Roots of V-To-C Movement in Romance: Investigating Late Latin Grammar

Traces of History Conference, March 9-10, 2015

Roots of V-to-C Movement in Romance: Investigating Late Latin Grammar

André Antonelli Universidade Federal de Pelotas, Brazil [email protected]

1 Introduction b. Como posso eu caber aí? how can I to.fit there A long standing claim in the generative literature is that medieval ‘How can I fit there?’ (18th century) Romance languages are characterized as verb second (V2) grammars (see, among many others, Adams 1987 and Roberts 1993 for old French; The word order in (1) is similar to what is found in standard V2 languages, Benincà 2006 and Ledgeway 2008 for medieval Italian dialects; Ribeiro like German (see Vikner 1995). 1995 and Galves 1997 for old and classical Portuguese; Fontana 1993 and 1 Pinto 2011 for old Spanish). A much less discussed topic derived from (2) Welches Buch hat Peter gelesen? this hypothesis concerns the origin itself of the V2 phenomenon in the which book has Peter read Romance branch and, more specifically, the development of the V-to-C ‘Which book has Peter read?’ syntax. The standard view is that most of the Germanic impact on the devel- ◮ One hypothesis: the V2 property in Romance evolves from a opment of Portuguese was restricted to the lexicon, without substantial Germanic influence (see Mathieu 2009 in relation to V2 word order in old influence on syntax (see, for instance, Head and Semënova-Head 2013). French). So, how to explain the similar syntactic pattern between (1) and (2)?

◮ Problem: Not all medieval Romance varieties which manifested V2 ◮ Goal: to discuss the syntactic structure of wh-interrogative clauses in traces experienced substantial contact with languages of the Germanic late Latin.2 branch. ◮ Hypothesis: The V-to-C property seen in interrogative clauses in old ⇒ Wh-interrogative main clauses in classical Portuguese (see Lopes-Rossi Romance can be traced back to late Latin, regardless of Germanic contact. 1996).

2 (1) a. Que dizeis vós, Humildade? Here, we understand late Latin as a grammatical period extending from the 3rd century up to the 6th century. For a detailed discussion on the diﬃculties in deﬁning what say-2pl you-2pl Humility this period, see Adams (2011). ‘What do you say, Humility?’ (16th century)

1For views against the V2 hypothesis in old Romance, see, among others, Kaiser (1999), Eide (2006), Rinke (2009), Rinke and Meisel (2009).

1 2 Corpus d. quid dicit illi responsum divinum what say-3sg to.him answer divine ⇒ Primary source: Jeromes’s Latin translation of the Bible (The Vul- ‘What was God’s answer?’ (Romans 11:4) gate).3 If preverbal, subjects tend to appear before the wh-element, as illustrated ⇒ Selected books: Old Testament books of 1st and 2nd Samuel and 1st in (5). The licensing of a subject breaking the adjacency between the wh- and 2nd Kings as well as the books comprising the New Testament. phrase and the verb is quite marginal. In our corpus, we found just one example, shown in (6). ⇒ Used edition: 5th edition of the Biblia Sacra Iuxta Vulgatam Versionem (5) tu quid dicis de eo 3 The Data you-2sg what say-2sg of him ‘What have you to say about him?’ (John 9:17) One first point to be emphasized is that Late Latin is not a wh-in-situ quid scriptura language. (6) enim dicit what then scripture say-3sg (3) a. quare scidisti vestimenta tua ‘What does the scripture say?’ (Romans 4:3) why tear-PAST-2sg clothes your ‘Why have you torn your robes?’ (2nd Kings 5:8))4 N % b. quid faciet agricolis illis Wh-V-S 14 77.78 what do-FUT-3sg tenants-DAT those-DAT S-Wh-V 3 16.67 ‘What will he do to those tenants?’ (Matthew 21:40 Wh-S-V 1 5.56 c. quem vultis dimittam vobis Total 18 100 who-ACC want-2pl release-1sg to.you-2pl ‘Which one do you want me to release to you?’ (Matthew 27:17) Table 1: Distribution of overt subjects in quid-clauses A second point concerns word order differences between object wh- sentences and adjunct wh-structures. In the first group, we observe a In adjunct wh-sentences, postverbal subjects are also attested, just like in strong tendency to show subject-verb inversion. quid-clauses. (4) a. quid dixerunt viri isti (7) a. quare percussit nos Dominus hodie what say-PAST-3pl men those why defeat-PAST-3sg us Lord today ‘What did those men say?’ (2nd Kings 20:14) ‘Why did the Lord bring defeat upon us today?’ (1st Samuel b. quid faciemus et nos 4:3) what do-FUT-1pl also we b. quare succenderunt servi tui segetem meam ‘What should we do?’ (Luke 3:14) why fire-PAST-3pl servants your field my c. quid vult seminiverbius hic dicere ‘Why have your servants set my field on fire?’ (2nd Samuel what want-3sg babbler this to.say 14:31) ‘What is this babbler trying to say?’ (Acts 17:18) c. quare maledicit canis hic moriturus domino meo why curse-3sg dog this dead lord-DAT my-DAT 3This work dates from the 4th century. 4All the translations are from the New International Version (NIV). ‘Why should this dead dog curse my lord?’ (2nd Samuel 16:9)

2 d. quare fremuerunt gentes (10) a. quid enim feci why rage-PAST-3pl nations what for do-PAST-1sg ‘Why do the nations rage?’ (Acts 4:25) ‘But what have I done? (1st Samuel 29:8) b. quid ergo vult ut faciam ei However, SV structures are quite common, in particular the linear order what so want-3sg that do-1sg to.her in which the subject appears between the wh-element and the verb. ‘What can be done for her?’ (2nd Kings 4:14) (8) b. quare dominus meus ﬂet c. quid igitur faciam de Iesu why lord my weep-3sg what so do-FUT-1sg of Jesus ‘Why is my lord weeping?’ (2nd Kings 8:12) ‘What shall I do, then, with Jesus?’ (Matthew 27:22) b. quare discipuli tui transgrediuntur traditionem d. quid autem habes why disciples your break-3pl tradition what but have-2sg seniorum ‘What do you have?’ (1st Corinthians 4:7) of.elders ‘Why do your disciples break the tradition of the elders?’ In fact, not only discourse markers appear as intervening material in quid- (Matthew 15:2) clauses. We also ﬁnd 6 examples of clitic-like elements, particularly non- nominative personal pronouns (see (11)), and only one example of fronted c. quare nos non potuimus eicere eum XP (see (12)), clearly a quite marginal case. As a matter of comparison, why we not can-PAST-1pl drive.out it discourse markers appear as breaking constituents in 16 examples. ‘Why couldn’t we drive it out?’ (Mark 9:28) (11) quid tibi dixit Heliseus The word order in which the subject precedes quare is attested in only one what to.you say-PAST-3sg Elisha example, shown in (9). ‘What did Elisha say to you?’ (2nd Kings 8:14) (9) tu quare spernis fratrem tuum (12) quid in via tractabatis you-2sg why despise-2sg brother your what in road argue-PAST-2pl ‘Why do you look down on your brother?’ (Romans 14:10) ‘What were you arguing about on the road?’ (Mark 9:33)

In quare-sentences, we also find discourse markers breaking the linear ad- N % jacency between the wh-operator and the verb. Wh-V-S 9 45 S-Wh-V 1 5 (13) a. quare ergo peccas in sanguine innoxio Wh-S-V 10 50 why so sin-2sg in blood innocent Total 20 100 ‘Why then would you do wrong to an innocent man?’ (1st Samuel 19:5) Table 2: Distribution of overt subjects in quare-clauses b. quare ergo contempsisti verbum Domini why so despise-PAST-2sg word of.Lord There is another striking difference between quid and quare-sentences. In ‘Why did you despise the word of the Lord?’ (2nd Samuel 12:9) the former, even though subjects cannot break the wh-adjacency with the finite verb, we attest that some coordinated conjunctions, usually called However, structures with quare impose no restriction on the presence of discourse markers (see Kroon 1998) can appear as intervening constituents. fronted XPs between the wh-phrase and the verb.

3 (14) a. quare iuxta murum accessistis (15) ForceP why near wall approach-PAST-2pl ‘Why did you get so close to the wall?’ (2nd Samuel 11:21) Forceo TopP b. quare hoc fecisti why that do-PAST-2sg ‘Why do you behave as you do?’ (1st Kings 1:6) Topo IntP c. quare cum publicanis et peccatoribus manducat why with publicans and sinners eat-3sg magister vester Into FocP teacher your-2pl ‘Why does your teacher eat with tax collectors and sinners?’ o (Matthew 9:11) Foc TopP d. quare in parabolis loqueris eis why in parables speak-2sg to.them Topo FinP ‘Why do you speak to the people in parables?’ (Matthew 13:10)

⇒ A summary of the main facts: Fino IP quid-clauses quare-clauses 1. Wh-phrase in the left periphery X X We propose that quid and quare are positioned in the CP-domain, in wh 2. Wh-V-S word order X X accordance with the fact that late Latin is a -movement language. 3. Wh-S-V word order X ⇒ Analysis of quid-clauses: 4. Intervening discourse markers X X 5. Fronted XPs as intervening elements X • Both the wh-phrase and the verb are moved to the left periphery, o Table 3: Summary of main diﬀerences and similarities between quid and the former reaching SpecFocP and the latter being raised to Foc quare-clauses. (as a result of the Wh-Criterion, for instance), in a paradigm similar to what has been proposed for object wh-clauses in Italian (see Rizzi 1997).

• The subject either remains in a lower layer (SpecIP, for instance), thus accounting for the Wh-V-S word order, or is generated in a Topic position 4 The Analysis above FocP, thus deriving the S-Wh-V sequence.

In our analysis, we assume a split-CP view (see Rizzi 1997, 2001, 2004), • The Spec-head relation between quid and the ﬁnite verb also explains as the one schematized in (15). the impossibility of fronted intervening XPs.

4 • We assume that intervening non-nominative personal pronouns are clitic (18) Perché, il mio libro, Gianni lo ha portato via? elements adjoined to the verbal constituent.5 In this case, their presence to ‘Why, my book, Gianni took it away?’ the left of the verb does not contradict the hypothesis of V-to-C movement. • We propose that quare is in SpecIntP, without manifesting a Spec-head ⇒ Analysis of quare-clauses: relation with the verb, since in this context there would be no V-to-C movement. As in Italian, this proposal derives the possibility of preverbal • We propose that quare functions as perché “why” in Italian. subjects and the presence of fronted XPs breaking the adjacency between quare and the ﬁnite verb. Rizzi (2001) shows that perché does not target SpecFocP because its oc- • Concerning postverbal subjects, as shown in (7), it could be assumed curence is compatible with the presence of a focalized constituent to its 7 right, as the contrast in (16) exempliﬁes. He proposes the existence of that they are in-situ, remaining inside the VP-layer. an additional layer above FocP specialized in hosting wh-operators like 6 ⇒ perché, namely IntP. What about intervening discourse markers?

(16) a. Perché QUESTO avremmo dovuto dirgli, no qualcos’altro? In adjunct wh-clauses, the presence of an intervening discourse marker is ‘Why this we should have said to him, not something else?’ not a problem, since we could assume that it is located in the CP-domain, b *QUESTO perche avremmo dovuto dirgli, non qualcos’altro? occupying either a spec position or a head position below IntP. However, in object wh-sentences, this would not fit with the hypothesis that quid For Rizzi, the position of perché in the CP-domain is SpecIntP. In addition and the verb are in a Spec-head relation, thus weakening the hypothesis to that, he also proposes that V-to-C movement does not apply in this of V-movement to the CP-field in these clauses. type of wh-structure. Such a claim would explain why the SV word order perché A point to be noted is that discourse markers behave as second position is possible in -clauses (see (17)). Since the verb is not in the left 8 periphery, a subject in SpecIP would occupy a position structurally higher clitics. than that where the verb is landed. (19) a. quid ergo dicemus (17) Perché Gianni è venuto? what so say-FUT-1pl ‘Why has Gianni left?’ ‘What shall we say, then?’ (Romans 6:1) b. tu ergo quid dicis A second advantage of this kind of analysis is that it accounts for why you-2sg so what say-2sg fronted XPs are allowed to appear between perché and the finite verb (see ((16a) and (18)). Again, since the verb is not in a Spec-head relation ‘Now what do you say?’ (John 8:5) with the wh-operator, specialized positions for a focus element or topic constituents can be activated below IntP. The literature presents a lot of accounts for second position clitics (see, 5 For a view that classical Latin already had non-nominative personal clitic pronouns, among many others, Anderson 1993, Halpern 1995, Bošković 2000, Legate see Adams (1994). 6Rizzi shows that come mai “how come” behaves like perché, as confirmed by the 2008) and it is not our aim here to discuss this phenomenon in detailed pair of sentences in (i) and (ii). technical terms. In the case of discourse markers in late Latin, we propose i. Come mai IL MIO LIBRO gli ha dato, non il tuo? three assumptions: ‘How come MY BOOK you gave to him, not yours?’ 7 It is beyond the scope of this paper to discuss why subjects occur pre and postver- ii. *IL MIO LIBRO come mai gli ha dato, non il tuo? bally. One possibility would be to say that discursive factors are at stake, in the same ‘How come MY BOOK you gave to him, not yours?’ line of what Belletti (2001, 2004) has proposed for the alternation SV/VS in Italian declarative clauses. 8See Adams (1994) for other possible second position clitics in Latin.

5 (20) i. As second position clitics, discourse markers must have 5 Conclusion phonological material to their left; • Our results show that the process of V-to-C movement is already present ii. discourse markers are left-adjoined to the highest head in an specific type of wh-interrogative in late Latin. preceded by phonological material; iii discourse markers are the last elements to adjoin to their target • This is particularly welcome considering those medieval Romance va- head. rieties which, despite showing V2 traces, did not undergo a substantial impact from Germanic languages. Our findings show that V2 properties ◮ Derivation of (19a): in Romance can be traced back to late Latin, a more natural predecessor than the hypothesis of Germanic influence. We could say that the discourse marker is adjoined to the finite verb in Foco. This would apply after V-to-C movement. References (21) FocP Adams, J. N. (1994). “Wackernagel’s Law and the Position of Unstressed Personal Pronouns in Classical Latin.” Transactions of the Philo- quid logical Society, 92(2), 103-178. ergo-V-Foco . . . Adams, J. N. (2011). “Late Latin.” In J. Clackson (ed.), Latin Language. Oxford: Wiley-Blackwell, 257-283. ◮ Derivation of (19b): Adams, M. (1987). Old French, Null Subjects, and Verb Second Phenom- ena. PhD dissertation, UCLA. We could say that the subject pronoun tu “you" occupies the specifier of a Anderson, S. R. (1993). “Wackernagel’s Revenge: Clitics, Morphology, and topic position above FocP. Under this configuration, the discourse marker the Syntax of Second Position.” Language, 68-98. would be adjoined to the null head of this topic projection. Belletti, A. (2001). “‘Inversion’ as Focalization.” In A. Hulk & J.-Y. (22) XP Pollock (eds..), Subject Inversion in Romance and the Theory of Universal Grammar. Oxford: Oxford University Press, 60-90. Belletti, A. (2004). “Aspects of the Low IP Area.” In L. Rizzi (ed.), The tu Structure of CP and IP. The Cartography of Syntactic Structures, ergo-Xo FocP vol. 2. New York: Oxford University Press, 16-51. Benincà, P. (2006). “A Detailed Map of the Left Periphery of Medieval Romance.” In R. Zanuttini, H. Campos, E. Herburger & P. Portner quid (eds.), Crosslinguistic Research in Syntax and Semantics. Negation, o Tense, and Clausal Architecture. Washington, D.C.: Georgetown V-Foc . . . University Press, 53-86. In both cases, the discourse marker meets the requirements proposed for Biblia Sacra Iuxta Vulgatam Versionem. (1994). Stuttgart: Deutsche their structural derivation, namely, to have phonological material to their Bibelgesellschaft, 5th edition. left, to be adjoined to the highest head preceded by phonological material Bošković, Ž. (2000). “Second Position Cliticisation.” In F. Beukema and and to be the last element to adjoin to their target head. The point M. den Dikken (eds.), Clitic Phenomena in European Languages. to be emphasized, however, is that these second position clitics are fully Amsterdam/Philadelphia: John Benjamins, 71-120. integrated with the hypothesis of V-to-C movement in quid-clauses.

6 Eide, K. (2006). Word Order Structures and Unaccusative Verbs in Clas- (eds.), Proceedings of the Workshop “Null-subjects, Expletives and sical and Modern Portuguese. PhD dissertation, University of Oslo. Locatives in Romance”. Arbeitspapier 123. Fachbereich Sprachwis- Fontana, J. (1993). Phrase Structure and the Syntax of Clitics in the senschaft, 93-130. History of Spanish. PhD dissertation, University of Pennsylvania. Rizzi, L. (1996). “Residual Verb Second and the Wh-criterion.” In A. Galves, C. (1997). “Do Português Clássico ao Português Europeu Mod- Belleti & L. Rizzi (eds.), Parameters and Functional Heads. New erno: uma Análise Minimalista.” Estudos Linguísticos e Literários, York: Oxford University Press, 63-90. 19, 105-128. Rizzi, L. (1997). “The Fine Structure of the Left Periphery.” In L. Haege- Halpern, A. (1995). On the Placement and Morphology of Clitics. Center man (ed), Elements of Grammar. Dordrecht: Kluwer, 281-337. for the Study of Language (CSLI). Rizzi, L. (2001). “On the Position ‘Int(errogative)’ in the Left Periphery Head, B. F. and Semënova-Head, L. (2013). “Vestígios da Presença Sueva of the Clause.” In G. Cinque & G. Salvi (eds.), Current Studies no Noroeste da Península Ibérica: na Etnologia, na Arqueologia e in Italian Syntax: Essays oﬀered to Lorenzo Renzi. Amsterdam: na Língua.” Revista Diacrítica, 27(2), 257-277. Elsevier, 287-296. Kaiser, G. (1999). “A Ordem das Palavras e a Posição do Verbo Finito Rizzi, L. (2004). “Locality and Left Periphery.” In A. Belletti (ed.), Struc- no Português Antigo.” In P. Ferenc (ed.), Actas do Congresso In- tures and Beyond: the Cartography of Syntactic Structures, vol. 3. ternacional por Motivo dos Vinte Anos do Português no Ensino New York: Oxford University Press, 223-251. Superior. Budapest: Departamento de Língua e Literatura Por- Roberts, I. (1993). Verbs and Diachronic Syntax. Dordrecht: Kluwer. tuguesas da Faculdade de Letras da Universidade Eötvös Loránd The Holy Bible. New International Version de Budapeste, 248-261. . (1984). Grand Rapids: Zon- dervan. Kroon, C. (1998). “A Framework for the Description of Latin Discourse Markers.” Journal of Pragmatics, 30(2), 205-223. Vikner, S. (1995). Verb Movement and Expletive Subjects in the Germanic Languages. New York: Oxford University Press. Ledgeway, A. (2008). “Satisfying V2 in Early Romance: Merge vs. Move.” Journal of Linguistics 44, 437-470. Legate, J. A. (2008). “Warlpiri and the Theory of Second Position Clitics.” Natural Language & Linguistic Theory, 26(1), 3-60. Lopes-Rossi, M. A. (1996). A Sintaxe Diacrônica das Interrogativas-Q do Português. PhD dissertation, UNICAMP. Mathieu, E. (2009). “On the Germanic Properties of Old French.” In P. Crisma and G. Longobardi (eds.), Historical Syntax and Linguistic Theory. New York: Oxford University Press, 344-357. Pinto, C. F. (2011). Ordem de Palavras, Movimento do Verbo e Efeito V2 na História do Espanhol. PhD dissertation, UNICAMP. Ribeiro, I. (1995). A Sintaxe da Ordem no Português Arcaico: o Efeito V2. PhD dissertation, UNICAMP. Rinke, E. (2009). “Verb Placement in Old Portuguese.” In A. Dufter & D. Jacob (eds.), Focus and Background in Romance Languages. Amsterdam/Philadelphia: John Benjamins, 309-332. Rinke, E. and Meisel, J. (2009). “Subject-inversion in Old French: Syn- tax and Information Structure.” In G. Kaiser and E-M Remberger