Information structure and choice of perspective in Hungarian narrative discourse: a developmental study Gabriella Fekete Dynamique Du Langage (UMR 5596 CNRS & Université Lyon 2) email@example.com Syntax of the World’s Languages lll, Free University of Berlin, September 25-28, 2008 Narrative production • Organization of events by linguistic expressions • Multi-propositional structure • Coherence • Guide of attention flow in the story Mastery of many linguistic tools BUT Difficulties in the construction of a narration Several linguistic options for the organization of the information flow (Jisa et al. 2002) ↓ Constructions in competition for the same function Berman & Slobin (1994) Dimensions of event construal: (a) selection of topic ; (b) selection of loci of control and effect ; (c) selection of event view ; (d) selection of degree of agency. Distribution of information: - Choice of elements - Attribution of salience - Selection of foreground or background Foundations of a basis of reference order of access important ↓ • “Privilege” of the initial element (Gernsbacher & Hargreaves 1992, Croft 1994) • Initial focus of attention (Langacker 1998) • “Starting point” (MacWhinney 1977) Problem with the terminology „starting point” Languages with fixed word order (English, French): First element = subject/agent = topic = starting point ↓ Equivalents Languages with flexible word order (Hebrew, Spanish, Turkish, Hungarian): First element = subject/agent / direct objet / indirect object In Hungarian: • Not obligatory topic → clauses beginning with the verb • Pro drop+object marking in the verb → clauses containing a verbal form S/A, starting point, topic = not equivalents ↓ Use of the term « perspective » Several devices for the manipulation of perspective (Berman & Slobin 1994): Transitivity (1) a.The boy was frightened because an owl came out. b. The boy was afraid of the owl. c. The owl frightened the boy. Reference form (2) The boy hung on to the antlers of a deer. The deer/he/which/this one ran away. Voice (3) a. The bees chased the dog. b. The dog was chased (by the bees). c. (fr.) Le chien s’enfuit. = The dog ran away. Topicalization, Word order (4) a. As for the frog, the boy saw it. b. (hu) A békát nézte a fiú. = ‘The frog(acc.) saw the boy.’ How do Hungarian children and adults organize the components of information? Which participant do they prefer to take as the perspective? Methodology Subjects • 5 age groups : 3, 5, 7/8, and 11/12 years of age, and adults • 15 subjects in each group • Monolingual Hungarian speakers from middle class backgrounds Task A series of pictures with no text ↓ Elicitation of the narrative 4 episodes treated here 3-yearolds 5-yearolds 7-8yearolds 11-12yearolds adults Total 15 15 15 15 15 75 43 63 68 76 78 328 Mean clauses per subject 2.87 4.2 4.53 5.07 5.2 Range 1.6 2.7 2.8 1.9 3.9 Mean of episodes not mentioned per subject 1.6 0.87 0.53 0.53 0.33 Number of subjects who do not mention all episodes 12 8 7 3 3 Range 1.3 1.2 1.2 2.3 1.2 n Total number of clauses Table 1. Number of subjects, number of clauses encoding the 4 targeted events, mean and range of clauses coded per subject. Characteristics of the Hungarian Language • Agglutinative language of the Finno-Ugrian language family • Pro-drop • Case-marked grammatical relation for every argument (17 cases) • No gender • Object marker in transitive verb forms (2 types of conjugation). (5) a fiú meg-ijeszt-ett egy bagly-ot def boy prev-to frighten-past.3S indef owl-ACC the boy frightened an owl a-ki le-lök-t-e a fá-ról rel-animate prev-to push-past-3SO def tree-delative which pushed [him] out of the tree • No passive construction • Lexicalized verbal form for the middle voice • Left-dislocation → another register • SVO (subject-verb-object) canonical word order - Very flexible - Pragmatically determined (topicfocus-comment) 3 syntactic positions = 3 pragmatic functions • Sentence-initial position → topic • Immediately preverbal position → focus • Postverbal position → background information (comment) • Topic = definite and/or animate NP • Focus = the most information-bearing element Identification : - the strongest accent of the sentence - pre-verb moved after the verb • Post-verbal position = backgrounded / defocused NP Restrictive hierarchy of the position of the argument in perspective Argument in perspective Initial (topic) (6) a. viszont egy ideges vakond but indef nervous mole meg-csíp-t-e prev-to bite-past-3SO az orr-á-t def nose-poss-ACC but a nervous mole has bitten his nose (19;06.d) b. itt meg a kutyá-t here and def dog-ACC a def el-kerget-ik prev-to chase-present-3PO legy-ek fly-pl and here the dog, the flies are chasing it (5;08.f) Argument in perspective Initial (topic) Grammatical (subject/agent) (7) a. mert meg-harap-t-a a orr-á-t because prev-to bite-past-3SO def nose-poss-ACC because [it] bit his nose (5;07.b) b. és itt le-dob-ja and here prev-to throw-present-3SO and [it] throws him here (3;07.c) Argument in perspective Initial (topic) Grammatical (subject/agent) Grammatical (object) (8) ugyanis kerget-ik ideed to chase-present-3PO a def méh-ek bee-pl. indeed, the bees are chasing [it]. (21;07.n) Argument in perspective Initial (topic) Grammatical (subject/agent) Grammatical (object) Post-verbal (9) ott le-dob-t-a a over there prev-to throw-past-3SO def a def szarvas deer kis-fiú-t little-boy-ACC over there, the deer has thrown the little boy (8;01.a) Argument in perspective Initial (topic) Grammatical (subject/agent) Grammatical (object) Post-verbal Pre-verbal (focus) (10) mert az odú-ból egy bagoly jött elő because def hole-elatif indef owl to come.past.3S prev because it was an owl that came out o the hole (11;08.f) Results 70,00% 60,00% 50,00% 3-year-olds 40,00% 5-year-olds 30,00% 7/8-year-olds 11/12-year-olds 20,00% adults 10,00% 0,00% intrans intrans with obl trans other Graph1. Mean (%) of the distribution of intransitive versus transitive clauses in the 4 episodes • Intransitive constructions decrease (F(4,65)=2.323,p=.0658) • Transitive options increase (F(4,65)=2.045,p=.0984) • Intransitive clauses with obliques increase (F(4,65)=.588,p=.6726) Only clauses with at least two participants (transitive clauses, intransitive clauses with oblique(s)) One device alternating perspective in Hungarian ↓ Variations in word order 70,00% 60,00% 50,00% 40,00% 3-year-olds 5-year-olds 30,00% 7/8-year-olds 20,00% 11/12-year-olds adults 10,00% 0,00% initial gramm post-verbal Graph 2. Mean (%) of the distribution of positions of the actor/agent perspective in the clauses with two participants in the 4 episodes • 3-year-olds: grammatical forms (F(4,65)=.768,p=.5496) • 5 and 7/8-year-olds : initial position (F(4,65)=3.022,p=.0238) • 7/8-year-olds: post-verbal position (F(4,65)=2.075,p=.0942) • 11/12-year-olds and adults: alternance of initial and grammatical positions 70,00% 60,00% 50,00% 40,00% 30,00% 20,00% 3-year-olds 10,00% 5-year-olds 7/8-year-olds 0,00% lex lex prim pro sec initial gramm flexion gramm flexion prim sec grammatical lex 11/12-year-olds sec adults post-verbal Graph 3. Mean (%) of the distribution of the position, the characters and the linguistic means used for the actor/agent perspective in the clauses with two participants in the 4 episodes. • Secondary characters = actor/agent (F(4,65)=5.172,p=.0011) • 3-year-olds: grammatical options (F(4,65)=.522,p=.7199) • 5 and 7/8-year-olds: lexical noun phrases (F(4,65)=3.126,p=.0205) • 11/12-year-olds and adults: alternance of grammatical and lexical devices • 11/12-year-olds: pronominals in remarkable proportion (F(4,65)=5.409,p=.0008) 70,00% 60,00% 50,00% 40,00% 3-year-olds 5-year-olds 30,00% 7/8-year-olds 11/12-year-olds 20,00% adults 10,00% 0,00% initial gramm post-verbal verbal Graph 4. Mean (%) of the distribution of positions of the oblique/patient perspective in the clauses with two participants in the 4 episodes • 3-year-olds: grammatical forms (F(4,65)=3.187,p=.0188) • 5 and 7/8-year-olds: initial position (F(4,65)=1.222,p=.3103) • 11/12 ans and adults: initial position 70,00% 60,00% 50,00% 3-year-olds 40,00% 5-year-olds 30,00% 7/8-year-olds 11/12-year-olds 20,00% adults 10,00% 0,00% lex pro gramm flexion prim prim initial grammatical lex pro prim post-verbal lex pré-verb sec prim verbal Graph 5. Mean (%) of the distribution of the position, the characters and the linguistic means used for the oblique/patient perspective in the clauses with two participants in the 4 episodes. • • • • • Primary characters = oblique/patient (F(4,65)=1.322,p=.2713) 3-year-olds: grammatical options (F(4,65)=3.187,p=.0188) 5 and 7/8-year-olds: lexical noun phrases (F(4,65)=.685,p=.6050) 11/12-year-olds and adults: lexical noun phrases Adults : pronominals in significative proportion (F(4,65)=1.700,p=.1607) Discussion • Clauses with two participants: increase with age • 3 and 5-year-olds: intransitive clauses • 7/8 and 11/12-year-olds: intransitive and transitive clauses • Adults: transitive clauses • Secondary characters = actor/agent • Primary characters = oblique/patient ↓ Secondary characters = do the action Primary characters = affected by the action • Oblique/patient perspective → increases with age ! 3-year-olds = appearance of word order which take the oblique/patient in perspective • 3-year-olds = grammatical forms for the perspective • 5 and 7/8-year-olds = lexical noun phrases whatever the perspective • 11/12-year-olds and adults = alternation of the 2 linguistics tools for the actor/agent, lexical noun phrases for the oblique/patient • Pronominal oblique/patient at the beginning of sentences in 11/12-year-olds and adults = surprising ↓ In Hungarian, personal pronouns used with a tonic function ↓ Synthesis of parallel actions of the two protagonists, thus contrasted • Different linguistic tools depending on the age groups → no mastery of the conventional rules of referential coherence until the age of 11/12 years ↓ • Resort to different strategies : - thematic subject strategy (pronominal forms to refer to the main character irrespective of the function), - nominal strategy (full nominal even for maintaining characters) ) - and anaphoric strategy ( pronominals for maintaining reference but nominals for switching). (Karmiloff-Smith 1981, Wigglesworth 1997). • Position of the arguments in perspective → link to the strategies mentionned above 3-year-olds = actor/agent or oblique/patient integrated in the verbal form 5 and 7/8-year-olds = actor/agent or oblique/patient in initial position • Post-verbal position attested in the 7-8 year olds ↓ Actor/agent taken in background ↓ Strong topicalization is compensated 11/12-year-olds and adults = initial and grammatical positions for the actor/agent and initial position for the oblique/patient Conclusion • 3-year-olds = attempt to alternate perspectives but exclusively with verbal forms integrating the affected character • 5-year-olds = mastery already unsteady of the use of the different ways to encode the actions • 7/8-year-olds = “true” variation of the canonical word order for pragmatic reasons • From 7/8-year-olds = initial position favoured for the argument in perspective or its integration in the verbal form → choice depends on the discursive function of the argument. • Linguistic means selected to package the information properly encoded to discursive functions → difficult to control before 11/12-year-olds. ↓ • The establishment of the referential coherence not perfectly mastered by the children ↓ • Use of different strategies (Karmiloff-Smith1981, Wiglesworth, 1997, Fekete 2008) - toddlers = pronominal forms (thematic strategy) - oldest children = nominals (nominal strategy) - adults = coordination of these two strategies (anaphoric strategy) • 7/8-year-olds = particular concerning the combination of the linguistic means favoured and the position employed for the argument in perspective ↓ Post-verbal position for the lexical AC/AG ↓ At the same time resort to the nominal strategy, and try to compensate the difficulties of the referential task with the help of the pragmatic functions of word order. ↓ Solution for the excessive lexicalization at the beginning of the sentence → manipulation of the referents’ order ↓ This is another solution, which they already master, to put the chosen element in background. • Capacity of all the children to put the patients of the action in perspective, using different linguistic and pragmatic tools • Most difficulties in the application of the conventional rules of narration • Berman, R. Slobin, D. I. (Eds.) (1994) Relating Events in Narrative: A Crosslinguistic Developmental Study. Hillsdale, NJ:Erlbaum. • Croft, W. (1994) Voice: beyond control and affectedness. In Hopper, P. & Fox, B. Voice: Form and Function. pp. 89-117. Amsterdam: John Benjamins. • Fekete, Gabriella (to appear 2008). Referential cohesion in Hungarian: a developmental study. • GERNSBACHER, M. A., HARGREAVES, D. (1992) The privilege of primacy: Experimental data and cognitive explanations. In Payne, D. L. Pragmatics of word order flexibility. pp. 83-116. Philadelphia: John Benjamins. • Jisa, H., Reilly, J., Verheoven, L., Baruch, E. & Rosado, E. (2002) "Crosslinguistic perspectives on the use of passive constructions in written texts." Journal of Written Language and Literacy, 5, 163-81. • Karmiloff-Smith, A. (1981) The grammatical marking of thematic structure in the development of language production. In Deutsch, W. (Ed.). The child’s construction of language. New York: Academic Press, 121-147. • Langacker, R. W. (1998) Conceptualization, Symbolization, and Grammar. In Tomasello, M. The new psychology of language : Cognitive and functional approaches to language structure. pp. 1-39. Mahwah, NJ : Lawrence Erlbaum • MacWhinney, B. (1977) Starting points. In Language, 53. pp. 152-168. • Mayer, M. (1969) Frog, Where are you? Amsterdam : Dial Press. • Strömqvist, S., Verhoeven, L. (Eds.) (2003) Relating events in narrative – typological and contextual perspectives. Mahwah, New Jersey : Lawrence Erlbaum Associates Publishers. • Wigglesworth, G. (1997) Children’s individual approaches to the organization of narrative. In Journal of Child Language 24: 279-309.