The syntax and prosody of Focus: the Bantu-Italian connection

 Ibe ia: An In e na ional Jou nal o Theo e ical Linguis ics ol 2.1, 2010, 131–168 pp. h p://www.si .us.es/ibe ia/index.php/ij/index ISSN 1989–8525 The Syn ax and P osody o Focus: he Ban u-I alian Connec ion 1 Ma ía Luisa Zubiza e a zubiza @usc.edu Uni e si y o Sou he n Cali o nia Abs ac : This pape p o ides an o e iew o he li e a u e on he syn ax and p osody o ocus in some o he Ban u languages (Kima uumbi, Chimwiini, Chichewa) and in I alian, and i a gues ha , despi e hei ypological dis ance, hey sha e much in common wi h espec o bo h he syn ax and p osody o ocus: 1) bo h language ypes ha e an ac i e low Focus posi ion (Belle i 2004, Aboh 2007); and 2) he Focus posi ion igge s he inse ion o a s ong p o sodic bounda y, which gi es ise o a ‚ ipple e ec ‛ in ha ph ases o he igh o Foc a e simila ly lanked by a compa able p osodic bounda y. The iew ou lined he e a gues in a o o a s onge syn ax-p osody connec ion han is gene ally ecognized in cu en app oaches. Keywo ds: p osodic ph asing, s ong p osodic bounda y, weak p osodic bounda y, ocus, high ocus posi ion, low ocus posi ion, unc ional domain, lexical domain, p osodic ema gina ion, igh -disloca ion. Resumen: Es e a ículo analiza los p incipales abajos sob e la sin axis y p osodia del oco en lenguas ban úes (Kima uumbi, Chimwiini, Chichewa) y en i aliano, p oponiendo que, a pesa de la dis ancia ipológica, ambos ipos de lengua compa en muchas de las p opiedades que ca ac e izan a la sin axis y p osodia del oco: 1) Compa en la exis encia de una posición in e io de Foco ac i a (Belle i 2004, Aboh 2007); y 2) es a posición de Foco equie e la inse ción de una ma ca p osódica ue e, que causa un ‚e ec o 1 This pape was de eloped wi hin he amewo k o a semina augh conjoin ly wi h And ew Simpson (on he in e ace o p osody and syn ax) a USC in he Sp ing o 2010. I hank he pa icipan s o his semina , as well as he audience o he USC Syn ax Plus G oup, he audience a he 40 h Linguis ics Symposium on Romance Languages a he Uni e si y o Washing on (Sea le) and he Sp ing 2010 Linguis ics Semina a Consejo Supe io de In es igaciones Cien í icas/Cen o de Ciencias Humanas y Sociales (Mad id). Ve y special hanks o Anna Ca dinale i o gene ously p o iding da a and insigh s on I alian.  Ibe ia: An In e na ional Jou nal o Theo e ical Linguis ics ol 2.1, 2010, 131–168 pp. h p://www.si .us.es/ibe ia/index.php/ij/index ISSN 1989–8525 132 The Syn ax and P osody o Focus: he Ban u-I alian Connec ion onda‛ de mane a que los cons i uyen es si uados a la de echa de Foc es a án igualmen e lanqueados po una ma ca p osódica simila . El análisis p esen ado aquí apoya la exis encia de una conexión en e la sin axis y la p osodia mucho más es echa de lo que gene almen e se asume en los es udios ele an es. Palab as cla e: aseo p osódico, ma ca p osódica ue e, ma ca p osódica débil, oco, posición supe io de oco, posición in e io de oco, dominio uncional, dominio léxico, ‘ma ginación’ p osódica, dislocación a la de echa. Resumo: Es e a igo ap esen a uma isão ge al da li e a u a sob e a sin axe e p osódia de Foco em algumas línguas ban as (Kima uumbi, Chimwiini, Chichewa) e no i aliano, e de ende que, apesa da sua dis ância ipológica, es as pa ilham aspec os no que diz espei o à sin axe e à p osódia de oco: 1) ambos os ipos de línguas êm uma posição de Foco baixa ac i a (Belle i 2004, Aboh 2007); 2) a posição de oco desencadeia a inse ção de uma on ei a p osódica o e, que d{ o igem a um ‚e ei o onda‛ no qual os sin agmas à di ei a do Foco são igualmen e acompanhados po uma on ei a p osódica compa á el. A isão aqui expos a a gumen a a a o de uma conexão sin axe-p osódia mais o e do que aquela ge almen e econhecida em abo dagens co en es. Pala as-cha e: aseamen o p osódico, on ei a p osódica o e, on ei a p osódica aca, oco, posição de oco al a, posição de oco baixa, domínio uncional, domínio lexical, ma ginalização p osódica, deslocação à di ei a. 1. In oduc ion This pape has wo goals. Fi s , i aims o highligh he simila i ies be ween I alian and ce ain Ban u languages (Kima uumbi, Chimwiini, and Chichewa) wi h ega ds o he syn ax and p osody o ocus. Aboh (2007) has al eady p oposed o ex end o Ban u he ype o s uc u es p oposed by Belle i (2004) o I alian. Al hough Aboh’s p oposal is a e y insigh ul one, a p ecise syn ac ic compa ison o he ocused s uc u es in he wo language ypes s ill emains o be a icula ed, as well as he e ec s o hose s uc u es on p osodic ph asing. This pape is an a emp o begin o ill ha gap. 2 Second, as I ad ance a p oposal o he syn ax-p osody mapping o he languages unde discussion, I 2 As he eade will no ice hough, he e is a ce ain imbalance be ween he dep h o ou knowledge ega ding he syn ax o he Ban u languages unde discussion and he syn ax o I alian. A his poin , we know much mo e abou he la e han we know abou he o me .  Ibe ia: An In e na ional Jou nal o Theo e ical Linguis ics ol 2.1, 2010, 131–168 pp. h p://www.si .us.es/ibe ia/index.php/ij/index ISSN 1989–8525 133 Ma ía Luisa Zubiza e a p opose a mo e di ec syn ax-p osody connec ion han is gene ally assumed in P osodic Phonology (e.g. T uckenb od 1999, Kane a 1990, F asca elli 1999, 2000 o he languages unde discussion). 3 I has long been ecognized ha ules ha a ec he du a ional na u e o segmen s wi hin a wo d, as well as ules ha a ec he hy hmical o ganiza ion o wo ds, a e sensi i e o he p esence and s eng h o p osodic bounda ies. On he o he hand, he ques ion o wha is he algo i hm ha de e mines he posi ioning o such junc u es is s ill e y much unde deba e. P osodic phonology assumes ha he e exis s a p osodic cons i uen hie a chy abo e he p osodic wo d, which consis s o he phonological ph ase (di ided in o mino and majo ph ases), he in ona ional ph ase, and he u e ance ph ase. Wi hin he Edge-Alignmen heo y o P osodic Phonology (p oposed by Selki k 1986, 1995, T uckenb od 1999), i is assumed ha le o igh edges o P-ph ases (depending on he language) a e hen aligned wi h he le o igh edges o syn ac ic ph ases. I u he mo e assumes a gene al condi ion ( he Lexical Ca ego y Condi ion) ha igno es unc ional ca ego ies; only lexical ca ego ies a e aligned wi h p osodic cons i uen s. The iew ha we pu o h he e will no assume he exis ence o p osodic cons i uencies (independen o he syn ax), which a e hen aligned o ma ched wi h syn ac ic s uc u e. Fu he mo e, we allow o he exis ence o g adien p- ph asing, as de ined by s eng h o p-bounda ies, and we c ucially assume ha unc ional ca ego ies a e isible o he algo i hm ha inse s p-junc u es. In pa icula , we will a gue ha he Focus head (which is a unc ional head) can igge inse ion o a p-bounda y. The p oposal we pu o h he e is en a i e o he ex en ha he da a unde discussion s ill awai s an in-dep h in ona ional s udy, bu I hope i will se e he pu pose o os e ing u he hough s and discussions on his e y impo an and ascina ing opic. Indeed, i phonology can cue ce ain aspec s o he syn ax, he implica ions o syn ac ic acquisi ion and p ocessing a e imminen ly signi ican . The pape is o ganized as ollows. In 2, we in oduce some p elimina ies on he syn ax and p osody o ocus. In 3, we discuss he Ban u da a, p ima ily 3 Fo in luen ial wo ks in P osodic Phonology, see Selki k (1986, 1995), Selki k & Shen (1990), Nespo & Vogel (1986), Inkelas & Zec (1990), and e e ences ci ed he ein.  Ibe ia: An In e na ional Jou nal o Theo e ical Linguis ics ol 2.1, 2010, 131–168 pp. h p://www.si .us.es/ibe ia/index.php/ij/index ISSN 1989–8525 134 The Syn ax and P osody o Focus: he Ban u-I alian Connec ion Kima uumbi and Chichewa (based on wo k by Odden 1996, Kane a 1990), wi h sho e e ences o Chi umbuka (Downing 2006) and Chimwiini (Kissebe h 2010), and p opose a eanalysis o T uckenb od ’s (1999) in luen ial p oposal (see he Appendix o a sho summa y and discussion o his wo k). In 4, we ex end he analysis p oposed o Ban u o he I alian da a (as desc ibed by Ca dinale i 2002 and F asca elli 1999, 2000). In 5, we commen b ie ly on Spanish and i s place in he gene al ypology o p-ph asing. In 6, we conclude and d aw he implica ions o he p oposed analysis ega ding he domain o applica ion o p-junc u e inse ion. 2. Some P elimina ies on he Syn ax and P osody o Focus Clausal s uc u e is assumed o be di ided in o h ee main ields: he high CP ield, he in e media e In lec ional (IP) ield, and he low e bal ( P) ield. Wi hin his iew o syn ax, i has long been ecognized by Romance syn ac ians ha he e exis s a high ocus posi ion, immedia ely abo e IP, which is ac i e in many o he Romance languages (see He nanz and B uca 1987, Cinque 1990, Zubiza e a 2001, Rizzi 1997, Belle i 2004, among many o he s). (1) Qualcosa, di sicu o, io a ò. (Cinque 1990: 15) SOMETHING, su ely, I will-do Aboh (2007) has shown ha he Kwa languages and some o he Ban u languages (all SVO languages, like I alian) also ha e an ac i e high ocus posi ion (example below is om Gungbe, a Kwa language). (2) Mó ò w Dòsú kù wá. (Aboh 2007) Ca Foc Dosu d i e come ‘Dosu came by CAR’. Mo e no el om he Romance poin o iew is he exis ence o a low ocus posi ion, immedia ely abo e he P (Belle i 2004). As shown by Aboh (2007), he Ban u languages p o ide s ong e idence o such a posi ion. See he examples om Aghem (ci ed by Aboh 2007: 90 and a ibu ed o Biloa 1997), whe e he ocused ph ase appea s immedia ely o he igh o he e b. As Aboh shows, his is no a Case- ela ed posi ion and any ype o ph ase (including Ad e bs) appea s in his posi ion when ocused. (3) a. A mɔ zi n ɔ bɛ -ˈkɔ {n ʻsóm Expl Pas ea Inah u u in a m ‘INAH a e u u in he a m.’  Ibe ia: An In e na ional Jou nal o Theo e ical Linguis ics ol 2.1, 2010, 131–168 pp. h p://www.si .us.es/ibe ia/index.php/ij/index ISSN 1989–8525 135 Ma ía Luisa Zubiza e a b. Fil a-mo-zi ang wo be’-ko F iends SM-P2-ea wi h hands u u ‘I was wi h hei HA NDS ha he iends a e u u.’ c. T -b -b gh| mɔ z n bɛ -kɔ Dogs wo Pas ea oday u u ‘The wo do gs a e u u TODAY.’ I he low ocused posi ion is immedia ely abo e P, hen i mus be ha he Ve b has mo ed ou o he P in o he In lec ional ield. Aboh sugges s ha in Aghem, whe e ense is a ee-s anding mo pheme, he Ve b mo es o Aspec , which is also pa o he In lec ional ield, al hough dis inc om Tense. He e we will abs ac away om he ine-g ained s uc u e o he In lec ional ield (which con ains Tense, Mood, and Aspec p ojec ions), and we will si ua e he Ve b unde he gene ic ca ego y In l o all he Ban u languages unde discussion (an assump ion o which Kima uumbi p o ides some compelling e idence). The s uc u e below shows he wo syn ac ic posi ions o ocus: he high ocus, abo e IP, and he low ocus, be ween IP and P. A language wi h a syn ac ically ac i e high o low Foc posi ion mo es he ocused cons i uen o he sen ence o he Spec o Foc o e ly. Languages wi h a syn ac ically non- ac i e Foc posi ion can be assumed o mo e he ocused cons i uen o Spec o Foc ‚co e ly‛. As no ed by Aboh, many Ban u languages ha e an o e mo phological ma ke o Foc. In he case o Gungbe, his mo pheme is w  ; see (2) abo e. (4) [ Spec [ Foc [ Spec [ In l [ Spec [ Foc [ VP Kima uumbi, a Ban u language spoken in Tanzania and analyzed by Odden (1996), p o ides an in e es ing a gumen in a o o he iew ha he low ocus posi ion is selec ed by he in lec ional node Tense. I is no ed he e ha in Kima uumbi, ocal dis inc ions a e ma ked in he ense sys em, which is a undamen al pa o he in lec ional sys em o Kima uumbi e bs (Odden 1996: 60–66). Mo e p ecisely, he ype o ense may impose pa icula ocusing equi emen s on he clause. Thus, noun- ocal enses equi e ha some non- e b in he clause be he ocus o he clause, and he ocused cons i uen mus appea immedia ely a e he e b. On he o he hand, e b- ocal enses selec he e b as he ocused cons i uen o he clause. Such selec ional ela ions be ween ype o ense and ocus ype is immedia ely cap u ed by he syn ac ic s uc u e in (4), whe e he low Foc posi ion is selec ed by Tense, and Tense is pa o he ine-  Ibe ia: An In e na ional Jou nal o Theo e ical Linguis ics ol 2.1, 2010, 131–168 pp. h p://www.si .us.es/ibe ia/index.php/ij/index ISSN 1989–8525 136 The Syn ax and P osody o Focus: he Ban u-I alian Connec ion g ained s uc u e o In l. In his pape , we will be mainly conce ned wi h he low Foc posi ion. On he p osodic side, i is known ha he e a e wo main ways o ealizing p ominence a he pos -lexical le el: 1) culmina i ely, ia main sen ence s ess (which in ol es associa ing he hy hmically s onges elemen o a me ical s uc u e wi h he Nuclea Pi ch Accen o NPA); and 2) dema ca i ely, ia a p osodic junc u e (ma ked ia leng hening, pauses, and/o dedica ed ph asal ones). Ge manic is known o be o he i s ype, and Japanese and Ko ean o he la e ype. Some languages ma k pos -lexical p ominence by bo h means (Bengali). See Jun (2005) o de ailed discussion. Many Ban u languages clea ly belong in he dema ca i e ype o ca ego y. This becomes pa icula ly clea in he manne in which ocus is p osodically ma ked in hese languages. As we will see, ocus igge s he inse ion o a p osodic bounda y, namely a igh edge p osodic bounda y. As ema ked by Downing (2006, 2008), in many o he Ban u languages unde discussion, he inal o penul ima e syllable o a phonological ph ase (o p- ph ase) is long. And he one in he las p-ph ase is he longes and can he e o e be iden i ied wi h main sen ence (o Nuclea ) s ess. Ye , i is he p-bounda y, and no he Nuclea S ess which se es o p osodically iden i y he ocused cons i uen . As o I alian, we will a gue ha i is bo h a culmina i e and a dema ca i e ype o language, and ha he e ec s o ocus on p-ph asing ound in some o he Ban u languages a e ema kably simila o hose ound in I alian. We p opose ha in de ining he domain o p-bounda y inse ion, i is use ul o dis inguish be ween wo dis inc syn ac ic domains: he Fu(nc ional) domain and he L(lexical) domain. A syn ac ic cons i uen belongs o he L- domain o lexical head H i i is L- ela ed o H and i belongs o he Fu-domain i i is Fu- ela ed o unc ional head F, as summa ized below. 4 (5) a. XP is L- ela ed o head Y i XP is a speci ie o , complemen o , o adjunc o Y, whe e Y heads a lexical p ojec ion YP. b. XP is Fu- ela ed o unc ion head Fu i i is a speci ie o Fu o adjoined o FuP. 4 A p ecu so o his idea is ound in Nespo & Vogel (1986), who sugges ha opics and p e e bal subjec s a e lanked by an in ona ional ph ase bounda y, unlike pos e bal complemen s, which a e lanked by a phonological ph ase bounda y.  Ibe ia: An In e na ional Jou nal o Theo e ical Linguis ics ol 2.1, 2010, 131–168 pp. h p://www.si .us.es/ibe ia/index.php/ij/index ISSN 1989–8525 137 Ma ía Luisa Zubiza e a Fo languages ha make a dis inc ion be ween weak and s ong p- bounda ies, weak p-bounda ies ge inse ed a he edge o L- ela ed ca ego ies and s ong p-bounda ies ge inse ed a he edge o Fu- ela ed ca ego ies. Following a no a ion p oposed by T uckenb od (1999), we will use a single (le o igh ) p-b acke o anno a e a weak p-bounda y and a sequence o le and igh p-b acke s o anno a e a s ong p-bounda y. (6) a. Weak p-bounda ies: ...XP ) YP) ZP) o (XP (YP (ZP... b. S ong p-bounda ies: ...XP) ( YP) (ZP... As we shall see, he (low) Foc head in (4) igge s he inse ion o a p- bounda y in Ban u and in I alian. Being a unc ional ca ego y, i igge s he inse ion o a s ong p-bounda y. Fu he mo e, he Foc- igge ed p-bounda y gi es ise o a ipple e ec , such ha a s ong p-bounda y ge s copied a he le edge o each ph ase ha ollows he ocused cons i uen . 3. P osodic ph asing in some Ban u languages We discuss i s he case o Kima uumbi (as desc ibed by Odden 1996), hen Chimwiini (as discussed by Kane a 1990), and inally Chichewa (as desc ibed by Kane a 1990). Fo each o he languages, we discuss he p- ph asing in wide ocus con ex s and hen he p-ph asing in cases o na ow ocus. 5 3.1. Kima uumbi (Odden 1996) As we shall see below, Kima uumbi is a language ha makes a dis inc ion be ween inne s ong and weak p-bounda ies. The ollowing ules gi e us he basic p-ph asing: (7) WPB inse ion: applies wi hin an L-domain and inse s a igh b acke o he igh o an XP in ha domain. (8) SPB inse ion: applies wi hin a Fu-domain and inse s a sequence o igh and le p-b acke s be ween an XP and a Fu-p ojec ion. Fu he mo e, we assume he gene al closu e con en ion in (9), which gene a es p-b acke s a he ou e mos edges o he sen ence. (These a e 5 The Kima uumbi and Chichewa da a discussed he e ha e been analyzed by T uckenb od (1999), which we b ie ly summa ize and discuss in he Appendix.  Ibe ia: An In e na ional Jou nal o Theo e ical Linguis ics ol 2.1, 2010, 131–168 pp. h p://www.si .us.es/ibe ia/index.php/ij/index ISSN 1989–8525 138 The Syn ax and P osody o Focus: he Ban u-I alian Connec ion p obably he s onges p-bounda ies bu hey a e no ele an o ou p esen conce ns.) (9) Closu e con en ion ( o ex e nal edges): ou e mos edges o he sen ence mus be closed by a le / igh p-bounda y. Odden (1996) discusses a a ie y o sandhi ules in Kima uumbi, wo o which a e pa icula ly ele an o iden i y inne p-bounda ies (also discussed by T uckenb od (1999); see he appendix). Rule (10a) conspi es o p ese e long owels a he igh -edge o p-ph ases. Rule (10b) iden i ies s ong p- bounda ies by he p esence o a igh -edge High one. (10) a. Vowel Sho ening sho ens a long owel o a s em i i is no igh -adjacen o a p- bounda y. b. Ph ase Tone Inse ion (PTI) inse s a High one a he igh -edge o a p-ph ase i ollowed by a s ong p-bounda y (i.e. a sequence o igh and le p-b acke s). We illus a e he ac ion o he abo e ules wi h some examples. As was said ea lie , i is assumed ha V mo es o In l, an assump ion ha is suppo ed by he p esence o Neg (a unc ional ca ego y) be ween he e b and he objec in (11a). In his example, all he XPs o he igh o he e b, including he empo al ad e b, a e V- ela ed. (Tempo al ad e bs a e known o beha e as VP- ad e bs; see Pese sky 1995). The WPB Inse ion ule hen inse s a igh p- b acke o he igh o all he V- ela ed XPs; as shown in (11b). No e ha he e bal s em o he objec does no unde go Vowel Sho , which es i ies o he p esence o a p-bounda y a e he i s objec . I he second objec we e o con ain a s em wi h a long owel, such as kikóloombe ‘shell’ (as in ‘I did no gi e Mamboondo he shell on F iday’), he la e would no unde go Vowel Sho ei he , es i ying o he p esence o a WPB a e he second objec (Odden 1996: 243). No e ha he absence o a igh edge-High one on he i s and second objec s in he abo e example indica es ha hese XPs a e igh - lanked by a WPB and no by a SPB. (11) a. naamp i lí Mamboondo ki wikilyo í i ju ma I-him-ga e neg Mamoondo co e on-F iday ‘I didn’ gi e Mamboondo a co e on F iday.’ b. [IPIn l-V lí [[ P V DP DP] Ad ]] (naampéi lí Mamboondo) kiwikilyo) iíjuma) Odden con as s empo al ad e bs wi h sen ence le el ad e bs, such as modal ad e bs. When hese a e loca ed a he igh -edge o he sen ence, hey  Ibe ia: An In e na ional Jou nal o Theo e ical Linguis ics ol 2.1, 2010, 131–168 pp. h p://www.si .us.es/ibe ia/index.php/ij/index ISSN 1989–8525 139 Ma ía Luisa Zubiza e a a e p eceded by a SPB, as in he example below. This is shown by he ac ha no only he e bal s em ails o unde go Vowel Sho , bu i also ca ies a igh - edge High one, inse ed by PTI. This is as expec ed, since he modal ad e b belongs o he Fu-domain. The p-ph asing is as in (12b). The sequence o le and igh p-b acke s a he igh edge o he IP is in oduced by he SPB Inse ion Rule in (8). The ou e mos b acke s a e in oduced by he Closu e Con en ion in (9) (Odden 1996: 241). (12) a. aak{laang kwa{li he-will- y pe haps ‘He will pe haps y.’ b. [ [IP In l-V ] Ad ] (aak{laang ) (kwa{li ) A p e e bal subjec in Spec o IP also belongs o he Fu-domain and he e o e i is igh - lanked by a SPB, as illus a ed in (13). No only he s em o he subjec Mamboondo ails o unde go Vowel Sho , bu i also ca ies a igh - edge High one. (13) a. Mamboond aawí i le ‘Mamboondo died.’ b. [IPDP [ In l-V[(V)]]] (Mamboond ) (aaw i le) P e- e bal opicalized XPs a e also igh - lanked by a SPB, as shown by he p esence o a High one a he igh edge o hese XPs, as shown in (14). This is as expec ed since hey belong o he Fu-domain. (We omi he Top head ha heads he TopP p ojec ion in he s uc u e below o sake o conciseness.) (14) a. I í ju m ki yógoy Mamboond naam i lí . on-F iday kikoyogo-bi d Mamboondo I-ga e no ‘I did no gi e Mamboo ndo a kiyogoyo bi d on F iday.’ b. [TopDP [TopDP [Top DP [IPIn l-V Neg[(V) .....] (I ju m ) (ki yógoy ) (Mamboond ) (naam i l ) We u n nex o he case o ocused s uc u es. While Odden does no p o ide a ull p osodic desc ip ion o he ele an s uc u es, enough is he e ha we can d aw some conclusions. Recall ha Kima uumbi has ocus-sensi i e enses, which selec ei he o a e bal ocus o a nominal ocus. The ansi i e examples in (15) illus a e cases o e bal ocal enses, in which he e b is ocused. These examples sugges ha he e is a p osodic bounda y be ween he e b and he objec (Vowel Sho ails o apply o he e bal s em). Fu he mo e,  Ibe ia: An In e na ional Jou nal o Theo e ical Linguis ics ol 2.1, 2010, 131–168 pp. h p://www.si .us.es/ibe ia/index.php/ij/index ISSN 1989–8525 146 The Syn ax and P osody o Focus: he Ban u-I alian Connec ion We e e o he abo e phenomenon as he ‚ ipple e ec o ocus,‛ which is illus a ed by he examples below whe e he ocused e b is ollowed by mo e han one L-domain cons i uen , wo complemen s in (36a), plus a empo al ad e b in (36b). As indica ed by penul ima e owel leng hening (and one e ac ion), he complemen s a e lanked by p-bounda ies. (36) a. [Wha did he do o he house wi h he ock?] (anaméenya)P (nyuúmba)P (ndí mwáála)P he-hi house wi h ock ‘He hi he house wi h he ock.’ b. [Jus wha did he do o he house wi h he ock yes e day?] (anaméenya) (nyuúmba) (ndí mwáála) (dzuulo) To illus a e he phenomenon wi h a de i a ion, conside he s uc u e o (36b), shown in (37). Foc igge s inse ion o SPB (i.e. a sequence o le and igh p-b acke s) be ween he ocused e b and he ocused NP objec nyuúmba. The e o e, he la e (a V- ela ed XP) is now lanked on i s le by a SPB. This SPB is hen copied o he le o he ollowing XPs wi hin he V-domain. (37) [IP V-In l [(V) Foc [VP [VP NP (V) PP ] Ad ] (Anaméenya ) ( (nyuúmba ndí mwáála dzuulo) Foc ule:  nyuúmba ) (dí mwáála ) (dzuulo) Copy:   We a ibu e his p-bounda y ‚copying‛ o he Uni o mi y Condi ion in (38). I is possible ha his Uni o mi y Condi ion is due o mo e gene al eu y hmici y conside a ions, which equi e balanced and compa able p- ph asing wi hin a domain. To de e mine i his is indeed he case, u he esea ch on he in e ac ion o ocus and ‚weigh ‛ (as de e mined by leng h) is needed. 7 con as i e ocused p-ph ase appea s o ha e a ela i ely highe pi ch ange han i would o he wise ha e. 7 No e he o mal simila i y be ween he copying o PBs in ocused s uc u es in Chichewa and he copying o ph asal accen s in non- ocused s uc u es in Chimwiini. We suspec ha ul ima ely he copying ules in he wo languages ha e he same pu pose. In he case o Chichewa, he Uni o mi y Condi ion imposes a ha mony o ph asal one bounda ies wi hin he L-domain in he case o ocused s uc u es (since his language lacks mic o-p osody in non- ocused s uc u es). In he case o Chimwiini, i imposes a ha mony o ph ase accen ypes o ocus-neu al cases (since his language does ha e mic o-p osody in such s uc u es). I is possible ha he  Ibe ia: An In e na ional Jou nal o Theo e ical Linguis ics ol 2.1, 2010, 131–168 pp. h p://www.si .us.es/ibe ia/index.php/ij/index ISSN 1989–8525 147 Ma ía Luisa Zubiza e a (38) I a PB is inse ed a he edge o a cons i uen C in domain D (e.g. L-domain), hen all cons i uen s wi hin ha domain mus be equally lanked by he same ype o PB. One mo e p ope y o Chichewa ha is wo h men ioning is ha , in his language, o e mo emen o a ocused cons i uen o he low Foc posi ion is mo e pa simonious han in o he Ban u languages (Kima uumbi, Chimwiini, Aghem). While VP medial XPs, when ocused, mus mo e o Spec o Foc, a ocused XP in VP- inal posi ion in Chichewa s ays in-si u, as exempli ied in (39). In his example, he VP inal ins umen al ocused PP emains in-si u. (39) Wha did he hi he house wi h? (anaményá nyumbá ndí mwáála) He-hi he house wi h a ock ‘I is wi h a ock ha he hi he house.’ I hus appea s ha Chichewa is a language in which he syn ac ic licensing o a ocused cons i uen may be co e o o e . I VP medial, he licensing mus be o e ; i.e. i mus mo e o Spec o FocP. The eason is ha his is he only way ha i can be p osodically licensed by a SPB o i s igh . In ha posi ion, Foc ule (20) ensu es jus ha . On he o he hand, i i is VP- inal, he syn ac ic licensing is (o may be) co e , because in ha posi ion, i is al eady igh - lanked by a ‚s ong‛ bounda y, in oduced ei he by he Fu-domain ule (5b), i he VP is ollowed by IP-le el ma e ial, as in he cases wi h igh - disloca ed XPs, o o he wise by he Closu e Con en ion in (9). To ecapi ula e, he p oposal pu o h he e is ha in Chichewa, an in-si u VP medial ph ase canno be p osodically iden i ied as ocused, bu a VP- inal ph ase can. Consequen ly, he o me bu no he la e mus mo e o e ly o Foc posi ion. 3.4. Summa y To summa ize, we ha e p o ided an analysis o he Ban u ac s ha assumes an in ima e ela ion be ween he syn ax and he p osody. We ha e sugges ed ha he inse ion o p osodic bounda ies is sensi i e o he unc ional/lexical dis inc ion. Mo e p ecisely, we p oposed ha wi hin he inne domain o he clause, s ong PBs ypically iden i y he unc ional-domain and weak PBs ypically iden i y he L-domain o he clause. This is no o say ha we ne e ind SPBs wi hin he L-domain, no does i mean ha we ne e ind Uni o mi y Condi ion applies selec i ely as o maximize he con as be ween ocus neu al and na ow ocused cases.  Ibe ia: An In e na ional Jou nal o Theo e ical Linguis ics ol 2.1, 2010, 131–168 pp. h p://www.si .us.es/ibe ia/index.php/ij/index ISSN 1989–8525 148 The Syn ax and P osody o Focus: he Ban u-I alian Connec ion WPBs wi hin he Fu-domain. In ac we ha e al eady seen ha he ipple e ec s o ocus gi e ise o SPBs wi hin he L-domain. And as we shall see in he case o I alian, as speech can also weaken a SPB wi hin he Fu-domain. In he Ban u languages, he e is a low Foc posi ion ha is syn ac ically ac i e. The Ve b, when i is ocused, mo es h ough i on i s way o In l. When a DP is ocused, i mo es o Spec o FocP. Ban u being a dema ca i e ype o language, Foc sys ema ically igge s he inse ion o a PB, in ac a SPB, gi en i s s a us as a Fu-ca ego y. The case o Chichewa u he mo e illus a es he p osodic ipple e ec s o ocus, in ha he p-bounda y inse ion igge ed by Foc is copied on o he le edge o each XP ha ollows i . We a ibu ed his copying phenomenon o he Uni o mi y Condi ion on p-ph asing (38) and sugges ed ha his condi ion migh be de i able om mo e gene al eu y hmici y conside a ions. We u n nex o I alian, which also appea s o ha e an ac i e low Foc posi ion. We a gue ha in I alian, as in Ban u, he Foc posi ion igge s inse ion o a SPB, which also gi es ise o a ipple e ec wi h espec o p-ph asing. 4. I alian low Foc posi ion: syn ax and p-ph asing The ques ion a ises as o whe he I alian, a Romance language wi h an ac i e low Foc posi ion (Belle i 2004), also makes a dis inc ion be ween ‚s ong‛ and ‚weak‛ p o sodic bounda ies, l ike Kima uumbi. F asca elli (1999, 2000) discusses wo sandhi ules ha seem o p o ide e idence o ‚s ong‛ PBs: Go gia Toscana (GT) and In e ocalic Spi an iza ion (IS) –bo h a e lenis ules ound in Tuscan and o he Cen al a ie ies o I alian. Following Nespo & Vogel (1986), he au ho e e s o hem as In ona ional-domain ules and con as s hem wi h wo o he sandhi ules ha a e said o be sensi i e o phonological ph ase bounda ies: Radoppiamen o Sin a ico (RS), ound in mos cen al and sou he n a ie ies o I alian, and he Rhy hm Rule (RR), ypical o No he n I alian bu also ound in o he a ie ies (as epo ed by F asca elli). RS and RR apply in exac ly he same en i onmen . RS s eng hens an ini ial consonan o a wo d when i is immedia ely p eceded by a wo d ha ends wi h inal s ess; RR applies o a wo d wi h inal s ess (shi ing s ess le -wa d) when i is immedia ely ollowed by an ini ially s essed wo d. Fo he pu pose  Ibe ia: An In e na ional Jou nal o Theo e ical Linguis ics ol 2.1, 2010, 131–168 pp. h p://www.si .us.es/ibe ia/index.php/ij/index ISSN 1989–8525 149 Ma ía Luisa Zubiza e a o illus a ion, we will only use he GT ule (an In ona ional- ph ase domain ule), o mula ed in (40), and he RR ule (a phonological-ph ase domain ule), o mula ed in (41). 8 We e e he in e es ed eade o F asca elli’s wo k o illus a ion o he o he wo ules. 9 (40) Gio gia Toscana (GT): I changes he oiceless s ops /p, , k/ in o he ica i es *φ, θ, h+ be ween wo sono an s bo h wi hin and ac oss wo ds. (41) Rhy hm Rule (RR): in a sequence o wo wo ds, when he las syllable o he i s wo d and he i s syllable o he second wo d ca y main s ess, he main s ess o he i s wo d shi s le wa d. The so-called phonological ph ase bounda ies a e ‚weake ‛ han he so- called In ona ional Ph ase bounda ies in ha GT and IS can apply ac oss phonological ph ase bounda ies, bu no ac oss In ona ional Ph ase bounda ies. On he o he hand, RS and RR only apply ac oss phonological ph ase bounda ies (bu see no e 9). In o de o enhance he simila i ies be ween he Ban u languages discussed ea lie and I alian, we will ep esen he sen ence in e nal ‚weak‛ PB as a single p-b acke (which ypically applies wi hin he L- domain) and he sen ence in e nal ‚s ong‛ PB as a sequence o le and igh p- b acke s, which ypically appea s wi hin he Fu-domain. Fu he mo e, we will show ha hinking o junc u es in e ms o ela i e s eng h ( a he han in e ms o ca ego ical I-domain and Phonological ph ase-domain junc u es) will allow us o econcile he syn ac ic and p osodic p ope ies o I alian ‚ema ginazione‛ phenomenon. 10 8 RR has been epo ed o exis in many o he languages, including English. The e is some dispu e as o whe he he Rhy hm Rule is a s ess e ac ion o a s ess enhancemen ule. Fo ou p esen pu poses, his is no impo an . Wha is impo an is ha he phenomenon is sensi i e o ph asing. 9 F asca elli men ions ha he e is dialec al a ia ion as o he con ex o applica ion o some o hese ules, such as he GT ule. This au ho also men ions ha RS applies obliga o ily a he Phonological ph ase le el (independen ly o speech a e), bu ha i may op ionally apply a he In ona ional ph ase le el in as speech. 10 We has en o add ha he in ona ional p ope ies o he I alian ‚s ong‛ PB and ‚weak‛ PB s ill emain o be in es iga ed and would equi e ca e ul in es iga ion o he sup asegmen al ea u es associa ed wi h each o hem. The e o e, he p oposal pu o h he e is en a i e, bu hope ully, in he igh di ec ion.  Ibe ia: An In e na ional Jou nal o Theo e ical Linguis ics ol 2.1, 2010, 131–168 pp. h p://www.si .us.es/ibe ia/index.php/ij/index ISSN 1989–8525 150 The Syn ax and P osody o Focus: he Ban u-I alian Connec ion 4.1. I alian p-ph asing in ocus-neu al con ex s I alian has a iable p-ph asing, subjec o a e o speech and weigh conside a ions (Nespo & Vogel 1986, Ghini 1993, F asca elli 1999, 2000). Thus, in an ou -o - he blue con ex , he sen ences below a e said o ha e he ollowing wo ypes o p-ph asing (depending on a e o speech). These gi e ise o an audible in ona ional hia us, and hei p esence is mani es ed by he ac ha in (42c), bu no in (42b), RR applies o he e b, and in (43c), bu no in (43b), RR applies o he head o he NP objec . (42) a. Ques a se á mange ó pas a pe cena. This e ening ea -Fu -1sg pas a o dinne b. (ques a se a) (mange ó) pás a) pe cena) c. (ques a se a) (mánge o pás a) pe cena) (43) a. Non mi piacciono le ci á no diche. b. (non mi piacciono) le ci á) nó diche) c. (non mi piacciono) le cí a nó diche) As we can see in (42b) and (43b), a igh PB can appea a e he lexical head in I alian (unlike Kima uumbi). We will he e o e assume ha in I alian he L-domain ule may inse a igh p-b acke ei he o he le o o he igh o an XP, whe e he choice is dependen on weigh conside a ions as well as a e o speech. F asca elli no es hough ha , in no mal speech a e, he lexical head ends o ge ph ased wi h i s complemen (on i s igh side) unless he e a e hy hmical ac o s ha a o he al e na i e ph asing; see also Ghini (1993). In he unc ional domain, in pa icula hose in ol ing opics, he na u e o he PB is also dependen on weigh and a e o speech conside a ions (F asca elli 2000: 45-52). In pa icula , F asca elli no es ha he GT lenis ule does no apply be ween opics o be ween a on ed opic and he adjacen e b in slow speech (44a), bu i may do so a as e speech a e (44b). Thus, i would seem ha SPB ( ep esen ed he e as a sequence o le and igh p-b acke s), in oduced by he Fu-domain ule, may be “weakened” in o a single ( igh ) p-b acke in ela i ely as speech. This is he case in (44b), whe e he le p-b acke has been dele ed and only he igh p-b acke be ween he opicalized objec and he ollowing e b su i es, allowing o [k] o u n in o he ica i e [h]. (44) a. (ques o li b o) ( *k+onosco) l’au o e) che l’ha sc i o) b. (ques o lib o) *h+onosco) l’au o e) che l’ha sc i o) ‘This book (I) know he au ho ha has w i en i .’  Ibe ia: An In e na ional Jou nal o Theo e ical Linguis ics ol 2.1, 2010, 131–168 pp. h p://www.si .us.es/ibe ia/index.php/ij/index ISSN 1989–8525 151 Ma ía Luisa Zubiza e a We u n nex o he e ec s o ocus on p osodic ph asing. 4.2. I alian ocus and p osodic ‚ema ginazione‛ Since he classic pape by An inucci and Cinque (1977), i is known ha ocus (i.e. con as i e ocus) has an e ec on p osodic ph asing in I alian. An inucci and Cinque e e ed o his phenomenon as ‚ema ginazione‛ (o ma ginaliza ion). The ema gina ed cons i uen ypically ollows a con as i e ocused cons i uen , as he example below illus a es ( om Ca dinale i 2002: 40). In (45B), he pos e bal subjec Ma a is con as i ely ocused (p onounced wi h empha ic s ess) and is sepa a ed om he di ec objec la macchina by an in ona ional hia us o sho pause (indica ed by he au ho wi h a comma). Ca dinale i a gues ha he ema gina ed cons i uen s (p onounced wi h educed pi ch) a e in-si u and a e dis inc om igh -disloca ed cons i uen s. (We indica e he con as i e ocused cons i uen wi h caps and he ema gina ed one wi h i alics.) (45) A: Posso guida e io du an e il iaggio? Can d i e I du ing he ip B: No, non mi piace come guidi: po e à MARA, la macchina. No, [I] no like [you] as d i e : will-d i e Ma a, he ca In o ma ional ocus does no appea o be compa ible wi h he cons uc ion ha in ol es p osodically ‚ema g ina ed‛ cons i uen s. Ex amples ha con ain an ema gina ed objec , bu wi h no con as i e p esupposi ion, a e said o be unna u al (Ca dinale i 2002: 40). This is illus a ed by he example below, whe e he con ex ques ion is na u ally in e p e ed as gi ing ise o an in o ma ional ype o ocus in he co esponding answe -s a emen s (indica ed by unde lines). (46B), whe e he in o ma ional ocused subjec is ollowed by an ema gina ed objec , is said o con as , on he one hand, wi h (46B’), whe e he ocused cons i uen is ollowed by a igh -disloca ed objec (as indica ed by he ac ha i is doubled by a e bal p onominal cli ic) and, on he o he hand, wi h (45B), whe e he ocus is con as i e/empha ic in na u e. (46) A: Chi po e à la macchina? Who will-d i e he ca ? B: ??Po e à Ma a, la macchina. (Ema gina ed objec ) B’: La po e | Ma a, la macchina. (RD objec ) I will-d i e Ma a, he ca .  Ibe ia: An In e na ional Jou nal o Theo e ical Linguis ics ol 2.1, 2010, 131–168 pp. h p://www.si .us.es/ibe ia/index.php/ij/index ISSN 1989–8525 152 The Syn ax and P osody o Focus: he Ban u-I alian Connec ion In I alian, in o ma ional ocused cons i uen s a e p osodically inden i ied by i ue o being loca ed in he posi ion ha ge s assigned Nuclea S ess (NS), namely he igh mos posi ion in he VP (Cinque 1993, Zubiza e a 1998). The con as i e/empha ic na u e o he ocused cons i uen ha p ecedes he ema gina ed ph ases hen sugges s ha such cons i uen is no loca ed a he igh -mos edge o he VP (i.e. adjacen o he igh IP bounda y), bu a he ha i is loca ed in a sen ence in e nal posi ion. This in u n en ails ha he ‚ema gina ed‛ cons i uen s a e loca ed in-si u (wi hin he VP). Ye , i has been a ma e o deba e whe he he p osodically ema gina ed ph ases a e indeed in-si u. F asca elli (2000) assumes ha hey a e igh - disloca ed (RD). On he o he hand, Ca dinale i (2002) a gues ha hey a e in- si u (wi hin he VP). One o he a gumen s p o ided by Ca dinale i is ha ema gina ed ph ases, unlike igh -disloca ed ones, mus espec he base wo d o de (Ca dinale i 2002: 34–35). In (47) he complemen s a e RD (as indica ed by he p esence o p onominal cli ics) and hese may appea in ei he o de . Compa e wi h (48), whe e only he base wo d o de is na u al. (47) a. Ce l’ha nascos o il bambino, il lib o, so o il le o. (RD) The e i -has hidden he child, he book, unde he bed b. Ce l’ha nascos o il bambino, so o il le o, il lib o. (RD) (48) a. Ha nascos o il BAMBINO, il lib o, so o il le o. (EMARG) Has hidden he child he book, unde he bed. b. ??Ha nascos o il BAMBINO, so o il le o, il lib o. (EMARG) The examples in (49) u he exempli y he base wo d o de equi emen in he case o ema gina ed cons i uen s. The ema gina ed DO and he in ini i al complemen mus appea in ha sequence when p osodically ma ginalized; c . (49a–c) and (49d–e). (49) a. Che cosa hai con in o MARIO, a a e? Wha ha e (you) con inced Ma io o do? b. Che cosa hai CONVINTO, Ma io, a a e? Wha ha e (you) con inced Ma io o do? c. Che cosa hai con in o TU, Ma io, a a e? Wha ha e YOU con inced Ma io o do? d. *Che cosa hai CONVINTO, a a e, Ma io? e. *Che cosa hai con in o TU, a a e, Ma io?  Ibe ia: An In e na ional Jou nal o Theo e ical Linguis ics ol 2.1, 2010, 131–168 pp. h p://www.si .us.es/ibe ia/index.php/ij/index ISSN 1989–8525 153 Ma ía Luisa Zubiza e a Ca dinale i gi es a numbe o o he syn ac ic es s ha dis inguish ema gina ed om igh -disloca ed cons i uen s. Ema gina ed cons i uen s can be quan i ied, while igh -disloca ed ones canno . The geni i e cli ic can be ex ac ed om an ema gina ed objec bu no om a igh -disloca ed one. Wi h espec o p onominal binding, an ema gina ed complemen clause is shown o be wi hin he scope o he pos e bal subjec , as opposed o a RD complemen clause. A pa icula ly compelling a gumen o he iew ha an ema gina ed ph ase is wi hin he VP is p o ided by he pos e bal subjec in he Ancona dialec o cen al I aly (Ca dinale i 2002: 50–52). In his dialec , pos e bal subjec s wi hin he VP do no ag ee wi h he e b (50). Ema gina ed pos e bal subjec s do no ag ee wi h he e b ei he and, as expec ed, hese mus obey he base wo d o de . This is illus a ed by he con as be ween (51a), in which he ema gina ed cons i uen s obey he base wo d o de , and (51b), in which he ema gina ed cons i uen s do no obey he base wo d o de . Also, compa e (51b) wi h (51c), whe e he objec and he subjec a e igh -disloca ed, and need no obey he base wo d o de . (50) Ques o disegno l’ha a o quei bambini lí. his d awing i has done hose child en he e (51) a. Ques o disegno l’ha a o IERI, quei bambini lí. b. *Ha a o IERI, il disegno, quei bambini lí. has done yes e day, his d awing, hose child en he e c. L’ hanno a o IERI, il disegno, quei bambini lí. i - ha e done yes e day, his d awing, hose child en he e C . Quei bambini hanno/*ha a o il disegno ie i. When a pos e bal unaccusa i e subjec is ema gina ed, i has he ypical p ope ies o unaccusa i e pos e bal subjec s, namely geni i e cli iciza ion om he pos e bal subjec is possible, as illus a ed in (52b); compa e wi h he RD coun e pa in (52c). Fu he mo e, an ema gina ed pos e bal subjec may be quan i ied (53a), bu no so a RD subjec ; see he eply pa in (53b). (52) a. Ne è enu o uno. (in-si u, VP-in e nal subj, neu al in ona ion) O - hem has come one b. Ne è enu o IERI, uno (EMARG subj) o - hem has come yes e day, one c. ??Gliene è s a o egala o, a Gianni, uno (RD subj) o-him o - hem has been gi en, o Gianni, one  Ibe ia: An In e na ional Jou nal o Theo e ical Linguis ics ol 2.1, 2010, 131–168 pp. h p://www.si .us.es/ibe ia/index.php/ij/index ISSN 1989–8525 154 The Syn ax and P osody o Focus: he Ban u-I alian Connec ion (53) a. Può già RITIRARE, ogni agazzo, la (sua) macchina. (EMARG subj) Can al eady go-and- ake, e e y boy, he (his) ca b. A: Chi ha incon a o, ogni s uden e? Who has me , e e y s uden ? ‘Whom has e e y s uden me ?’ B:*Ha incon a o il p eside, ogni s uden e. (RD subj) [he] has me he dean, e e y s uden ‘E e y s uden has me he dean’ I is o be no ed ha , while ema gina ed cons i uen s mus obey he base wo d o de , he con as i e ocused cons i uen , which appea s igh -adjacen o he e b, does no do so necessa ily. This is exempli ied in (54a), whe e he PP loca i e complemen is con as i ely ocused and p ecedes he ema gina ed objec . Compa e also (48b) and (54b), whe e a empo al ad e b is con as i ely ocused and p ecedes bo h he ema gina ed objec and loca i e complemen (Ca dinale i 2002). (54) a. Il bambino ha nascos o so o il LETTO, il lib o. b. Il bambino ha nascos o IERI, il lib o, so o il le o. To he ex en ha ema gina ed cons i uen s a e in-si u and appea immedia ely o he igh o a (con as i e) ocused cons i uen and ha he ocused cons i uen in u n appea s immedia ely o he igh o he in lec ed e b o he e bal pa iciple (known o be ou side he VP and in he In lec ional ield; see Belle i 1990), i is emp ing o conclude ha I alian has an ac i e low ocus posi ion, loca ed below In l and abo e he VP (Belle i 2004). 11 I indeed he cons i uen s ha ollow he con as i e ocused cons i uen (loca ed in he low Foc posi ion) a e wi hin he VP (as a gued by Ca dinale i), he ‚p osodic ema gina ion‛ p ope ies o hese pos - ocal cons i uen s would seem o be compa able o hose ound in he Ban u languages. We elabo a e on his below. 4.3. I alian ocus and he na u e o he p-bounda y F asca elli (2000) no es ha he p-bounda y be ween he ocused and he ema gina ed cons i uen s blocks he applica ion o RR. Recall ha RR is blocked 11 Belle i (2004) sugges ed ha he low ocus posi ion only lodges in o ma ional ocused ph ases and he high ocus posi ion (abo e IP) lodges he con as i e ocused posi ion. Bu his does no squa e wi h wha we know o be he case in o he languages, in pa icula he Ban u languages. Fu he mo e, i is unlikely ha syn ac ic posi ions a e labelled o ype o ocus.  Ibe ia: An In e na ional Jou nal o Theo e ical Linguis ics ol 2.1, 2010, 131–168 pp. h p://www.si .us.es/ibe ia/index.php/ij/index ISSN 1989–8525 155 Ma ía Luisa Zubiza e a by any ype o PB, including ‚weak‛ PB. Compa e (42c) wi h (55b), whe e he e b is ocused. (55) a. Ques a se a, MANGERÓ, pas a, pe cena. b. * Ques a se a, MÁNGERO, pas a, pe cena. As was men ioned ea lie , F asca elli a gues ha GT can apply ac oss a ‚weak‛ PB. This is exempli ied by he ocus-neu al sen ence in (56a) (F asca elli 2000: 35). In his example, GT weakens he i s consonan o he p eposi ion ha heads he PP con gli amici, which is sepa a ed by a WPB om he p eceding loca i e complemen . On he o he hand, GT ails o apply in he con ex o a con as i ely ocused cons i uen , as shown in (56b). This sugges s ha he PB ha ollows he con as i ely ocused cons i uen is indeed a ‚s ong‛ PB bounda y, compa able o he PB ha p ecedes igh disloca ed ph ases; see (57) (F asca elli 2002: 29). In his example, he objec is an in o ma ional ocused cons i uen , which bea s NS, ollowed by a igh -disloca ed PP complemen . GT does no apply o he p eposi ion ha heads his PP complemen ei he . (No e ha he PP in (57) is a case o RD, despi e he absence o an an icipa o y cli ic. Indeed, igh -disloca ed PPs a e only op ionally doubled by a cli ic, and o his PP in pa icula , he e is no co esponding cli ic.) (56) a. (and ó) al cinema) [h]on i miei amici) Go-FUT-1SG o- he cinema wi h he my iends b. (and ò al CINEMA) ( [k]on i miei amici) *[h]on (57) Q: Cosa mange à Ca lo con i suoi amici? ‘Wha will Ca lo ea wi h his iends?’ A: (mánge a pás a) ([k]on gli amici) *[h]on ea -FUT-3SG pas a wi h his iends I is his simila i y be ween he PB ha p ecedes igh -disloca ed elemen s and he PB ha p ecedes ema gina ed cons i uen s ha led F asca elli o conclude ha ‚ema gina ed‛ cons i uen s a e syn ac ically igh -disloca ed, in ha hey a e bo h lanked by s ong PBs, o he ype associa ed wi h he unc ional domain. We p opose o esol e his appa en con adic ion be ween he non- disloca ed p ope ies o ema gina ed cons i uen s and i s p osodic p ope ies by assimila ing he case o I alian o ha o he Ban u languages. The con as i ely ocused cons i uen is loca ed in he spec o low FOC and i is he FOC unc ional head ha igge s he inse ion o he s ong PB. To summa ize,  Ibe ia: An In e na ional Jou nal o Theo e ical Linguis ics ol 2.1, 2010, 131–168 pp. h p://www.si .us.es/ibe ia/index.php/ij/index ISSN 1989–8525 162 The Syn ax and P osody o Focus: he Ban u-I alian Connec ion ‚ ipple‛ e ec o ocus, due o he ac ion o he Uni o mi y Condi ion on p- ph asing. The ques ion a ises as o why in I alian (bu no in Ban u) he sen ence in e nal low Foc posi ion only hos s con as i e ocused cons i uen s and no in o ma ional ocused ones. The answe o e ed o his ques ion is as ollows. Because I alian, unlike he Ban u languages, is bo h a culmina i e and a dema ca i e language, he ocused cons i uen mus ca y main sen en ial s ess (which co esponds o he Nuclea Accen a he in ona ional le el). In I alian, as in o he Romance languages, in o ma ional ocus mus ca y Nuclea S ess (NS) and u he mo e, NS is assigned o he las cons i uen wi hin he VP in hese languages; i.e. he cons i uen igh -adjacen o he IP bounda y. On he o he hand, sen ence-in e nal ocused cons i uen s ecei e s ess om he Empha ic Rule, which applies eely (including a he subwo d le el), hus associa ing sen ence in e nal ocus wi h an empha ic/con as i e in e p e a ion. The low Foc posi ion in I alian, when s ic ly sen ence-in e nal, can only ecei e main s ess ia he Empha ic ule. Two syn ac ically ac i e posi ions o ocus ha e been iden i ied: immedia ely abo e IP and immedia ely abo e P. I is possible ha hese wo syn ac ic posi ions o ocus a e no acciden al. The e could be in e p e a ional easons. On he LF side, i ocus is quan i ica ional (He bu ge 2000, Rizzi 1997), hen he domain o ocus is expec ed o be he same as ha o o he (s ong) quan i ie s, namely, he big clause (IP) o he small clause ( P) (May 1985). On he PF side, possibly hese wo domains co espond o he phase-domains (in he sense o Chomsky 1995). The ques ion o whe he phases, namely CP and P, a e he domain whe e p-ph asing should be de ined (as p oposed by K a ze & Selki k 2007, Ishiha a 2007) will no be esol ed he e. I will simply ema k ha he p- ph asing igge ed by Focus in languages like Chichewa, Chimwiini, and I alian canno be eadily cap u ed by a phase-based heo y i we assume ha P (o i s complemen ) cons i u es he domain o spell-ou . In hese languages, we obse e ha he low Foc posi ion, loca ed abo e P, igge s he inse ion o a (s ong) p-bounda y and ha his has an e ec on he mic o-p osody o he ma e ial con ained wi hin he phase i sel . The e o e, i we a e o assume ha phases a e ele an in de ining he domain o p-ph asing, i mus be he case  Ibe ia: An In e na ional Jou nal o Theo e ical Linguis ics ol 2.1, 2010, 131–168 pp. h p://www.si .us.es/ibe ia/index.php/ij/index ISSN 1989–8525 163 Ma ía Luisa Zubiza e a ha hese include ope a o s (like Foc), which immedia ely domina es he P. We lea e his ques ion un esol ed. Fu he and mo e in-dep h in ona ional analysis o he languages discussed he e and o o he languages is ye needed, one ha includes o he ope a o s han ocus, such as he s anda d s ong quan i ie s wi h a iable scope. In his pape , I ha e examined he e only a small subse o languages and o he ypologically di e en languages emain o be examined in ligh o he p oposal pu o h he e. Pa icula ly ele an a e he head inal languages, like Ko ean, Japanese, and Bengali (see Jun 2005 and e e ences ci ed he ein). I would seem ha head inal languages may, o some ex en , iola e he cons ain in (74) (Jun 2010). I his ypological co ela ion u ns ou o be co ec , i has impo an implica ions o he heo izing o he syn ax/p osody in e ace. APPENDIX T uckenb od ’s (TR’s) n lysis o Kim uumbi nd Chichew . A sho summa y. TR’s p o posal is couched wi hin an OT e sion o P osodic Phonology (along he lines pu o h by Selki k & Chen 1990 and Selki k 1995). The basic assump ion o P osodic Phonology is ha phonological ules apply o p osodic cons i uen s and no o syn ac ic cons i uen s ( he so-called Indi ec Re e ence hypo hesis). I is u he mo e assumed ha p osodic ca ego ies a e o ganized along a bo om-up hie a chy o syllables and ee below he P osodic Wo d, on he one hand, and Phonological Ph ases (o p-ph ases), In ona ional Ph ases, and U e ance Ph ase abo e he P osodic Wo d. Unde he End-based heo y p oposed by Selki k (1995), i is assumed ha he ules ha align p osodic (P) cons i uen s wi h syn ac ic cons i uen s only make e e ence o edges o syn ac ic ph ases (and no o syn ac ic ela ions). Two basic alignmen ules a e p oposed: (76) a. Align-XP, R: Fo each XP, he e is a P such ha he igh edge o XP coincides wi h he igh edge o P. b. Align-XP, L: Fo each XP, he e is a P such ha he le edge o XP coincides wi h he le edge o P. The End-based heo y, couched wi hin an Op imali y Theo e ic app oach, pos ula es ha he e a e undomina ed (in iolable cons ain s), such  Ibe ia: An In e na ional Jou nal o Theo e ical Linguis ics ol 2.1, 2010, 131–168 pp. h p://www.si .us.es/ibe ia/index.php/ij/index ISSN 1989–8525 164 The Syn ax and P osody o Focus: he Ban u-I alian Connec ion as he Exhaus i i y Cons ain (77). O he cons ain s, such as he Non-Recu si i y Cons ain (78) and *P-ph ase cons ain (79) a e said o be iolable 16 and a e a iably anked ac oss languages. Ranking o *P-ph ase below (76a) and abo e (76b) ende s a igh -alignmen g amma , and he e e se ende s a le - alignmen g amma . TR adds ye ano he ( iolable) cons ain , namely W ap XP (80). No e ha he unc ion o W ap XP o e laps wi h ha o he Align, XP ules in (76), gi ing ise o a pa ial edundancy wi hin he sys em. (77) Exhaus i i y Cons ain (Selki k 1995): Pa sing on e e y p osodic le el mus be exhaus i e. (78) Non-Recu si i y (Selki k 1995): punishes a p osodic cons i uen ha con ains ano he p osodic cons i uen o he same le el. (79) *P-ph ase cons ain : A oid P-ph ases. (80) W ap XP: Each XP is con ained in a phonological ph ase. I is u he mo e assumed ha , as s a ed in he Lexical Ca ego y Condi ion below, he ules and cons ain s ha go e n he mapping be ween p osodic and syn ac ic cons i uen s do no apply o unc ional heads and hei p ojec ions, an assump ion ha we ha e challenged. (81) Lexical Ca ego y Condi ion (LCC) (T uckenb od 1999, based on Selki k 1995) Cons ain s ela ing syn ac ic and p osodic ca ego ies apply o lexical syn ac ic elemen s and hei p ojec ions, bu no o unc ional elemen s and hei p ojec ions, o o emp y syn ac ic elemen s and hei p ojec ions. In he Ban u languages unde discussion (Kima uumbi and Chichewa), Align-XP, R is anked highe han *P-ph ase and Align-XP, L is anked lowe han *P-ph ase. Conside he p-ph asings possibili ies in (82a-d) o a lexically headed ph ase. The p-ph asing in (82a) iola es Exhaus i i y and is no an op ion. On he o he hand, he op ions in (82b-d) a e all a es able p-ph asings. The p-ph asing in (82c) is ob ained unde he anking gi en in (83a), and i is a gued o be he case o Kima uumbi. The p-ph asing in (82d) is ob ained unde he anking gi en in (83b), and i is a gued o be he case o Chichewa. 16 T uckenb od p oposes a g adien o mula ion o Non- ec: Any wo-ph ases ha a e no disjoin in ex ension a e iden ical in ex ension. See his wo k o mo i a ion and de ails.  Ibe ia: An In e na ional Jou nal o Theo e ical Linguis ics ol 2.1, 2010, 131–168 pp. h p://www.si .us.es/ibe ia/index.php/ij/index ISSN 1989–8525 165 Ma ía Luisa Zubiza e a (82) [XP X YP ZP] a.* ( ) ( ) non-exhaus i e ph asing: iola es Exhaus i i y b . ( ) ( ) ( ) ph asing wi h addi ional s uc u e: iola es W ap XP c. (( ) ) iola es Non-Rec, bu obeys W ap XP d. ( ) obeys Non- ec & W ap XP; iola es Align-XP,R (83) a. Align-XP, R > W ap XP > Non- ecu si i y. (Kima uumbi) b. Non-Recu si i y > W ap XP >Align XP, R. (Chichewa) In TR’s sys em, he Align-XP ule accomplishes he same ac ion as ou L- domain inse ion ule does, and W ap XP accomplishes, o some ex en , wha ou Fu-domain inse ion ule does. F om he empi ical poin o iew, he main p oblem wi h TR’s analysis is ha i assumes he w ong syn ax. We ha e gi en wo pieces o e idence ha in Kima uumbi he e b is no in he VP bu in In l. The pos e bal ocused subjec is ph ased wi h he e b, clea ly showing ha he subjec is wi hin he scope o he e b; i is no igh -disloca ed. Assuming ha subjec s a e gene a ed highe han he e b wi hin he e bal ph ase, i mus be he case ha he e b has mo ed ou o he VP and in o an In l p ojec ion. The o he piece o e idence ha he e b in Kima uumbi is in In l is p o ided by he ac ha he Neg unc ional ca ego y (loca ed abo e he VP) can su ace be ween he e b and he objec . I he e b is indeed in In l, W ap XP will ail o inse he le p-b acke o he le o he e b in all he e bal s uc u es discussed in 3.1. Indeed, because o he LCC, unc ional ca ego ies (such as In l) a e in isible o p osodic ules and cons ain s. Samek-Lodo ici (2005), who ca ies o e TR’s sys em o I alian, no ices his p oblem bu does no p opose a solu ion. I submi ha his is a non- i ial p oblem ha canno be ixed wi hou sabo aging he LCC. As o ocused s uc u es, TR p oposes ha ocused cons i uen s a e ma ked wi h a ocus- ea u e and ha his ocus- ea u e igge s he inse ion o a p-bounda y: (84) Align- ocus: a ocused cons i uen is igh -aligned wi h a p-bounda y. Align- ocus is highly anked. The p-ph asing in (82b) obeys bo h Align- ocus and Non-Rec, while iola ing W ap XP. This is a possible p-ph asing in Chichewa  Ibe ia: An In e na ional Jou nal o Theo e ical Linguis ics ol 2.1, 2010, 131–168 pp. h p://www.si .us.es/ibe ia/index.php/ij/index ISSN 1989–8525 166 The Syn ax and P osody o Focus: he Ban u-I alian Connec ion gi en ha in his language W ap XP is anked lowe han Non- ec; see (83b). 17 TR’s p oposal c ucially assumes ha he p-bounda y ha lanks he ocused cons i uen is igge ed by a ea u e on he ocused ph ase. As we ha e shown in 4, I alian p o ides c ucial e idence agains his assump ion. REFERENCES Aboh, Enoch Oladé. 2007. Le wa d ocus e sus igh wa d ocus: he Kwa- Ban u conspi acy. SOAS Wo king Pape s in Linguis ics 15, 81–104. An inucci, F ancesco & Guglielmo Cinque. 1977. Sull’o dine delle pa o le in i aliano: l’ema ginazione. S udi di G amma ica I aliana 6, 121-146. Belle i, Ad iana. 1990. Gene alized Ve b Mo emen . Aspec s o Ve b Syn ax. To ino: Rosenbe g & Sellie . Belle i, Ad iana. 2004. Aspec s o he low IP a ea. In Luigi Rizzi (ed.), The S uc u e o CP and IP. 16–51. Ox o d/New Yo k: Ox o d Uni e si y P ess. Biloa, Edmond. 1997. Func ional ca ego ies and he syn ax o ocus in Tuki. Lincom S udies in A ican Linguis ics 02. München: Lincom Eu opa. Ca dinale i, Anna. 2002. Agains op ional and null cli ics. Righ disloca ion s. ma ginaliza ion. S udia Linguis ica 56 (1), 29–57. Chomsky, Noam. 1995. The Minimalis P og am. MIT P ess. Cinque, Guglielmo. 1990. Types o A’-dependencies. MIT P ess. Cinque, Guglielmo. 1993. A null heo y o ph ase and compound s ess. Linguis ic Inqui y 24, 239–298. Downing, Lau a, Al M enje & Be nd Pompino-Ma schall. 2004. P osody and in o ma ion s uc u e in Chichewa. ZAS Pape s in Linguis ics 37, 167–186. Downing, Lau a. 2006. The p osody and syn ax o ocus in Chi umbuka. ZAS Pape s in Linguis ics 43, 55–79. Downing, Lau a. 2008. Focus and p ominence in Chichewa, Chi umbuka, and Du ban Zulu. ZAS Pape s in Linguis ics 49, 47–65. Face, Timo hy. 2001. Focus and ea ly peak alignmen in Spanish in ona ion. P obus 13, 223–246. Face, Timo hy. 2002. Local in ona ional ma king o Spanish con as i e ocus. P obus 14, 71–92. F asca elli, Ma a. 1999. The p osody o ocus in I alian. P obus 11 (2), 209–238. 17 No e ha such mic o-p osody iola e *P-ph ase, so his cons ain mus be anked lowe han bo h Non-Rec and W ap XP.  Ibe ia: An In e na ional Jou nal o Theo e ical Linguis ics ol 2.1, 2010, 131–168 pp. h p://www.si .us.es/ibe ia/index.php/ij/index ISSN 1989–8525 167 Ma ía Luisa Zubiza e a F asca elli, Ma a. 2000. The Syn ax-Phonology In e ace in Focus and Topic Cons uc ions in I alian. Do d ech : Kluwe Academic Publishe s. Ghini, Mi co. 1993. - o ma ion in I alian: A new p oposal. To on o Wo king Pape s in Linguis ics 12, 41–78. Depa men o Linguis ics, To on o. On a io. Hayes, B uce & Adi i Lahi i. 1990. Bengali In ona ional Phonology. Na u al Language and Linguis ic Theo y 9, 47–98. He bu ge , Elena. 2000. Wha Coun s: Focus and Quan i ica ion. MIT P ess. He nanz, Ma ía Luisa & José Ma ía B uca . 1987. La Sin axis (1). Ba celona: Edi o ial C í ica. Inkelas, Sha on & D aga Zec. 1990. The Phonology-Syn ax Connec ion. The Uni e si y o Chicago P ess. Ishiha a, Shinichi o. 2007. Majo ph ase, ocus in ona ion, mul iple spell-ou . The Linguis ic Re iew 24, 137–167. Jun, Sun-Ah. 2005. P osodic ypology. In Sun-Ah Jun (ed.), P osodic Typology: The Phonology o In ona ion and Ph asing. Chap e 16. Ox o d/New Yo k: Ox o d Uni e si y P ess. Jun, Sun-Ah. 2010. P osodic ma kings o complex NP ocus and p e-/pos - ocus s ing. Pape p esen ed a he 28 h Wes Coas Con e ence on Fo mal Linguis ics, Uni e si y o Sou he n Cali o nia, Los Angeles, Ca. Kane a, Jonni. 1990. Focusing on phonological ph ases in Chichewa. In Sha on Inkelas & D aga Zec (eds.), 145–161. Kissebe h, Cha les W. 2010. Op imali y heo y and he heo y o phonological ph asing: The Chimwiini e idence. In Nomi E eschik-Shi & Lisa Rochman (eds.), The Sound Pa e ns o Syn ax, 217-246. Ox o d/New Yo k: Ox o d Uni e si y P ess. K a ze , Angelika & Elisabe h Selki k. 2007. Phase heo y and p osodic spell- ou : he case o e bs, The Linguis ic Re iew 24, 93–135. May, Robe . 1985. Logical Fo m. MIT P ess. Nespo , Ma ina & I ene Vogel. 1986. P osodic Phonology. Ams e dam: Fo is Publica ions. Nibe , Holly. 2000. Phone ic and Phonological E idence o In e media e Ph asing in Spanish In ona ion. Unpublished Phd Disse a ion, Uni e si y o Illinois a U bana-Champaign. Laka, I. 1990. Nega ion in Syn ax: On he Na u e o Func ional Ca ego ies and P ojec ions. Ph.D. disse a ion, MIT. Odden, Da id. 1996. The Phonology and Mo phology o Kima uumbi. Ox o d: Ox o d Uni e si y P ess.  Ibe ia: An In e na ional Jou nal o Theo e ical Linguis ics ol 2.1, 2010, 131–168 pp. h p://www.si .us.es/ibe ia/index.php/ij/index ISSN 1989–8525 168 The Syn ax and P osody o Focus: he Ban u-I alian Connec ion Pese sky, Da id. 1995. Ze o-Syn ax: Expe ience s and Cascades. MIT P esss. P ie o, Pila . 2007. Phonological ph asing in Spanish. In Fe nando Ma inez-Gil & Sonia Colina (eds.), Op imali y-Theo e ic S udies in Spanish Phonology, 39–61. Be lin/New Yo k: John Benjamins Publishing Co. Rizzi, Luigi. 1997. The Fine S uc u e o he Le Pe iphe y. In Liliane Haegeman (ed.) Elemen s o G amma , 281–337. Do d ech : Kluwe . Samek-Lodo ici, Vie i. 2005. P osody-syn ax in e ac ion in he exp ession o ocus. Na u al Language and Linguis ic Theo y 23, 687–755. Selki k, Elisabe h. 1986. On de i ed domains in sen ence phonology. Phonology Yea book 3, 371–405. Selki k, Elisabe h & Tong Shen. 1990. P osodic domains in Shangai Chinese. In Sha on Inkelas and D aga Zec (eds.), 313–337. Selki k, Elisabe h. 1995. The p osodic s uc u e o unc ion wo ds. Uni e si y o Massachuse s Occasional Pape s 18: Pape s in Op imali y Theo y, 439–469. GLSA, Uni e si y o Massachuse s, Amhe s . T uckenb od , Hube . 1999. On he ela ion be ween syn ac ic ph ases and phonological ph ases. Linguis ic Inqui y 30: 219-256. Zubiza e a, Ma ía Luisa. 1998. P osody, Focus, and Wo d O de , MIT P ess. Zubiza e a, Ma ía Luisa. 2001. The cons ain on p e e bal subjec s in Romance in e oga i es: A minimali y e ec . In Aa ke Hulke & Jean-Y es Pollock (eds.), Subjec In e sion in Romance and he Theo y o Uni e sal G amma , 183– 204. Ox o d/New Yo k: Ox o d Uni e si y P ess. Ma ía Luisa Zubiza e a zubiza @usc.edu Depa men o Linguis ics, GFS 301 Uni e si y o Sou he n Cali o nia Los Angeles, Ca. 90089-1693