24.962/DS Page 1 5/23/05 Correspondence basics (revised, augmented) Plan: •review basic correspondence notions and constraints •why any theory of phonology sooner or later recognizes a theory of correspondence•loan adaptation•meter and metrical tension•rhyming•segments vs. features as the unit of correspondence•input vs. output context in correspondence constraints.(1) UR /sIkl/ - SR [sIkl]` as against SR [sIk], [sIkIl], [sIkt] [sIko], [sIlk] (2) Ranking of correspondence constraints yielding SR [sIkl`] from a possible UR /sIkl/ (3) No correspondence rankings needed in a rule-based grammar: [+son] -> [+syllabic]/ [-son]_# (4) Two misapprehensions: •Correspondence Theory is the direct consequence of rejecting rules •…and ordering or cycles(5) Two arguments for ranked correspondence constraints in any grammar: • loan adaptation • poetic meter and rhyme(6) Terms: •L1, L2: native language, and second language, respectively •loan(word), borrowing: a form from L2 adapted and imported into L1 •loan adaptation: process thru which loans abide by phonological structure of L1 •loan phonology: the entire system that allows borrowers to adapt L2 words (7) Korean Hemnit < [hœmlEt]; kÓApÓÈlli – kÓApÓÈ|i < [kÓOpli]; pÓEksÈ < [fœks] •K [l] is either a coda or a geminate [ll]; [|] alternates with [l], is ok in V_V •some general processes of K:a. nl -> ll b. Cl -> Cn applies to some loans (Hamlet), not all (Copley). c. CN -> C[+nas] N applies to some loans (8) The mysteries of loan phonology: •Most rules operative in loan phonology seem absent from L1. There’s no epenthesis in K •And many L1 rules actually don’t apply to loanwords, or not always. • But loans tend to abide by the output constraints of L1. No Cl, CN on the surface in K loans from English No complex onsets, or impermissible codas in K. Most sounds in loanwords are native sounds, even if L2 source contained nonnative sounds: E. [f] -> K. [pÓ] If L1 rules don’t apply to loans, then in what sense are loans adapted to L1’s phonology? And where do the L1 speakers come up with the novel rules they do apply to loans?24.962/ds Page 2 5/23/05 (9) OT provides a different view of this: L1 phonology is a. a set of active or undominated phonotactic (markedness) constraints: e.g. *Complex onset b. a set of correspondence constraints: e.g. MAX C/ IO, DEP V/IO c. rankings within (a), and within (b), and between (a) and (b). e.g. *Complex onset >> MAX C (Japanese); MAX C >> *Complex onset (English) OT explains how loans are adapted to L1 phonology without undergoing L1 “rules”. •Loan adaptation = insuring that phonotactic constraints of L1 are satisfied in loans. •What differs between L1 phonology and adaptation from L2 to L1? Ranking among correspondence constraints. Something we won’t address: why the ranking of correspondence comes to differ between L1 and L2-> L1. (10) Same phonotactics, different means of satisfaction in L1 and in loans: Korean, E-> K Coda cluster resolution; coda modification; other cluster changes in K UR Final or pre-C Pre-V Gloss a. kulm kum kulm-starve b. palp pap palp-tread on c. hÈlk hÈk hÈlk-soil d. tols tol tols-anniversary e. haltÓ hal haltÓ-lick g. moks mok moks-share h. kaps kap kaps-price i. antS an antS-sit j. nas nat nas-sickle k. nas nat nas-sickle l. natS nat natS-daytime m. patÓ pat patÓ-field 8. Corresponding rules in K(L1), applied in this order: i. [+lateral] -> Ø/__[-syllabic, -coronal]]s ii. [-syllabic] -> Ø/[-syllabic]_]s iii. [+cons] -> [-cont, -spread glottis, -strident]/ _]s 9. Other general properties of K, whether or not supported by alternations: • no CC codas/onsets; no diphthongs; no fricatives other than s.•no consonantal continuants (e.g. s) in coda, or affricates (e.g. tS) •the only C in coda are unaspirated, voiceless stops or nasals.224.962/ds Page 3 5/23/05 10. Coda cluster resolution in E->K (L2-> L1), and other C transformations English Nativized in K as Not as a. pÓœs <pass> pÓEsÈ pEt c. grœf <graph> kÈrapÓÈ kÈrap d. lOv <love> robÈ rop e. fœks <fax> pÓEksÈ pÓEk, pÓEkÈt f. pÓølp <pulp> pÓOlpÓÈ pÓOp, pÓOl, pÓOlÈp 11. Some “rules” involved in E->K(L2 -> L1): i. [+cont,-voice, +cons] -> [+spread glottis, -cont] (f -> pÓ) ii. [+cont,+voice, +cons] -> [-spread glottis, -cont] (v -> b) iii. [-cont, -voice] -> [+spread glottis] (p, t, k -> pÓ, tÓ, kÓ) iv. [+lateral] -> [+long]/ V_V (film -> pÓillÈm) -> [-lateral]/#_ (love -> robÈ) v. Ø-> È /[+strident, +anterior]_{C, #} (pass -> pÓEsÈ) vi. Ø-> È /[+cont, +voice]_ # (love -> robÈ) [+cont, +spread glott]_ # (graph -> kÈrapÓÈ) 12. A different picture in an OT analysis: •Korean undominated phonotactics: *Complex onset; *Complex coda, If [–son]]s then [-cont, -voice, -sp.gl]; If [+cont, -son] then [+strident] *Cl, *[+nasal][-syll, -nasal] These and others are always satisfied by adapted loans. •Korean L1 system *Complex onset MAX IO *Complex coda If C]s then [-cont, -voice, -sp.gl] MAX/DEP/Ident IO DEP IO Ident [±voice], [±sp.gl] IO If [+cont, +cons] then [+strident] •English-to-Korean (L2 -> L1) system MAX IO, DEP V≠ È, Ident F IO *Complex onset DEP È *Complex coda If C]s then [-cont, -voice, -sp.gl] If [+cont, +cons] then [+strident] 13. A preliminary analysis of graph -> kÈrEpÓÈ L2 grœf K-phonotactics Ident F DEP È a. krEf b. kep *!(r) c. kÈrEp * Fd. kÈrEpÓÈ ** MAX C *!* (kr, f) *! (f -> p) 324.962/ds Page 4 5/23/05 14. A preliminary analysis of fax-> pÓEksÈ L2 fœks K-phonotactics Ident F DEP È a. pÓEks s, s]s) b. pÓEk *!(s) Fc. pÓEksÈ * MAX C *!* (CC]15. Summary of this point •The phonotactics that were undominated in the K (L1) system remain so in the KE (L1-> L2) system. It’s in this sense that the loans have been submitted to the L1 grammar. •The K (L1) differs from the KE (L1-> L2) in the rankings of correspondence constraints relative to each other: a grammatical description lacking correspondence constraints cannot even describe what it is that K and KE share and how they differ. 16. Syntagmatic and paradigmatic correspondence •syntagmatic: strings in correspondence cooccur within the expression being evaluated•paradigmatic: they don’t. 17. Correspondence in meter: between the abstract metrical