flavors based on which features (or feature combinations) require the direct object to be marked. with an adposition (and DO as a strong pronoun) as in Catalan. It might be. Me-first (*2/3. 180). [animate] objects and [specific] objects marked—Kannada (Lidz 2006) What is clear from these facts is that the PCC and DOM are both modulated by language- specific restrictions on valued features on nominals. on the lower of two nominals. FUT Intended: ‘Lucy will introduce us (DO) to them (IO). such that the DO is licensed in some non-canonical way. such that the IO is licensed in some non- canonical way. Let’s start with PCC repairs.DO)—Sambaa (Riedel 2009) d.) d.IO>1/2.DO)—Greek (Anagnostopoulou 2003) c. 3. such as 3.5 (19) Some varieties of DOM a. [human. where object-marking is in fact affected by the subject’s features. 953. Rezac 2011. *Lucille nous leur présent-era.. [specific] objects marked—Senaya (Kalin 2014) b.4 Repairs Finally. or (ii) the repair could target the DO.ACC them.DAT introduce-3 SG . languages make use of several different strategies that can target the higher or lower nominal in the configuration (Bonet 1991. starting from the observation that both the PCC and DOM can be seen as ruling out “violations” of certain featural configurations (cf. Hungarian (Bárány 2015). However. In order to express the featural configurations that are ruled out by the PCC.IO>1. (PCC violation) Lucy 1PL. with an adposition (and IO as a strong pronoun) as in French. Super strong (*1/2/3. Strong (*1/2/3. (20) (Rezac 2011. Laura Kalin (18) Some varieties of the PCC a. cited by Walkow 2013. there are languages with global case and/or agreement splits.

as found in PCC repairs in Arabic). *Te m’ha recomanat per a la feina la subdirectora.’ Other DO repairs (i. often take the form of an adposition. Lucille nous présent-era à elles. *Bes-ó María.).’ (21) a.g. cliticization (e.g. it is well-known that features like specificity on an object can induce ergative marking on a subject. FUT to them ‘Lucy will introduce us (DO) to them (IO). and Eastern Ostyak (Gulya 1966. Ne inu kofe a Sione. (DOM violation) kiss-3 SG .’ 6 This is also compatible with the interpretation “the deputy director has recommended me to you for the job”. no ERG) PAST drink coffee ABS Sione ‘Sione drank coffee. (22) a.. what this looks like is differential marking of subjects. other attested repairs involve using a camouflage reflexive (e. but at first glance seem to target only the lower nominal. Turning now to DOM “repairs”. Swahili. PAST P Mary ‘He kissed Mary. (23). cited in Baker 2014a). Turkish. P. (PCC violation) 2 1-has recommended for the job the deputy.e...director ‘The deputy director has recommended you (DO) to me (IO) for the job.e.PL. like a PCC repair can. as found in PCC repairs in Georgian and Greek). (20)? While not referred to as DOM.g.’6 While repair-by-adposition seems to be the most common.director ‘The deputy director has recommended you (DO) to me (IO) for the job. DOM strategies) include case-marking (e. We see this sort of ergative marking in Niuean (Massam 2000.c.g. Senaya). An argument for valued features as derivational time-bombs b.. cf. (24).’ b. Bes-ó a María.. . 16). i. (PCC repair) Lucy 1. M’ha recomanat a tu per a la feina la subdirectora. Ne inu e Sione e kofe.ACC introduce-3 SG .’ b. like PCC repairs. German. (indefinite DO. and Hebrew). cited in Woolford 2015). p. and agreement (e. with an IO repair. (DOM repair) kiss-3 SG . Take. (PCC repair) 1-has recommended P 2 for the job the deputy.. these. Can DOM repairs also target the higher nominal in a DOM configuration.. Macedonian). the Spanish DOM repair in (22) (Rodríguez-Mondoñedo 2007.’ b. for example. ERG) PAST drink ERG Sione ABS coffee ‘Sione drank the coffee. but it is actually still determined by the lower nominal’s features. (23) a. but the DO-repair interpretation is preferred (Victoria Mateu. PAST Mary Intended: ‘He kissed Mary. and using Case with a strong pronoun (e.g. (definite DO.

and if it is removed. P.g. While the PCC and DOM have too much in common to be completely unrelated. DOM envi- ronments are those where any nominal that requires licensing (bears any valued nominal feature that is a derivational time bomb) will require a special licenser to merge for the derivation to converge. German Niuean.3 PL / SBJ . DO.1 PL ‘We put them (pots of berries) beside a big tree. (indef. if DP2 is high in definiteness/animacy. DP1 or DP2 must be specially licensed. The PCC and DOM are triggered by valued features and are general.PAST.’ Taking into account the full range of repairs above highlights this final parallel. (25) Repair lower nom. E.COM berry pick. and so any nominal with participant features will require a special licenser to merge for the derivation to converge.sister. DOM: In the context DP1 >DP2 where DP1 is a non-defective intervener.’ b. OBJ . Arabic French. What. Where this leaves us At the right level of abstraction. SBJ . no ERG) we. then. Ostyak 4. . Note that one of the reasons these parallels have failed to be drawn in the past may be that the PCC is named for the violation.DUAL . The PCC and DOM arise due to intervention: the higher argument is typically immune from the effect. with overlapping strategies of “repair” surfacing to lift the restriction. Laura Kalin (24) a. (26) a. configurational effects. exam- ples like those in (1) show that a complete unification is not possible. then DP1 or DP2 must be specially licensed. if DP2 is 1st/2nd person. PCC environments are those where only partial licensing is possi- ble. The details of such an account are left for future work. because (e. the effect goes away. while DOM is named for the repair.1 PL ‘I went to pick berries with my younger sister.PAST. Repair higher nom. (25). M@ -N@n l@G@ @ll@ juG kanNa am@GaloG. a number of commonalities between DOM and the PCC become apparent. PCC Catalan. (definite DO. b. Basque DOM Spanish. Mä t’@käk@Glämnä ula m@nGäl@m. The PCC and DOM are about licensing: both involve the lower of two arguments being restricted featurally. The PCC: In the context DP1 >DP2 where DP1 is a defective intervener.) specific nominals are not blocked in all environments that block 1st/2nd person. I propose that DOM and the PCC are conspiring to tell us that valued nominal features are what crucially matter for licensing—they can be derivational time bombs—and that certain syntactic positions are (at least partially) inaccessible to canonical licensing. ERG) we-ERG them large tree beside put. differ- entiates the PCC from DOM? The following restatements of the PCC and DOM bring out their commonalities while revealing the direction forward for an account. NOM younger.

Ana Teresa Pérez-Leroux & Yves Roberge. 1988. Mark. Judith. Massachusetts Institute of Technology. & Susana Béjar. & Lisa McNair. Mark. Cambridge. & Nadezhda Vinokurova. Bonet. 2012. Baker. Syntax 10:2–37. In The morphology-syntax connection. 1991. & Andrew Nevins. 1991. Bhatt. 2010. In Phi theory: Phi features across interfaces and modules. Georg. 33–52. The syntax of ditransitives: Evidence from clitics. Locating agreement in grammar: an argument from agreement in conjunctions. 2007. Baker. Mark C. In Romance linguistics: theory and acquisition. & Elena Anagnostopoulou. Natural Lan- guage and Linguistic Theory 21:435–483. Daniel Harbour. Bonet. 2015. 1996. 2008. Anagnostopoulou. ed. 2014b. 49–62. Object shift and specificity: evidence from ko-phrases in Hindi. & Manuel Pérez Saldanya. Baker. David Adger. Dordrecht: Springer. 143–170. David. Linguistic Inquiry 45:341–379. Doctoral dissertation. Rajesh. Béjar. Rajesh. 2003. ed. Elena. Kora Singer. Chicago: Chicago Linguistic Society. Natural Language and Linguistic Theory 28:593–642. Bárány. & Milan Rezac. Differential object marking in Romance and beyond. 2013. Types of cross-linguistic variation in case assignment. Douglas A. Eulàlia. Morphology after syntax: Pronominal clitics in Romance languages. Bhatt. Rajesh. 2014a. Chicago: University of Chicago Press. Bonet. Clitització. ed. . Joan Mascaró. Aissen. Differential object marking: Iconicity vs. 2002. The syntax and syncretisms of the Person Case Constraint. 2003. Maria-Rosa Lloret. Mark. Talk presented at McGill. Unaccusativity and case licensing. The Hague: Mouton de Gruyter. Joan Solà. Lisa M. University of Cambridge. Morphotactics: Basque auxiliaries and the struc- ture of spellout. ed. Susana. & Martin Walkow. Incorporation: A theory of grammatical function changing. 1994. 2003. Karlos. Arregi. On dependent ergative case in Shipibo and its derivation by phase. Amsterdam/Philadelphia: Benjamins. 11–22. In Linguistic variation in a minimalist framework. In New analyses in Romance linguistics. Bossong. Barcelona: Editorial Empúries. The Person-Case Constraint: a morphological approach. & Daniel Harbour. Bobaljik. volume 1. Eulàlia. 2007. Differential Object Marking in Hungarian and the morphosyntax of case and agreement. Eulàlia. MA: MITWPL 22. M. ed. András. ed. In Papers from the 32nd regional meeting of the Chicago Linguistics Society. Oxford: Oxford University Press. Jonathan. In Gramàtica del Català contemporani. Carme Picallo. 933–989. Boston. An argument for valued features as derivational time-bombs References Adger. Baker. Natural Language and Linguistic Theory 31:951–1013. England. Kibbee & Dieter Wanner. Dobrin. Where’s phi? Agreement as a post-syntactic operation. Bhatt. MA. 295–328. economy. Person licensing and the derivation of PCC effects. Two modalities of case assignment: case in Sakha. New York: Oxford Univer- sity Press. John Benjamins. Doctoral dissertation.

Haspelmath. Eastern ostyak chrestomathy. Syntax 14:297–317. & Geoffrey Khan. In Agreement in natural language: Approaches. Lingua 122:225–240. The semantics of specificity. 1980. Abkhaz. 2011. Multiple agreement and case deletion: against φ -(in)completeness. Elizabeth. case marking and direct objects. Vicki. 2015. Gulya. Contrasting Ā-dependencies in Selayarese. & Daniel Currie Hall. Paula. The syntactic manifestation of nomi- nal feature geometry. Doron. ed. & Daniel Currie Hall. Linguistics 52:335– 364. ed. János. M. Definite and animate direct objects: a natural class. Bloomington. Linguistic Inquiry 11:1–46. 1988. Gabi. 1997. Sophie Burelle & Stana Somesfalean. Syntax 3:147–163. Cowper. Linguis- tics and philosophy 5:355–398. CSLI. Differential object marking in Neo-Aramaic. Danon. Martin. New York: Garland. Linguistica silesiana 3:13–21. ed. Universität Leipzig. Noam. Fenger. On binding. Objects and information structure. Elizabeth. Eleanor. Agreement vs. Doliana. 2014. This is definitely specific: Specificity and definiteness in article systems. MIT Press. volume 90 of Linguistische Arbeits Berichte. William. Mürvet. Mary. 1997. The typology of morphological ergativity in Neo- Aramaic. 177–202. 2014. & Irina Nikolaeva. 2006. 2002. Helen. 2011. 1982. . Enç. Bernard. Natural Language and Linguistic Theory 24:977–1008. Natural Language Semantics 14:175–234. Cowper. & Ivan Sag. de Hoop. 1979. Montreal: Cahiers linguistiques de l’UQAM. Danon. Linguistic Inquiry 22:1–26. 1979. Caseless nominals and the projection of DP. One needs to be properly fertilized to bear fruit: Impersonal pronouns in Asian languages. Amsterdam: North- Holland. Janet Dean. 1966. Daniel. In Interaction in grammar. In Proceedings of the 2002 annual conference of the Canadian Linguistics Association. Indefinites. Fabian Heck & Anke Assmann. 2013. Ionin. 2006. Natural Language and Linguistic Theory 15:677–728. Lingue e Linguaggio 13:63–82. Dalrymple. Chomsky. 1991. George. Croft. 159–179. 2012. Ferguson. Lingua Descriptive Series Vol 2. Referential and quantificational indefinites. Agreement and DP-internal feature distribution. Indefinite pronouns. Fodor. 2001. Comrie. 1992. descriptions. theories. Hewitt. The super strong person-case constraint: scarcity of resources by scale-driven impoverishment. Case configuration and noun phrase interpretation. 1996. 55–66. Barlow & C. Diesing. Tania. Cambridge University Press. Molly. Finer. Laura Kalin Carstens. The features and exponence of nominal number. Aaron. Presented at the 4th UConn Linguistics Graduate Roundtable. IN: University of Indiana Press. Coghill. Gabi. Oxford: Clarendon Press. Edit.

Preminger. Andrew Carnie & Eithne Guilfoyle. & Esther Torrego. Kornfilt. Andrew. 195–211. The grammar of accusative case in Kannada. 97–116. Oxford: Oxford University Press. Natural Language and Linguistic Theory 19:153–197. In Proceedings of the Twelfth Annual Meeting of the Berkeley Linguistics Society. 1993. ed. Lyons. chapter 262-294. Åshild. Stanford: Stanford University Press. German Westphal. Alec. & Asya Pereltsvaig. 2001. Javier. in honor of joseph e. 2015. choice functions. Lyutikova. Argument structure and the morphology of the Ojibwa verb. Jonathan Bobaljik & Colin Phillips. CA. Los Angeles. Ormazabal. Luis. In Differential Subject Marking. Vassiliki Nikiforidou. Natural Language and Linguistic Theory 25:273–313. Diane. Agreement. Lidz. Simin Karimi. John Benjamins. Aspect splits without ergativity: Agreement asym- metries in Neo-Aramaic. emonds. Differential Object Marking. Glyne. Diane. 2004. DOM and two types of DSM in Turkish. Doctoral dissertation. An argument for valued features as derivational time-bombs Kalin. Mary Niepokuj. Christopher. In Papers on case and agreement 1. Laura. Marianne. 2013. ed. ed. ed. Jeffrey. 2006. 2012. MIT Working Papers in Linguistics 18. Nevins. 2007. In Theoretical perspectives on Native American languages. Gerdts & Karin Michelson. Doctoral dissertation. & Deborah Feder. ed. Cambridge. Donna B. Omer. In The syntax of verb initial languages. Edith. The representation of third person and its consequences for person- case effects. University of California. Unergatives that assign ergative. Pesetsky. & Juan Romero. Aspect and argument licensing in Neo-Aramaic. Language 82:1–23. In Phrasal and clausal architecture: Syntactic derivation and interpretation. Mary VanClay. Laura. Los Angeles. & Wendy K. Case and licensing. Cambridge. New York: State University of New York Press. Definiteness. Lingua 114:1186–1212. Mithun. ed. case and agree- ment. When zero isn’t there. Canadian Journal of Linguistics 60:289–325. In Proceedings of the 8th Eastern States Con- ference on Linguistics (ESCOL 8). unaccusatives that assign accusative. What markedness marks: The markedness problem with direct objects. López. Marantz. 2008. Helen de Hoop & Peter de Swart. Jaklin. 1989. The Tatar DP. Pseudo noun incorporation in Niuean. 1999. Wilkins. NY: CLC Publications. Itziar. & Coppe van Urk. Springer. Mas- sachusetts Institute of Technology. Vida Samiian. The syntax of valuation and the interpretability of features. 149–172. 2014. Natural Language and Linguistic Theory 33:659–702. Indefinite objects: scrambling. MA. Agreement as a fallible operation. 176–208. Benjamin Ao. Moravcsik. & Hee-Rahk Chae. Cambridge: Cambridge University Press. ed. Næss. Ekaterina. 2015. Massam. Laka. 1991. ed. David. 234–253.2:221–239. 2011. Ithaca. 2007. 79–112. and differential mark- ing. MA: MIT Press. 2000. Massam. Piggott. 1986. Kalin. VSO and VOS: aspects of Niuean word order. In Universals of human language IV: Syntax. Joseph Greenberg. 1978. . Borealis: An International Journal of Hispanic Linguistics 2.

2009. Personality disorders and missing persons: Deriving the Person- Case Constraint without Case. Cambridge. Michael. ed. 1999. Springer. 2007. Claire Moore-Cantwell. Doctoral dissertation. 203–216. Canberra: Australian Institute of Aboriginal Studies. Doctoral dissertation. NJ: Africa World Press. Otero. Laura Kalin Rezac. Radboud University of Nijmegen. M. Riedel. Letter to Noam Chomsky and Howard Lasnik on “Filters and Control. Bare classifiers and definite- ness: A cross-linguistic investigation. Norvin. Esther. Ergativity and transitivity. Netherlands. 2008. The syntax of eccentric agreement: The Person Case Constraint and absolutive displacement in Basque. Marc. Joan Maling. Richards. Milan. 2008. In Proceedings of the 40th Annual Meeting of the North East Linguistic Society. Ms. Natural Language and Linguistic Theory 26:61– 106. Stegovec. 2010. University of Connecticut. Impersonal null subjects in Ice- landic and elsewhere. 1977. MA: MIT Press. 2011. Cambridge. In New dimensions in African linguistics and languages. R. & Höskuldur Thráinsson.edu . 2011. The syntax of objects: Agree and Differential Ob- ject Marking. 2014. 2008. 1998. MIT Press. Jean-Roger. MA: GLSA. Hooi Ling Soh. Milan. Linguistic Inquiry 46:489–531. Ellen. Storrs. Walkow. A unified analysis of the Person Case Constraint and 3-3-effects in Barceloní catalan. 2015. Laura Kalin lkalin@princeton. ed. Animating the narrow syntax. Amherst. Animacy hierarchy effects on object agreement. Martin. Miguel. 112–171. The dependencies of objects. Peter. In Foundational issues in linguistic theory. & Maria Luisa Zubizarreta. Defective agree. Leiden University. Woolford. The syntax of object marking in Sambaa: a comparative Bantu perspective. Studies in Language 35:168–193. Ellen. Phi-features and the modular architecture of language. Netherlands Nijmegen. Hierarchy of features and ergativity. Dixon. Silverstein. SigurDsson. & Verner Egerland. Vergnaud. 1976. Robert Freidin. Carlos P. & Hiroki Nomoto. University of Connecticut. Netherlands Nijmegen. Halldór Ármann. Martina. Paul Kotey. 2015. Cross-linguistic variation in object marking. Rodríguez-Mondoñedo. ed. In Grammatical categories in Australian languages. Zaenen. Torrego. Natural Language and Linguistic Theory 3:441–483. Case and grammatical func- tions: The Icelandic passive. Richards. Studia Linguistica 63:158–185. Wiltschko. Linguistische Arbeits Berichte 86:137–161. Annie. Netherlands.” April 17. de Swart. 2013. Ling- Buzz/002291. Seda Kan. and the prominence of person. Andrew. Doctoral disserta- tion. volume 81 of Studies in Natural Language and Linguistic Theory. W. Adrian. Trenton. case alternations. & Elizabeth Ritter. Nijmegen. Simpson. 1985. Rezac. Kristina. Leiden. & Robert Staubs. 2009. Uttering trees. MA: MIT Press. Woolford. 2007. ed. CT.