Assignment 7 Turkish: Some Aspects of Turkish Word Order I. Basic Phrase Structure NOTE on orthography: The problem god apparently does not want to mess with actual Turkish orthography, which has several symbols that he does not have on his typewriter; so he uses these substitutions: C = c cedilla = a voiceless palatal affricate (c = c = a voiced palatal affricate) S = s cedilla = a voiceless palatal fricative O = o umlaut = a low front rounded vowel U = u umlaut = a high front rounded vowel I = undotted i = a high back unrounded vowel G = g breve = positionally a consonant, but with no direct realization; alternates sometimes with /k/ NOTE on Vowel Harmony: in (most) Turkish suffixes, the backness of the suffix vowel is determined by the backness of the preceding vowel; for high vowels in suffixes, there is a similar assimilation in rounding. A result of this is that suffix vowels generally need to be specified only L (low) or H (high). There are some other fairly obvious alternations, such as that a suffix- initial stop assimilates in voicing to the preceding segment; and there is systematic syllable-final stop devoicing. A. Some Phrases (1) maymun "the monkey" (2) mutlu maymun "the happy monkey" (3) gUzel mutlu maymun "the beautiful happy monkey" (4) bu (Su, o) mutlu maymun "this (that, yonder) happy monkey" (5) *bu maymun mutlu } (6) *gUzel maymun mutlu } These are grammatical, but only (7) *mutlu bu maymun } as sentences, not as DPs or phrases (8) *bu Su maymun } of any type. (9) *maymun mutlu bu (10) muz "the banana" (11) lezzetli muz "the delicious banana" (12) pahalI araba "the expensive car" (13) bu gUzel araba "this beautiful car" (14) Su pahalI araba "this/that expensive car" (15) maymun-un muz-u "the monkey's banana" (16) maymun iCin "for the monkey" (17) bu pahalI araba ile "with this expensive car" (18) Su pahalI maymun-un uzun kuyruG-u gibi "like that expensive monkey's long tail" (19) gUzel kIz-In pahalI maymun-u-nun lezzetli muz-u iCin "for the beautiful girl's expensive monkey's tasty banana" (20) bir maymun "a monkey" (21) mutlu bir maymun "a happy monkey" (22) Cok mutlu bir maymun "a very happy monkey" (23) bir muz "a banana" (24) pek pahalI bir araba "too expensive a car" (25) biraz pahalI bir araba "a somewhat expensive car" (26) *bir mutlu maymun (27) *bir Cok gUzel kIz (28) *Su bir pahalI maymun This is just to give you a feel for what some simple phrases are like in this language. Looking ahead if necessary to section B, propose structures for (18) and (19). From what you see, what can you say about (a) head-complement order in this language? (b) specifier-head order in this language? (c) modified-modifier order in this language? B. We are going to look at some simple sentence constructions in Turkish, but before we do that here is a quick intro to parts of the morphology; Case: There are six morphological cases in Turkish: Nominative (NOM) -0 Genitive (GEN) -nHn Accusative (ACC) -yH Dative (DAT) -DL Ablative (ABL) -DLn Note that the Nominative (sometimes called "bare") case has no overt marking. Like most suffixes in Turkish, those with high vowels (H) undergo four-way vowel harmony (assimilation in backness and rounding) triggered by the preceding vowel; low vowels (L) are always unround and assimilate only in backness. There are also some consonant alternations, most notably the assimilation in voicing of a suffix-initial stop to the preceding segment. It will be useful to know that a suffix-initial /y/ always vanishes when preceded by a consonant. Verb morphology: Turkish has just two Tenses, Past and Conditional; and five Aspects: Future, Progressive, Aorist, Evidential, and Necessitative: T: Past (PST) -DH Cond (CND) -sL ASP: Future (FUT) -yLcLG Prog (PRG) -Hyor Aorist (AOR) -Lr (BUT -Hr after polysyllabic stems) Evid (EVD) -mHS Necess (NEC) -mLlH Subject Agreement: There are two sets of Agreement suffixes, depending on whether the predicate is "verbal" or not (diagnosed by whether it ends in a T morpheme): Verbal agreement (after PST or CND) Nonverbal agreement (elsewhere) sg pl sg pl 1st -m -k -Hm -Hz 2nd -n -nHz -sHn -sHnHz 3rd -0 -lLr -0 -lLr Genitive-Possessive Agreement There is a third pattern of agreement, which holds between possessive heads and their specifiers (e.g. maymun-un muz-u): sg pl 1st -m -mHz } These are the forms that the 2nd -n -nHz } possessive suffix takes in 3rd -(s)H(n) -lLrH } agreement with the possessor. C. Now we will look at some simple sentence structures in Turkish: (29) mutlu maymun gUlUyor "the happy monkey is laughing" (30) pahalI muz dUStU "the expensive banana fell" (31) Su gayet gUzel domuz Olecek "that rather beautiful pig will die" (32) bUyUk aGaC dUSmUS "the big tree fell (evidently)" (33) gUzel bir kIz geliyor "a beautiful girl is coming" (34) maymun muzu istiyor "the monkey wants the banana" (35) muz maymunu istiyor "the banana wants the monkey" (36) maymun muz ile oynuyor "the monkey is playing with the banana" (37) gUzel maymun mutlu domuzu gOrUyor "the beautiful monkey sees the happy pig" (38) maymun arabadan korktu "the monkey was afraid of the car" (39) gUzel kIz arabadan CIkIyordu "the beautiful girl was getting out of the car" (40) gUzel kIz o Cirkin maymun ile konuSacaktI "the beautiful girl was going to talk with that ugly monkey" (41) gUzel kIz sarhoS bir maymun gibi davranIyor "the beautiful girl is behaving like a drunken monkey" (42) maymun muza bakIyor "the monkey is looking at the banana" (43) gUzel kIz maymuna gUlUyor "the beautiful girl is laughing at the monkey" (44) maymun gUzel kIzIn muzu iCin aGaca CIkIyordu "the monkey was climbing the tree for the beautiful girl's banana" (45) gUzel kIz maymun iCin SarkIyI sOyleyecekse "if the beautiful girl was going to sing the song for the monkey" (46) maymun masada oturuyor "the monkey is sitting on the table" (47) maymun muz ile odaya giriyor [NOT ambiguous] "the monkey is entering the room with the banana" [monkey has banana] (48) gUzel kIz arabada oturuyor "the beautiful girl is sitting in the car" (49) Su muz masada kalacak "that banana will remain on the table" (50) Cok pahalI bir maymun o bUyUk aGaCtan dUSerse "if a very expensive monkey were to fall from that big tree" (51) o adam Cirkin bir maymuna benziyor "that man resembles an ugly monkey" (52) talebeler bu Odevden korkmalIlar "the students need to fear this assignment" (53) talebeler Odeve baSladIlar "the students began the lesson" (54) maymun mutlu "the monkey is happy" (55) muz masada "the banana is on the table" (56) Su muz maymun iCin "this/that banana is for the monkey" (57) bu su buz gibi "this water is like ice" (58) o maymun gayet gUzel bir hayvan "that monkey is a quite beautiful animal" (59) AySe Cok gUzel bir kIz "AySe is a very beautiful girl" (60) (sen) Cok iyi bir talebesin "you are a very good student" (61) (biz) bahCedeydik 'we were in the garden' (62) maymun her zaman domuzu sataSIyor 'the monkey always teases the pig' (63) domuz her zaman maymun tarafIndan sataSIlIyor 'the pig is always being teased by the monkey' Based on this, what would you say about basic sentence structure in Turkish? Would you be willing to offer a guess as to the probable structure of TP and VP? Any observations about subcategorization? What would account for the ordering relations observed between T and ASP? Where do you suppose the subject is in surface structure, and why? II. Turkish Passives A. Personal Passives (64) maymun bavulu aCtI 'the monkey opened the suitcase' (65) bavul maymun tarafIndan aCIldI 'the suitcase was opened by the monkey' (66) kasap eti kesti 'the butcher cut the meat' (67) et (kasap tarafIndan) kesildi (68) evimizi satacaGIz 'we're gonna sell our house' (69) evimiz satIlacak (70) herkes durumu anladI 'everybody understood the situation' (71) durum (herkes tarafIndan) anlandI (72) gUzel kIz maymunun muzunu gOrdU 'the beautiful girl saw the monkey's banana' (73) maymunun muzu (gUzel kIz tarafIndan) gOrUldU (74) maymun beni IsIrdI 'the monkey bit me' (7x) maymun tarafIndan IsIrIldIm (75) yIlan seni sokacak 'the snake will bite you' (76) yIlan tarafIndan sokulacaksIn From this, describe Turkish passives as best you can. B. Impersonal Passives (77) bu barda insanlar pek Cok baGIrIr 'in this bar people yell too much' (78) bu barda pek Cok (*insanlar tarafIndan) baGIrIlIr (79) insanlar pek Cok bu gOlde boGulur 'people drown too much in this lake' (80) bu gOlde pek Cok (*insanlar tarafIndan) boGulunur (81) harbin zamanInda Pariste insanlar OldUler 'in the time of the war people died in Paris' (82) harbin zamanInda Pariste (*insanlar tarafIndan) OlUndU (83) buradan bu dolmuS ile Universiteye CIkarIm 'from here with this taxi I'll go up to the university' (84) buradan bu dolmuS ile (*benim tarafImdan) Universiteye CIkIlIr (85) OksUz Cocuklar bu yetimhanede Cabuk bUyUrler 'orphan children grow up fast in this orphanage' (86) bu yetimhanede (*OksUz Cocuklar tarafIndan) Cabuk bUyUnUr (87) insan burada hastalanIr 'a person gets sick here' (88) burada (*insan tarafIndan) hastalanIlIr (89) talebeler derse baSladIlar 'the students began the lesson' (90) derse (*talebeler tarafIndan) baSlandI (91) herkes yIlanlardan korkar 'everybody fears snakes' (92) yIlanlardan (*herkes tarafIndan) korkulur Now what is your analysis of impersonal passives in Turkish? How does it compare with the analysis offered by Perlmutter in "Impersonal Passives and the Unaccusative Hypothesis"? (You probably haven't read that, but you should.) III. Scrambling (That's what it's called) A. If you take a simple Turkish sentence with four parts: (104) kIz maymuna muzu verdi believe it or not, all 24 possible orders are grammatical: (105) kIz maymuna muzu verdi maymuna kIz muzu verdi muzu kIz maymuna verdi kIz muzu maymuna verdi muzu maymuna kIz verdi maymuna muzu kIz verdi maymuna muzu verdi kIz muzu maymuna verdi kIz kIz muzu verdi maymuna muzu kIz verdi maymuna kIz maymuna verdi muzu maymuna kIz verdi muzu kIz verdi maymuna muzu kIz verdi muzu maymuna maymuna verdi kIz muzu maymuna verdi muzu kIz muzu verdi kIz maymuna muzu verdi maymuna kIz verdi kIz muzu maymuna verdi kIz maymuna muzu verdi muzu kIz maymuna verdi muzu maymuna kIz verdi maymuna kIz muzu verdi maymuna muzu kIz I won't give all the examples to show it, but the nominals in these examples (kIz, maymuna, muzu) are representative of (case-marked) DPs, and it's phrases that participate in this scrambling. They don't usually break up. B. The next task will be to try to sort out what is happening here. First, here are some sentences with embedded clauses: (106) (ben) (sen-in) maymun-a muz-u ver-diG-in-i san-dI-m I you-GEN monkey-DAT banana-ACC give-PRT-2SG-ACC think-PST-1SG 'I thought you gave the banana to the monkey.' I'm not going to gloss all of these examples, so study this one carefully. The only unfamiliar thing should be that -diG- suffix that is glossed PRT; it's a participle marker that fits a verb out to be embedded (something like a nominalizer). The subject of the embedded clause appears in the Genitive case and the participialized verb agrees with it, using the Gen-Poss agreement pattern. (107) (siz) (biz-im) yIlan-lar-dan kork-tuG-umuz-a SaSIr-mIS-sInIz you we-GEN snake-PL-ABL fear-PRT-1PL-DAT be-surprised-EVD-2PL 'You were apparently surprised at our being afraid of snakes.' (108) CocuklarIn imtihandan korktuklarI belli oldu. 'It became clear that the children feared the test.' (109) CocuklarIn imtihandan korktuklarInIn belli olduGunu beklettik. 'We expected that it would become clear that the kids feared the test.' Now let's embed (105): (110) (biz) kIzIn maymuna muzu verdiGini sandIk Here's an interesting fact: not all of the scrambled orders are permissible in the embedded clause. In fact, most of the orders exhibited in (105) are ungrammatical; here's one example: (111) *(biz) kIzIn muzu verdiGini maymuna sandIk Here are *all* of the grammatical orders in the embedded clause: kIzIn maymuna muzu verdiGini sandIk maymuna kIzIn muzu verdiGini sandIk muzu kIzIn maymuna verdiGini sandIk kIzIn muzu maymuna verdiGini sandIk muzu maymuna kIzIn verdiGini sandIk maymuna muzu kIzIn verdiGini sandIk What is the generalization? C. Here's another piece of the puzzle: maymuna muzu verdiGini sandIk kIzIn kIzIn muzu verdiGini sandIk maymuna kIzIn maymuna verdiGini sandIk muzu muzu maymuna verdiGini sandIk kIzIn maymuna kIzIn verdiGini sandIk muzu muzu kIzIn verdiGini sandIk maymuna muzu verdiGini sandIk kIzIn maymuna muzu verdiGini sandIk maymuna kIzIn maymuna verdiGini sandIk muzu kIzIn maymuna verdiGini sandIk kIzIn muzu kIzIn verdiGini sandIk muzu maymuna kIzIn verdiGini sandIk maymuna muzu verdiGini sandIk kIzIn maymuna muzu verdiGini sandIk maymuna kIzIn muzu verdiGini sandIk muzu kIzIn maymuna verdiGini sandIk kIzIn muzu maymuna verdiGini sandIk muzu maymuna kIzIn verdiGini sandIk maymuna muzu kIzIn These are all grammatical. What do you conclude from this? Now that you've earned it, here's a fact: any element that appears after the verb must be discourse-presupposed, i.e. some kind of old information recoverable from previous discourse or the situation. Placing an element in post-verbal position is sometimes called "Defocus". D. In a similar way, if anything other than the verb's internal argument is in immediate pre-verbal position, that element is "Focused": it bears the main sentence stress and is contrastive or new information. This true whether or not the clause is embedded: A: maymuna muzu kIz mI verdi, Cocuk mu? 'Did the GIRL gave the banana to the monkey, or the BOY?' B: maymuna muzu kIz verdi maymuna muzu kIzIn verdiGini sandIk kIzIn verdiGini sandIk maymuna muzu A focused element never gets to leave its own clause: (112) kIzIn maymuna muzu verdiGini biz sandIk (113) *kIzIn maymuna muzu biz verdiGini sandIk (114) *kIzIn maymuna verdiGini muzu sandIk Propose an analysis of Focus movement. Note any difficulties. E. A last piece of the puzzle: If an element is "Topical" (i.e. not new information in the discourse, but selected for current attention) it *can* sometimes leave its clause: (115) Hasan kIzIn maymuna muzu verdiGini sandI (116) maymuna Hasan kIzIn muzu verdiGini sandI (117) muzu Hasan kIzIn maymuna verdiGini sandI (118) muzu Hasan kIzIn maymuna verdiGinin belli olduGunu sandI (119) maymuna Hasan kIzIn muzu verdiGinin belli olduGunu sandI Assume that this is just like Topicalization in English, and propose an analysis that works equally well for both languages. Now what is "scrambling"? III. WH Questions (120) muzu kim yedi? kim muzu yedi? *muzu yedi kim? "Who ate the banana?" (121) maymun ne yedi? *ne maymun yedi *maymun yedi ne? "What did the monkey eat?" (122) kIz muzu kime verdi? *kime kIz muzu verdi? *kIz muzu verdi kime? "Who did the girl give the banana to?" (123) (siz) kIzIn maymuna ne verdiGini sandInIz? "What did you think the girl gave to the monkey?" (124) (siz) kIzIn muzu kime verdiGini sandInIz? "Who did you think the girl gave the banana to?" (125) (siz) kImIn maymuna muzu verdiGini sandInIz? (siz) maymuna muzu kImIn verdiGini sandInIz? "Who did you think gave the banana to the monkey?" The following are absolutely impossible: (123') *ne (siz) kIzIn maymuna verdiGini sandInIz? (123'') *(siz) kIzIn maymuna verdiGini sandInIz ne? (124') *kime (siz) kIzIn muzu verdiGini sandInIz? (124'') *(siz) kIzIn muzu verdiGini sandInIz kime? (125') *kimin siz maymuna muzu verdiGini sandInIz? (125'') *siz maymuna muzu verdiGini sandInIz kimin? What analysis would you propose for WH Questions in Turkish?