Truncation and Exponence How small can you get ? jollification - PowerPoint PPT Presentation

1 2 DFG Network Core Mechanisms of Exponence - o : Jan 11 – 12, 2008 aggravation aggr-o business bizz-o (Lappe 2007) Truncation and Exponence – How small can you get ? jollification joll-o Sabine Arndt-Lappe (Universität Siegen) 2 interesting questions: Note: A large part of this presentation (both data and analysis) is joint work with • What is the structure of outputs of truncation? ('word structure') Birgit Alber, Universitá degli Studi di Verona • Which part of the base form survives in the derivative? ('anchoring') I. Introduction (2) two research disciplines concerned with truncation (1) the phenomenon - examples a. Word-formation theory (e.g. Dressler & Merlini Barbaresi 1994, Dressler 2000, a. hypocoristics descriptive studies of individual languages, for English cf., e.g., Jespersen German: Italian: 1965repr., Marchand 1960, Adams 1973, Bauer 1983) Andreas Andi Andrea Andri => The structural characteristics of truncated forms are considered to be largely Dagmar Daggi Cristina Cri unpredictable; their status as a word-formation process is debated. Manuela Manu Simona Simo (Féry 1997, Wiese 2001, Alber 2006) (Alber 2006) Japanese: English: b. Phonological theory: Prosodic Morphology Mii- c an, Mido- c an Midori step 1: prosodic templates (McCarthy & Prince 1986 et seq., Weeda 1992) Patrick Pat, Patty Yoko- c an, Yoo- c an Yooko Melinda Lyn, Lindy, Linny step 2: Optimality Theory, Generalised Template Theory (GTT, McCarthy & Hana- c an, Haa- c an, Ha c - c an Hanako Elisabeth Bess, Beth, Bette, Prince 1994, 1998, 1999, Benua 1995, 1997) and others (e.g. Downing 2006) (Poser 1984, Mester 1990) Bessie, Betty, Liz, => makes profound predictions concerning the structure of truncated forms in the Lizzy, Libby, Ibby, Lisa, Elis languages of the world (Lappe 2007) BUT: Many of these predictions have neither been described systematically nor have they been tested empirically. To date there exist mainly studies of truncatory b. clippings patterns in individual languages. German: French: (Systematic studies exist mainly for reduplication, cf., e.g., McCarthy & Prince 1999, Downing 2006). Abitur Abi abréviation abrèv Lokomotive Lok habit bit Reproduktion Repro docu documentaire (3) this paper (Ronneberger-Sibold 1992) (Scullen 1997) Swedish: English: • overview: What does Prosodic Morphology predict concerning the structural elektricitet el accumulator ac characteristics and anchoring of truncated forms? raffinaderi raf business biz, bizzo • Testing (some of) the predictions against the data: word structure realisation rea cigarette cig, ret, ciggy, cigga (Nübling 2001) a. crosslinguistic evidence: Which structures are there, out in the world? (Lappe 2007) b. case studies: inventories • Testing (some of) the predictions against the data: anchoring c. compositionality I: fixed segments/affixes – examples from English a. crosslinguistic evidence: Which structures are there, out in the world? -y : no fixed segments b. case studies: inventories Victor Vick-y Victor Vic • theoretical implications: What kind of a theory do we need in order to account Bartholemew Bart-y Bartholemew Bart Melinda Lind-y Lin Melinda for the structure of truncations?

3 4 • focus: not so much the technical details of the OT analysis (cf. Alber & Lappe (7) truncations as minimal prosodic words 2007, Lappe 2007: chpt. 7), but: What is possible in general in truncation? How a. trochee, quantity insensitive: Spanish hypocoristics predictable is the structure of truncation? minimal word: (' σσ ) base hypocoristic Aristobulo Tobo (4) claims: Arminda Minda • Contra earlier claims in Prosodic Morphology, not all truncation corresponds to Umberto Beto the minimal prosodic word. Gilebaldo Balo (Pi Z eros 1998, 2000a,b, Roca & Feliu 2003) • Contra earlier claims in word-formation theory, however, this does not mean b. trochee, quantity-sensitive: Japanese tyan -suffixed hypocoristics that truncation is structurally unpredictable. minimal word: ('H), ('LL) base hypocoristic • The patterns observed are indeed expected if Midori Mido- č an, Mii- č an a. GTT is supplemented by a constraint favouring monosyllabicity and Yooko Yoko- č an, Yoo- č an b. ranking permutations of markedness and anchoring constraints active Hanako Hana- č an, Haa- č an, Ha č - č an in truncation are spelled out and taken seriously (factorial typology, Takako Taka- č an, Taa- č an, Ta č - č an contra earlier, templatic approaches to truncation). (Poser 1984a, b, 1988, Mester 1990: 479, Benua 1995: 117ff.) • Properties of an optimality-theoretic approach that can account for the structure c. iamb, quantity-sensitive: vocatives in Central Alaskan Yupik of truncation: minimal word: ('H), (L'H) a. non-templatic (i.e. no fixed, predefined template shape) base vocative b. process-specific faithfulness or correspondence constraints A  uka  naq A  , A  uk c. markedness ranking relevant for the truncatory process must resemble Nupi  aq Nup, Nupix, Nupik or correspond to the markedness ranking of the language A  iv  an A  if (5) the data Kalixtuq Ka  , Kalik (McCarthy & Prince 1998: 287f.) crosslinguistic (ongoing project, Alber & Lappe 2007) • basic observation no. 2 ( Generalised Template Theory , GTT): truncation patterns extracted from the literature, tagged according to template The minimal prosodic word corresponds to the unmarked prosodic word type/size and anchoring; lit. has been selected from different theoretical (McCarthy & Prince 1994 et seq.). frameworks In OT: The minimal prosodic word is the optimal structure if markedness So far: 76 patterns, spread over 22 languages constraints that are active in the language, are free to exert their influence ( The Emergence of the Unmarked ). case studies English (Lappe 2007), • Which markedness constraints are these? small case studies from Italian (cf. also Alber 2007), German, Estonian a. those constraints which are active in the metrical system of the language ('classic' GTT, e.g. McCarthy & Prince 1994 et seq., Benua 1995, II. Truncation in Prosodic Morphology – the basic assumptions on word 1997, cf. also, e.g., Féry 1997/Wiese 2001 for German, Pineros 1998 structure for Spanish) • 2 basic observations b. a constraint which – in effect – favours monosyllabic templates (cf. • basic observation no. 1: Lappe 2003, 2007 for the problem of generating monosyllabic The structure of truncations very often corresponds to the minimal prosodic templates in truncation; cf., e.g., Alber 2001, Downing 2006 for word of the relevant language (= one metrical foot, McCarthy & Prince 1986). alternative proposals) The minimal prosodic word functions as a template for the truncatory process.

5 6 shorthand: S IZE R EST ('size restrictors') Typically, truncated forms retain material which is prominent in the base form, mainly: (8) example: Spanish hypocoristics (one pattern, there are also others) a. the initial syllable of the base a. markedness constraints which are active in the metrical system of Spanish: b. the (main-)stressed syllable of the base T ROCH , F OOT B INARITY , P ARSE S YLLABLE , A LL F EET R IGHT (11) Spanish hypocoristics – main-stress anchoring b. If these constraints are not dominated by other active constraints, they render a Tóbo - Aristóbulo trochaic foot the optimal word structure. This is the case in hypocoristic formation. Bálo – Gilebáldo c. Ranking – interaction of S IZE R EST with two types of faithfulness constraints. Which constraints determine anchoring? Faithfulness constraints which call for M AX -IO: Every segment in the input has a correspondent in the output. prominent syllables in the base to be retained (A NCHOR - σ 1, A NCHOR - σ STRESS ) M AX -BT: Every segment in the base has a correspondent in the truncated form. example: Italian hypocoristics – initial-syllable anchoring (9) GTT-Ranking for truncation M AX IO >> S IZE R EST >> M AX BT (12) S IZE R EST , A NCHOR - σ 1 >> A NCHOR - σ STRESS >> M AX -BT (10) Spanish hypocoristics: Tobo < Aristobulo (Pi n eros 2000a): /Federíca/ S IZE R EST A NCHOR - A NCHOR - M AX - base: M AX -IO M AX -BT P RSE S YLL F T B IN ( SY A LL -F T -R σ 1 σ STRESS BT a.ris.to.u.lo LL ) � a. (fe.de) * rica a. 2 feet * ! lo b. (ri.ca) *! fede ( a.ris. )( to.o ) b. 1 foot, 1 * ! (13) 3 constraint families (informally): extrametrical syll. aris ( to.o )<lo> S IZE R EST : creates unmarked word structure (esp. foot, syllable) c. monosyllabic *! A NCHOR -P: retains prominent material from the base foot aris ulo M AX -BT : strives to retain as many segments as possible from the base  ( to ) � d . disyllabic foot aris lo ( to.o ) V. Word structure in truncation – the data e. no truncation A. The corpus *!*** * ( a.ris. )( to.u. ) lo (14) existing studies (Alber & Lappe 2007) - no truncation in the language in general: M AX -IO >> F T -B IN , P ARSE - σ , A LL F T L feet, bigger than one syllable: 44 patterns 57.9 % - truncatory template: one ( σσ )-foot: maximal foot: 6 patterns F T -B IN , P ARSE - σ , A LL F T L >> M AX -BT 2 feet: 1 pattern (?) monosyllables: 23 patterns 30.3 % subminimal foot: 2 patterns III. Which part of the base survives in truncation? Anchoring variable word structure: 9 patterns 11.8 % problem: In Prosodic Morphology, is of not much theoretical interest ------------ --------- 76 patterns 100 % => only few systematic studies of individul patterns, no systematic typological investigation