PCC I: Two aruments against one head -analysis

PCC I: Two aruments against one head -analysis Anagnostopoulou (2005); Adger and Harbour (2007); Richards (2008) Doreen Georgi EGG 2014, Debrecen July 31, 2014 Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 1 / 41

Outline 1 What is the PCC? 2 Two arguments against one head -analysis Anagnostopoulou (2003; 2005) Discussion Adger and Harbour (2007) Richards (2008) Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 2 / 41

Definition The PCC is a constraint on the possible combinations of weak elements (clitics, agreement markers) depending on their person value and their grammatical function. interaction between IO and DO, but also found between SU and DO marked combinations lead to ungrammaticality literature: Perlmutter (1971); Kayne (1975); Bonet (1991; 1994) Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 4 / 41

Versions of the PCC (ditransitives) Strong version; local effect (Bonet 1991: 182) In a combination of a weak direct object and an indirect object [clitic agreement marker or weak pronoun], the direct object has to be 3rd person. (1) IO DO 1/2 3 * 1/2 1/2 3 3 * 3 1/2 Weak version; global effect (cf. Bonet 1991) In a combination of a weak direct object and an indirect object [clitic, agreement marker or weak pronoun], the direct object must not be 1st/2nd person unless the indirect object is 1st/2nd person, too. (2) IO DO 1/2 3 1/2 1/2 3 3 * 3 1/2 For more versions, see Haspelmath (2004); Nevins (2007). Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 5 / 41

The PCC The PCC holds in a wide range of unrelated languages (Haspelmath 2004): Arabic (Semitic), Basque (isolate), Georgian (Kartvelian), Kiowa (Kiowa-Tanoan), Kera (East Chadic), Chambala (Bantu), Hakha Lai (Tibeto-Burman), Ojibwa (Algonquian), Hausa (West Chadic), Kambera (Malayo-Polynesian), Warlpiri (Pama-Nyungan), Passamaquoddy (Algonquian), Tetelcingo Nahuatl (Uto-Aztecan),... Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 6 / 41

Strong PCC: data (3) The PCC in Greek (Anagnostopoulou 2005): a. Tha mu to stilune FUT CL.GEN.1SG CL.ACC.3SG.NEUT send.3pl They will send it to me. b. Tha FUT su CL.GEN.2SG They will send him to you. c.*tha FUT to CL.ACC.3SG.MASC su me CL.GEN.2SG CL.ACC.1SG They will introduce me to you. d.*tha FUT tu CL.GEN.3SG.MASC They will send you to him. stilune send.3pl sistisune introduce.3pl se CL.ACC.2SG stilune send.3pl 1 3 2 3 2 1 3 2 Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 7 / 41

Strong PCC: data (4) The PCC in French (Perlmutter 1971): a. On me one 1SG.DAT They will show it to me. b.*on me lui one 1SG.ACC 3SG.DAT They will show me to him. c.*on me t a montré one 1sg.acc 2sg.dat has shown one has shown me to you. le montrera 3SG.ACC show.fut.3sg montrera show.fut.3sg 1 3 3 1 2 1 Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 8 / 41

Weak PCC: data (5) Weak PCC in Italian (Doliana 2013): a. Me l ha 1SG.DAT 3SG.ACC has shown He introduced him to me. b.*mi 1SG.ACC presentato gli ha 3SG.DAT has shown He introduced me to him. presentato c. Mi ti ha 1SG.DAT/ACC 2SG.ACC/DAT has shown He introduced me to you/you to me. presentato 1 3 3 1 1/2 1/2 Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 9 / 41

Weak PCC: data (6) Weak PCC in Catalan (Bonet 1994): a. Te m ha venut el mercader més important CL.2SG.ACC CL.1SG.DAT has sold the merchant most important The most important merchant has sold you to me. b. Vi ci manderà cl.2pl.dat cl.1pl.acc send.fut.3sg S/he will send us to you(pl). 1 2 2 1 Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 10 / 41

The PCC with transitive verbs: Icelandic occurs in Icelandic quirky subject constructions restriction on the person of the object DP (see e.g. Taraldsen 1995) The Person Restriction on Nominative Objects In the presence of a dative subject, the agreeing nominative object has to be third person. Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 11 / 41

The PCC in Icelandic (7) PCC in mono-clausal quriky subject constructions: a. Henni leiddust þeir she.dat bored.3pl they.nom She was bored by them. b.*henni leiddumst vid she.dat bored.1pl we.nom She was bored by us. 3 3 3 1 Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 12 / 41

A morphological or a syntactic phenomenon? Two views on the PCC in the literature: morphological phenomenon (Bonet 1991; 1994; Miller and Sag 1997): The ungrammatical combinations are not syntactically ill-formed but violate morphotactic constraints. syntactic phenomenon: The PCC is the result of competition in syntax (Anagnostopoulou 2003; 2005; Adger and Harbour 2007; Řezáč 2008; Richards 2008). morphological account: idiosyncratic co-occurrence restrictions; does not account for the fact that the PCC follows the Silverstein-scale; models repairs of the PCC. syntactic account: independent evidence from interactions with other syntactic operations (on repair strategies see e.g. Bonet (2008); Řezáč (2008); Arregi and Nevins (2012)) Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 13 / 41

Evidence for a syntactic account The PCC arises only if Dat c-commands Nom/Abs, but not if Nom/Abs c-commands Dat (see Řezáč 2008 on French and Basque). The PCC does not arise with non-argumental datives, e.g. ethical datives ( invoke the speaker or addressee as witness or vaguely affected party, see Jouitteau and Řezáč 2008). The combination of arguments banned by the PCC is grammatical if the arguments do not Agree with the verb, e.g. in non-finite clauses in Basque. movement of the Nom/Abs across the dative can circumvent a PCC violation, e.g. in Icelandic (see Řezáč 2008). Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 14 / 41

(8) French, acc dat basis: Je le i /vous i lui j crois e i infidèle e j. I 3SG.ACC/2PL.ACC her.dat believe unfaithful I believe him/you to be unfaithful to her. (9) French, ethical dative: Demain je me vous emmène en vacances tomorrow I 1SG.DAT 2PL.ACC take in vacations Tomorrow I will take you on vacation. Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 15 / 41

Overview 1 What is the PCC? 2 Two arguments against one head -analysis Anagnostopoulou (2003; 2005) Discussion Adger and Harbour (2007) Richards (2008) Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 16 / 41

Basic idea Weak elements: spell-out of features that a functional head acquires through Agree with an argument DP. The PCC is the result of competition. It arises if two DPs have to enter into Agree with a single head, i.e in a two arguments against one head -configuration. Agree applies twice (sequentially) and is subject to a locality condition. Result: Agree between the head and the higher DP can bleed subsequent Agree with the lower DP. Depending on the features of the lower DP, bleeding can result in ungrammaticality. In this account, person/animacy hierarchies play no role. [...] an account which dispenses with hierarchies and relies exclusively on feature checking procedures. (Anagnostopoulou 2005: 201) Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 17 / 41

Evidence 1 drop of agreement in Icelandic: PCC holds only if the dative subject occupies SpecTP and the nominative object agrees with the verb (mediated by T). The combination is grammatical with default agreement. 2 PP-repair: One of the elements is realized in a PP instead of as a weak element (clitic, agreement marker). PPs are not Agree controllers. (10) PCC repair in Kiowa (Adger and Harbour 2007): a. Thalíí hegó gyá-poobóótoo boy now I.to:you.SG.him-bring.FUT I ll bring the boy to you. b. HegÓ k!yátáík!ii gyá-poobóótoo now chief I:to.him:him-bring.FUT I ll bring him to the chief. c. HegÓ k!yátáík!ii-ei em-poobóótoo now chief-loc I:you.SG-bring.FUT I ll bring you to the chief. 2 3 3 3 3 2 Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 18 / 41

Assumptions (Anagnostopoulou 2003; 2005) A functional head H bears unvalued φ-features that need to be checked (=valued) under Agree with a DP that bears matching valued features. Person and number are separate probes: H { [ π: ], [ #: ] }. Locality condition: A head agrees with the closest DP that it c-commands. Number features: Datives are defective for number; they only bear person features. Person features: local person: [+person] 3rd person nom/acc: 3rd person dat: [ person] Case filter: Every DP must be assigned case. Condition on case assignment: All φ-features of a DP must participate in Agree. Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 19 / 41

Structure (11) Structure: FP F [ π: ], [ #: ] XP DP dat [π:v] X X ZP... DP nom/acc ([π:v]), [#:V]... For locality reasons, F first agrees with the closer dative DP; in a subsequent step, it can target the lower nominative DP. F = v (ditransitives) or T (transitive) Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 20 / 41

Analysis Since the dative DP is the closest argument for F and since it always bears a person feature ([ person]), it values the person probe on F. By assumption, datives lack number features. Hence, the number probe cannot be valued in the first Agree cycle. The number probe on F targets thus targets the lower nominative/accusative DP for number and is valued. But the person feature of the lower DP cannot enter into Agree with the head; the person probe has been valued previously by Agree with the dative DP. Consequence: If the lower DP bears person features, it will not be assigned case. Consequence: Only derivations in which the lower DP does not bear a person feature (i.e. is 3rd person) will converge strong PCC The PCC is a case filter effect. Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 21 / 41

PCC in ditransitives (12) Initial structure: v v [ π: ], [ #: ] IO [π:v] ApplP Appl Appl V VP DO ([π:v]), [#:V] Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 22 / 41

PCC in ditransitives (13) Person valuation: v v [ π: ], [ #: ] IO [π:v] ApplP Appl Appl V VP DO ([π:v]), [#:V] Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 23 / 41

PCC in ditransitives (14) Number valuation: v v [π:v], [ #: ] IO [π:v] ApplP Appl Appl V VP DO ([π:v]), [#:V] Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 24 / 41

PCC in transitives (15) Structure: T T [ π: ], [ #: ] v vp ApplP SU dat [π:v] Appl Appl V VP DO ([π:v]), [#:V] Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 25 / 41

The weak PCC Weak PCC: The lower argument can be local person of the dative argument is also local person. problem so far: the lower argument can never check person proposal: A head can enter into Multiple Agree with several arguments simultaneously, i.e. person is checked simultaneously with the two DPs parameter: languages allow for multiple agree - yes or no (16) A condition on Multiple Agree: Multiple Agree can take place only under non-conflicting feature specifications of the agreeing elements. Result: 1/2 1/2 ok (multiple Agree possible) 3 1/2 ungrammatical (multiple Agree impossible) Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 26 / 41

The weak PCC But what about the combinations 1/2/3 3?? These are grammatical but shouldn t be under multiple Agree (conflicting person values). Anagnostopoulou (2005: 221): [C]ombinations of 1st, 2nd and 3rd person indirect objects with 3rd person direct objects have the same analysis as comparable clusters in languages with the Strong Version of the PCC,... these combinations do not result from multiple Agree. Is multiple Agree optional or a repair that applies whenever simple Agree would lead to a violation of the case filter (then the analysis would contain an OT-component)? no formal implementation Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 27 / 41

Summary: Anagnostopoulou (2003; 2005) The PCC is the result of competition between two DPs that Agree with the same head. The head bears only one person and one number feature. The structurally higher DP Agrees with the head in person. The lower DP must cope with what is left after the first Agree cycle. If both DPs bear person features, one of them cannot check its person features. The PCC is a case filter effect. crucial: person specifications of the arguments variation: multiple Agree available yes or no Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 28 / 41

Discussion / problems encoding of the hierarchy? 3rd person = absence of person 3rd person 3rd person (dat vs. acc/nom), its representation varies with the case of a DP If case is assigned during the derivation, how can the representation of 3rd person be fixed when the DP enters the derivation? Possible solution: Case is checked, there is a value on the DP when it enters the derivation; but: checking and valuation approach of Agree mixed two different ways to Agree: multiple Agree vs. simple Agree (a unified account would say that a local person nom/acc DP cannot check case because it does not bear person features) How exactly is the weak PCC accounted for (1/2/3 3 combinations)? The analysis cannot account for ungrammaticality of 3-3 in languages like Kambera (Doliana 2013). Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 30 / 41

Adger and Harbour (2007) strong PCC in Kiowa The account does not make use of hierarchies. The PCC follows from the distribution of φ-features and how selection/agreement works. The PCC is not a case filter effect (as in Anagnostopoulou s analysis), but the approaches share their basic idea: two arguments relate to a single head and thus cannot have the same feature (a person feature). Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 32 / 41

Feature inventory person feature: [participant:1/2/ ] number feature: [number: sg/du/pl] presence of [participant:] entails animacy ( capable of mental experience ) local person: always animate bears [participant:] 3rd person animates: unvalued [participant:]; 3rd person inanimates: absence of [participant:] feature (17) a. [participant: 1, number: singular] b. [participant: 12, number: dual] c. [participant:, number: singular] (animates) d. [number: singular] (no entailment for animacy) IOs: always interpreted as semantically animate Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 33 / 41

Analysis Appl introduce the IO; it has a selection requirement: [ PART: ] PCC: The features which a functional head requires its specifier to bear cannot be used as probes in the head s complement domain. Reason: Probe-goal relations between a head and its specifier are only possible when the head s complement domain lacks a goal that matches the probing feature(s). Due to cyclicity, the goal in the complement domain would value the probe first (cf. Abels 2003) Same idea as in Anagnostopoulou s approach: IO always checks person on a funcional head, the DO must also relate to this head if it bears a person feature, but the head provides only a single feature. Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 34 / 41

Evidence for the relevance of animacy Multidimensional scales: In the leísta dialect of Spanish, not only person but also animacy plays a role in the PCC (Ormazábal and Romero 2007): lo/la/los/las for inanimates and animates in Standard Spanish, in this dialect only for inanimates, clitics for animates = le/les; if the DO is 3rd person, the PCC arises iff the DO is animate, but not if it is inanimate. (18) PCC in the leista dialect: a. te lo di 2SG.DAT 3SG.ACC gave.1sg I gave it to you. b.*te 2SG.DAT le 3SG.ACC I gave him/her to you. di gave.1sg Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 35 / 41

on the PCC in Icelandic quirky subject constructions 3rd person = absence of person features But (quirky) datives always bear a person features (justification: to circumvent defective intervention effects datives bear a minimal φ-probe + an unvalued case feature to make the dative active) T bears a person and a number probe. The dative, being the closest argument, will value person on T. Person Agree with the lower nominative argument requires matching = non-distinctness of the probe s and the goal s features. This derives the weak PCC: ok if DO = 3rd (no person feature present), if IO =1st/2nd, then DO = 1st/2nd. strong version: previously valued features do not count; the DO must check all of its features to receive case same basic idea as before: one person probe on a head, but two arguments that want to check person; PCC = case filter effect Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 37 / 41

Analysis (19) Initial structure: TP T [ π: ], [ #: ] vp QS [DAT], [π:3 case ] v v VP V DO [π:v], [#:V] case Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 38 / 41

Bibliography I Abels, Klaus (2003): Successive Cyclicity, Anti-Locality, and Adposition Stranding. PhD thesis, University of Connecticut, Storrs, Connecticut. Adger, David and Daniel Harbour (2007): Syntax and Syncretisms of the Person Case Constraint, Syntax 10(1), 2 37. Anagnostopoulou, Elena (2003): The Syntax of Ditransitives: Evidence from Clitics. Mouton de Gruyter, Berlin. Anagnostopoulou, Elena (2005): Strong and Weak Person Restrictions: A feature checking analysis. In: L. Heggie and F. Ordoñez, eds, Clitic and Affix Combinations. John Benjamins, Amsterdam/Philadelphia, pp. 199 235. Arregi, Karlos and Andrew Nevins (2012): Morphotactics: Basque Auxiliaries and the Structure of Spellout. Springer, Dordrecht. Bonet, Eulália (1991): Morphology after Syntax. PhD thesis, MIT, Cambridge, Mass. Bonet, Eulália (1994): The Person-Case Constraint: A morphological approach, MIT Working Papers in Linguistics 22, 33 52. Bonet, Eulalia (2008): The Person-Case constraint and repair strategies. In: R. d Alessandro, S. Fischer and G. Hrafnbjargarson, eds, Areement Restrictions. Mouton de Gruyter, Berlin, pp. 103 128. Interface Explorations 15. Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 39 / 41

Bibliography II Doliana, Aaron (2013): On the Person-Case Constraint: From the Giga to the Zero Version with Copy Impoverishment and Check. BA thesis, Universität Leipzig. Haspelmath, Martin (2004): Explaining the Ditransitive Person-Role Constraint: a usage-based account, Constructions 2. Jouitteau, Mélanie and Milan Řezáč (2008): The French ethical dative, 13 syntactic tests, Bucharest Working Papers in Linguistics IX(1), 97 108. Kayne, Richard S. (1975): French Syntax: The Transformational Cycle. MIT Press, Cambridge, Mass. Miller, Philip and Ivan Sag (1997): French clitic movement without clitics or movement, Natural Lanuage and Linguistic Theory 15, 573 639. Nevins, Andrew (2007): The representation of third person and its consequences for the person-case constraint, Natural Language and Linguistic Theory 25, 273 313. Ormazábal, Javier and Juan Romero (2007): The Object Agreement Constraint, Natural Language and Linguistic Theory 25, 315 347. Perlmutter, David (1971): Deep and surface structure constraints in syntax. Holt, Rinehart and Winston, New York. Řezáč, Milan (2008): The syntax of eccentric agreement: the Person Case Constraint and absolutive displacement in Basque, Natural Language and Linguistic Theory 26. Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 40 / 41

Bibliography III Richards, Marc D. (2008): Defective Agree, Case Alternations, and the Prominence of Person. In: M. Richards and A. Malchukov, eds, Scales. Vol. 86 of Linguistische Arbeitsberichte. Universität Leipzig, Institut f Linguistik, pp. 137 161. Taraldsen, Tarald (1995): On agreement and nominative objects in Icelandic. In: H. Haider, S. Olsen and S. Vikner, eds, Studies in Comparative Germanic Syntax. Kluwer Academic, Dordrecht, pp. 307 327. Doreen Georgi (Leipzig University, IGRA) EGG 2014, Debrecen July 31, 2014 41 / 41