Passives CAS LX 500 A1 Language Acquisition
Week 9a. BUCLD 33
Messenger English-speaking children's comprehension of the passive appears Crawford semantically constrained: 4-5-year-olds comprehend agent-patient The Maturation Hypothesis (MH) predicts cross-linguistic delay in the actional passives better than experiencer-theme non-actional passives acquisition of verbal passives (Borer and Wexler 1987). Demuth (1989) (Maratsos et al, 1985). These results might reflect children's difficulty argues against MH, finding early passive use in Sesotho. This paper interpreting non-actional event pictures, or they might have genuine adds experimental results to the debate. 11 Sesotho-speaking children difficulties interpreting such sentences. We present two experiments (5;0-6;0) and 10 adults took a Sesotho version of the two-choice investigating this: Experiment 1 tested children's comprehension of picture-selection task. Four actional verbs were tested. ANOVAs agent-patient, experiencer-theme and theme-experiencer verb active crossing voice (actives/long passives) with age show a main effect and passives in a sentence-picture-matching task. We found a (ME) of voice, ME of age, and interaction of voice and age. Comparing significant effect of structure (p<.001) and verb-type (p<.001): actives to short passives, there is a ME of voice, but no interaction Children understood actives better than passives and actional and between age and voice. ANOVAs crossing length and age reveal long theme-experiencer passives better than experiencer-theme passives. passives were disproportionately difficult. These data suggest Sesotho Experiment 2 tested the same factors using structural priming (Bock, children comprehend short actional passives, but not long actional 1986). We found a reliable effect of structure (p<.01) but not verb-type passives. I argue the poor performance on long actional passives is due (F<2): Children were more likely to produce a passive description to their being unambiguously verbal, while adult-like performance on after hearing a passive prime than an active, irrespective of the prime- short passives is due to differing agreement morphology. verb. Our priming results suggest task effects may confound the semantic effects in sentence-picture-matching experiments.
Pragmatics: de Villiers In relevance implicatures the central meaning in an exchange of Implicatures utterances is left unsaid and the observer has to fill it in from what s/ he knows about the topic under discussion and from reading the speakers' communicative intentions (Sperber & Wilson, 1986). So children with autism who have a delay in their theory of mind understanding may also have difficulty in understanding relevance implicatures despite the absence of a misleading literal meaning. In a study of the comprehension of relevance implicatures in 38 typically- developing children and 10 children diagnosed with "high functioning" autism there was strong developmental growth in the children's ability to report on the intended meaning of the speakers' responses. The children with autism lagged significantly behind the typically- developing children, though their vocabulary scores and non-verbal IQs were in the normal range. The best predictor of the autistic children's understanding of relevance implicatures was their theory of mind performance. !"#$%&%'%()*(+*,-.-/0)"-*123.%"0'$4-&*0)5*6(5$.04%'7% % )*+%,-.-/0)"-*%23.%"0'$4-8%% Verbuk %",%-.%/01%2'3%45"65%"%(06$%07%58$%(06$9:";$<%% %%%=,%>.%?@%935%'7%"%;'"5%"71%=6'3#85%2'3%'7$,A%% On the Gricean "language-based" account, conversational %;,%>B4%C$:$D"7;$%0E9:0;"536$.%@%101%7'5%45"65%"%(06$,%%%% implicatures are computed by engaging in linguistic reasoning, and % )F+%,-.-/0)"-*123.%"0'$4-8%%%%% are viewed as part of the language module. On Kasher's (1991) G"5%5':1%/'#H%?/'#H%7'I%05B4%2'36%5367%5'%1'%4'E$5807#,%@5B4%;80::2%8$6$,%J5"65%"%(06$%07%58$%% Rationality account, conversational implicatures are produced by the (06$9:";$H%9:$"4$,%@B::%#'%8"D$%4'E$5807#%5'%$"5%07%58$%K05;8$7,A%% non-linguistic competence, and are viewed as derived by rationality- based reasoning, also instrumental in deriving non-linguistic inferences. English-speaking children (5;1-8;1) were tested on computing Relevance implicatures and parallel in nature non- linguistic inferences. Because children engage in linguistic reasoning in computing Relevance implicatures but not non-linguistic %% % inferences, the Language-based account predicted the latter to be less )L8$%$M9$60E$75$6%48'I4%58$%;80:1%58$%90;536$H%58$7%6$E'D$4%05+,%%% %-(5$6%"%I80:$H%G"5%6$5367$1%5'%58$%:0D07#%6''E,%/'#%4"01H%?@%935%'7%"%;'"5%"71%=6'3#85%2'3%'7$,A%% challenging. On the Rationality-based account, computing both types %N$5B4%562%5'%(0#36$%'35%I8"5%8"99$7$1,%/'%2'3%5807K%/'#%45"65$1%"%(06$<%O'I%;"7%2'3%5$::<% of inferences requires performing rationality-based non-linguistic % reasoning; children were not predicted to perform better on )P+%9():.%);$%&'%"*1)+-4-)"-8%% G"5%5':1%/'#H%?/'#H%7'I%05B4%2'36%5367%5'%1'%4'E$5807#,%@5B4%;80::2%8$6$,%J5"65%"%(06$%07%58$%% computing non-linguistic inferences. Children performed significantly (06$9:";$H%9:$"4$,%@B::%#'%8"D$%4'E$5807#%5'%$"5%07%58$%K05;8$7,A%% better on computing!"#$%&%'%()*(+*,-.-/0)"-*123.%"0'$4-&*0)5*6(5$.04%' non-linguistic inferences, as the Language-based7% % )*+%account,-.-/0)"-*%23.%"0'$4-8 predicted. %%Reasoning about language, specifically, about the %",%-.%/01%2'3%45"65%"%(06$%07%58$%(06$9:";$<%%role of seemingly irrelevant utterances %%%=,%>.%?@%935%'7%"%;'"5%"71%=6'3#85%2'3%'7$,A%% in discourse, constitutes the %;,%>B4%C$:$D"7;$%0E9:0;"536$.%@%101%7'5%45"65%"%(0main acquisition challenge presented6$,%%%% by the Relevance implicatures. % )F+%,-.-/0)"-*123.%"0'$4-8%%%%% G"5%5':1%/'#H%?/'#H%7'I%05B4%2'36%5367%5'%1'%4'E$5807#,%@5B4%;80::2%8$6$,%J5"65%"%(06$%07%58$%% %% % (06$9:";$H%9:$"4$,%@B::%#'%8"D$%4'E$5807#%5'%$"5%07%58$%K05;8$7,A%% %)L8$%$M9$60E$75$6%48'I4%58$%;80:1%58$%90;536$H%58$7%6$E'D$4%05+,%%% %-(5$6%"%I80:$H%G"5%6$5367$1%5'%58$%:0D07#%6''E,%N$5B4%4$$%I8"5%8"99$7$1,%%%
%% % )L8$%$M9$60E$75$6%48'I4%58$%;80:1%58$%90;536$H%58$7%6$E'D$4%05+,%%% % %-(5$6%"%I80:$H%G"5%6$5367$1%5'%58$%:0D07#%6''E,%/'#%4"01H%?@%935%'7%"%;'"5%"71%=6'3#85%2'3%'7$,A%% Design%N$5B4%562%5'%(0#36$%'35%I8"5%8"99$7$1,%/'%2'3%5807K%/'#%45"65$1%"%(06$<%O'I%;"7%2'3%5$::<% %N$5B4%562%5'%(0#36$%'35%I8"5%8"99$7$1,%/'%2'3%5807K%/'#%45"65$1%"%(06$<%O'I%;"7%2'3%5$::<% % % J$:$;5$1%C$($6$7;$4% )P+%9():.%);$%&'%"*1)+-4-)"-8%% ThreeQ60;$H%!,%)*RPR+% conditions:!"#$%&'(%)("*&(+,-(./(+.0$' verbal RelevanceH%O"6D"61%S70D$64052%!6$44H%G"E=601#$H%T"44,%% implicatures; G"5%5':1%/'#H%?/'#H%7'I%05B4%2'36%5367%5'%1'%4'E$5807#,%@5B4%;80::2%8$6$,%J5"65%"%(06$%07%58$%% U"48$6H%-,%)*RRP+%10,23,"%4'5(60%"%4,7(6.)4&8"'9%N'71'7.%C'35:$1#$,%% (06$9:";$H%9:$"4$,%@B::%#'%8"D$%4'E$5807#%5'%$"5%07%58$%K05;8$7,A%% U"48$6H%-,%)*RR*+%?!6"#E"50;4%"71%58$%T'13:"6052%'( parallel non-verbal%T071,A%@7%J,%/"D04%$14,H% inferences; 10,23,"%4'H%VM('61% S70D$64052%!6$44H%WXFYWP&,%% scalar implicatures. *
Group one: !"#$%&%'(%&% Picture stories: 1 success, 2 failures. %% % %)L8$%$M9$60E$75$6%48'I4%58$%;80:1%58$%90;536$H%58$7%6$E'D$4%05+,%%% Stories: 2 successes, 1 failure. %-(5$6%"%I80:$H%G"5%6$5367$1%5'%58$%:0D07#%6''E,%N$5B4%4$$%I8"5%8"99$7$1,%%% 3 scalar implicature items.
Group two: Picture stories: 2 successes, 1 failure. Stories: 1 success, 2 failures. 3 scalar implicature items. % %N$5B4%562%5'%(0#36$%'35%I8"5%8"99$7$1,%/'%2'3%5807K%/'#%45"65$1%"%(06$<%O'I%;"7%2'3%5$::<% 10 % J$:$;5$1%C$($6$7;$4% Q60;$H%!,%)*RPR+%!"#$%&'(%)("*&(+,-(./(+.0$'H%O"6D"61%S70D$64052%!6$44H%G"E=601#$H%T"44,%% U"48$6H%-,%)*RRP+%10,23,"%4'5(60%"%4,7(6.)4&8"'9%N'71'7.%C'35:$1#$,%% U"48$6H%-,%)*RR*+%?!6"#E"50;4%"71%58$%T'13:"6052%'(%T071,A%@7%J,%/"D04%$14,H%10,23,"%4'H%VM('61% S70D$64052%!6$44H%WXFYWP&,%% *
!"#$%&%'(%&% Subjects Scenarios
28 children aged 5;1-8;1 were tested.
10 5-year-olds, 7 6-year-olds, 9 7-year-olds and 2 8-year-olds.
This is Cat and this is Dog. Cat and Dog are friends and live in the same house. One weekend, they decided to do a bunch of chores around the house. They decided that Cat would do half the chores, and Dog would do the other half of the chores. Cat and Dog had to do some cleaning, cooking, fix some furniture, and do some other stuff. 11 12 (1) Picture Story. THE PARROT. After a while, Dog asked Cat to come back to the kitchen. Cat told Dog, “Dog, you start doing things, and I’ll go next. Feed Let’s see what happened. the parrot the birdseed, please. I’ll sit in the living room and read.”
Do you think Dog fed the parrot? How can you tell from this 13 picture? 14
(1) Story. THE PARROT. After a while, Dog asked Cat to come back to the kitchen. Cat told Dog, “Dog, you start doing things, and I’ll go next. Feed Dog said, “I put the empty bowl back under the table.” the parrot the birdseed, please. I’ll sit in the living room and read.” Let’s try to figure out what happened.
Do you think Dog fed the parrot?
How can you tell? ************************************************************************
Dog’s utterance addresses the implicit QUD, “Did you feed the parrot?”
15 16
(2) Picture Story. THE WATERING CAN. After a while, Cat told Dog to come outside. Let’s see what Dog told Cat, “Cat, now it’s your turn to do something. Fix our happened. watering can, please. I’ll go and read a magazine in my room.”
Let’s try to figure out what happened. Do you think Cat fixed the watering can? How can you tell from this picture? 17 18 (2) Story. THE WATERING CAN. After a while, Cat told Dog to come outside. Cat said, “I am Dog told Cat, “Cat, now it’s your turn to do something. Fix our watering the grass from our sprinkler.” watering can, please. I’ll go and read a magazine in my room.” Let’s try to figure out what happened.
Do you think Cat fixed the watering can?
How can you tell?
19 20
(3) Picture Story. THE TV STAND. After a while, Cat asked Dog to come back into the house. Let’s Dog told Cat, “Cat, now it’s your turn to do something. Fix our see what happened. television stand, please. I’ll go outside and sit on the grass.”
Let’s try to figure out what happened. Do you think Cat fixed 21 the television stand? How can you tell from this picture?22
(3) Story. THE TV STAND. After a while, Cat asked Dog to come back into the house. Dog told Cat, “Cat, now it’s your turn to do something. Fix our Cat said, “I put the television on our big rug.” television stand, please. I’ll go outside and sit on the grass.” Let’s try to figure out what happened.
Do you think Cat fixed the television stand?
How can you tell?
23 24 (4) Scalar Implicature. THE WINDOW. Dog cleaned the right corner of the window.
Cat told Dog, “The window in the living room is really dirty, we Did Dog do everything Cat asked him to do? haven’t cleaned it since last year. Dog, clean the window, please.” How can you tell?
25 26
Results Children’s Performance on Computing Relevance Implicatures The Neo-Gricean account: children will do better on computing vs. Non-Verbal Inferences non-linguistic than linguistic inferences. 2.39 2.400 The Rationality account: children will do the same on both. 2.213
Children did significantly better on the non-verbal than the 2.025 verbal condition. ANOVA pairwise comparisons: p = .001. 1.82 The prediction of the neo-Gricean account was supported. 1.838
1.650 Mean # of correct responses out of 3 test items: Relevance verbal: M=1.8214; SD=.90487; N=28; Non-verbal non-verbal: M=2.3929; SD=.78595; N=28;
27 28
Children’s Performance on Computing Relevance Implicatures vs. Scalar Implicatures Lexical acquisition
3.0
2.5
2.0 1.82 1.85
1.5
1.0
0.5
0 Relevance Scalar
29 Gavruseva Family Recent inquiries into word learning (e.g. Mintz & Gleitman 2002) show Persian has a deceptively small repertoire of about 160 simple verbs that 2- and 3-year olds fail to extract and generalize adjectival meaning which belies an intricate system of light verb constructions (LVC). LVCs if multiple exemplars are labeled with the proforms 'one' or 'thing'. range from transparent to idiomatic, with most constructions occurring M&G's results are interesting because the syntactic frame "really somewhere in between. Only a handful of lexical items are available to ______" rules out nominal interpretations (*'really dog'). This paper productively express thousands of verbal notions. The current study asks whether children (ages 2-4) use syntactic cues to infer adjectival presents new data gathered on a weekly basis from four Persian- meaning by comparing their responses to the stimuli presented in speaking children between the ages of 1;11 and 5;0 over a six-month English and Russian. Another question is whether adjective-specific period. First-learned general verbs later serve as a basis for hundreds of morphology in Russian provides an edge in word learning. Two sets of different verbal notions. We calculated token and type frequencies of experiments were conducted, with children hearing novel words in LVCs as well as of simple verbs and describe their progressive emergence predicative position and a structure that requires like categories in different LVCs. We investigate the proper usage of LVs by Persian (coordination frame). The cross-linguistic comparisons indicate that the speaking children through analysis of errors, highlighting patterns in the type of cue (morphosyntactic or syntactic) does not have a significant acquisition of the semantic organization of LVCs in Persian. Our results effect on adjective acquisition if word learning occurs in the context of support a usage-based theory of language acquisition. multiple exemplars introduced by a lexical noun.
Foley An act-out task tested ten children (5;0-7;11) and ten adults in sentences like (1-2): Syntax: other (1) Tensed relative clause a. Eeyore pushes the boat [ behind which Pooh runs t ] b. Eeyore touches the boat [ which Pooh runs behind t ] (2) Infinitival relative clause a. Pooh picks the blanket [ under which to rest t ] b. Pooh picks the blanket [Op to rest under t ] A qualitative analysis showed that children interpreted pied-piping and preposition-stranding with similar success. However, responses revealed an unexpected eventive-type reading for (1a), but not (3). (3) Bert kicks the box [ under which Cheezer hides t ] Eventive reading, for (1a): Eeyore pushes boat; then Pooh begins running behind Eeyore and boat Non-eventive reading, for (3): Cheezer hides under box, then Bert kicks box Children thus draw on prepositional semantics and pragmatics in selecting relative clause attachment level (noun vs. clause).
References Brown, Roger W. (1973). A first language: The early stages. Cambridge, Mass.: Harvard University Press. Cazden, Courtney B. (1968). The acquisition of noun and verb inflections. Child Development, 39, 433-448. Chomsky, Noam. (1981). Lectures on government and binding: The Pisa lectures. Berlin, Federal Republic of Germany: Mouton de Gruyter. Pesetsky, David, & Torrego, Esther. (2001). T to C movement: Causes & consequences. In M. Kenstowicz (Ed.), Ken Hale: A life in language (pp. 355-426). Cambridge: MIT Press. Rizzi, Luigi. (1996). Residual verb second and the wh-criterion. In A. Belletti & L. Rizzi (Eds.), Parameters and functional heads: Essays in comparative syntax (pp. pp 63-90). New York: NY: Oxford U Press. Travis, Lisa. (1984). Parameters and effects of word order variation. Unpublished Doctoral Dissertation, MIT, Cambridge.
Grinstead V-to-I-to-C movement is investigated in child English speakers. Our
hypothesis is that children should mark verbs as finite to a greater or
equal extent than they invert subjects and verbs. We measure these abilities using paired grammaticality judgment tasks. In experiments
1 and 2, a group of child English-speakers (n = 63, mean age = 4;10, Figure 1 – Direct V-to-C Movement Prevented by the Head Movement Constraint range = 3;0-6;11, SD = 9.04 months) judges two constructions: root finite verbs and wh- questions. Most children's scores were higher for finiteness than for inversion (33 of 46 - 72%). A paired comparison of children's finiteness and inversion scores showed finiteness to be significantly better (paired t-test, t[45] = -2.788, p = . 008, two-tailed), consistent with independent verb movement to IP as a prerequisite for movement to CP. Strengthening this claim is the fact that children's finiteness and inversion scores were highly correlated (Pearson correlation = .565, p < .000).
Figure 2 – Inversion and Finiteness Scores (n = 46, Pearson Correlation = .565, p < .000) Kline Jin The question of when young children gain access to syntactic Language has various devices to convey sentence meaning. In English, knowledge has been hotly debated (c.f. Tomasello (2000), Fisher word order is one of the reliable cues to sentence meaning. However, (2002)). The present studies address this issue by investigating 2;6- languages like Korean, Turkish, and Japanese have relatively flexible year-olds' knowledge of transitive/intransitive verb alternations, word order and typically have case markers which indicate the using a novel verb learning task. This work provides the first direct grammatical roles of noun phrases. The current study examined whether comparison of 2;6-year-olds' ability to produce creative Korean 3-year-olds can use case markers when understanding sentences. generalizations with the two types of English intransitive verb Children heard either canonical word order (Subject-Object-Verb) (patient intransitives - 'Kim drops the box'/'The box drops' - and sentences or noncanonical word order (Object-Subject-Verb) sentences, agent intransitives - 'Kim paints the box'/'Kim paints'). Children while watching side-by-side videos about a bunny and a bear. In one showed knowledge of both these alternations: the majority made video, the bunny acted on the bear; in the other, the characters' roles creative generalizations with the novel verbs. Furthermore, some were reversed. The children were asked to choose which video matched children generated semantically appropriate agent intransitives even the sentence they heard. The children correctly understood both when discourse pressure favored patient intransitives, indicating canonical and noncanonical word order sentences relying on case that they had stronger command of the transitive/agent intransitive markers. These results suggest that at least by 3 years of age, Korean alternation. This bias was in line with frequency distributions in children have some knowledge of case marking system in Korean. adults' speech to children. These findings provide additional support for children's early access to syntactic knowledge.
Mahalingappa This study examines the acquisition of split-ergativity in Kurmanji Kurdish where there is a weakening of the ergative construction in the Lidz language. Data include spontaneous speech samples and How do infants use a Noun Phrase's syntactic context to make experimental data from children (2;0-4;6) and caretakers. Data from inferences about its thematic relation? In 2 experiments comparing caretakers show a limited use of the ergative-absolutive construction direct objects and prepositional objects, we show that syntax in the past tense, as well as ergative case-marking used more often on contributes differently for 16-month olds than for 19-month olds. The agents than patients and pronouns than nouns. These data suggest younger children rely more on the syntactic context than on the lexical highly variable case-marking as input for children. However, children representation of the verb, while the older children show the opposite used similar case-marking patterns to adults; there was no significant pattern. Because sensitivity to the structural context precedes difference in use between adults and children and children of sensitivity to lexical properties of the verb, these data argue against different ages. Thus it is unclear whether children are acquiring a theories in which the syntactic expression of thematic structure is split-ergative system or if they are using patterns seen in the adult derived by abstracting over knowledge of particular verbs. community; however, data suggest that when children are faced with variability present in their language input, they ultimately conform to the variability modeled by the adult community.
Novogrodsky Children with SLI show deficits in the comprehension and production Stringer of sentences derived by syntactic movement. This study tested the In an extension of research on universals in PP structure, this paper comprehension and production of 20 school-age Hebrew-speaking examines four types of modifying elements within spatial PP, and children with syntactic-SLI, and 20 children with typical language. identifies a fixed structural hierarchy, e.g. The girl ran [MEASURE The participants with SLI showed a considerable difficulty in the 20 yards [DEG {right/straight} [FLOW {on/back} [TRAJECT comprehension of (reversible) sentences derived by Wh- movement, {through/over} [PP into the room]]]]. This ordering in English finds along with good comprehension of sentences without Wh-movement; resonance crosslinguistically: when such modifiers appear, they are sentences in which an argument crossed another argument were always in the same order, and stack to the left irrespective of head- impaired (object relatives, object questions, topicalization), whereas complement directionality. In order to explore whether L2 learners sentences without such crossing were comprehended better (subject- have prior knowledge of this hierarchy, a preference task and a relatives, subject-questions); sentences in which both the moved grammaticality judgment task were designed in which an original, element and the crossed argument were referential (object which- computer-animated narrative was used to contextualize PPs and questions) were harder than sentences in which the moved element their modifiers. In advance of the main experimentation, two pilot was non- referential (object who-questions). The patterns of studies revealed that the hierarchy is not in evidence in the initial production and repetition are consistent with these patterns. We thus stages of L2A, and preliminary results suggest that accuracy on the suggest that the syntactic deficit in syntactic-SLI relates to the hierarchy is dependent on gradual acquisition of the lexicon. assignment of thematic roles to an element that underwent Wh- movement over another similar argument. Terzi Valois We studied (complex) locative Ps of 69 Greek speaking children, aged The goal of this paper is to show how child language data can 2-6, 8 age groups. Paradoxically, production showed that the three help discriminate between competing analyses of the same older groups use both epano apo 'above' and epano se 'on' correctly, morpho-syntactic phenomenon. Examination of two spoken while in the comprehension task they performed correctly on the French corpora, one cross-sectional (CS: 11 children aged former but at chance on the latter. We explain this mismatch as 1;8-2;12), and one longitudinal (L: Pauline, aged 1;2:20 – 2;6:13, follows: epano is lexical, modifying Place (Terzi 2005), apo is semi- in CHILDES,) leads to the conclusion that agreement is not at the lexical (van Riemsdijk 1990, 1998), carrying the meaning of Path/ source of noun ellipsis in French, e.g. Je veux le bleu 'I want the source (or distance), se is functional, checking the Case of the ground blue (one)', (corroborating results from Snyder et al. (2001) for argument (Botwinik-Rotem & Terzi 2008). Acquisition of epano is Spanish, and Ntelitheos and Christodoulou 2005 for Greek), and thus expected to precede acquisition of apo and apo to precede that the presence and nature of the determiner plays a central acquisition of se (confirmed by Alexaki et al. 2007). The correct role in this process. This supports Bouchard's (2002) analysis of production data do not reveal children's (incomplete) knowledge of French and English N-Drop in which atomization of the the spatial on, but their knowledge of Case requirements. denomination of a noun, which is accomplished through the Comprehension shows that children understand better spatial determiner in French, is the key factor licensing N-drop. notions with exclusively lexically conveyed meaning, i.e., above.
Takahashi Semantics: How do infants identify the phrase structure of their target language? Some have argued that semantic and/or prosodic Aspect correlates to phrase structure are necessary. Recent work has argued that surface distributional statistics may also contribute to syntactic acquisition. Thompson & Newport (2007) showed that adult participants can exploit transitional probabilities to acquire an artificial phrase structure grammar. This artificial grammar, however, contained phrases with no internal structure. Given that internally structured phrases are a hallmark of natural language syntax, these findings leave the issue of whether learners can detect statistical cues to internally structured phrases unresolved. Using the Headturn Preference Procedure, we find that 18-month-old infants can learn an artificial grammar with internally structured phrases based on the statistical distributions alone. These results add to a growing body of evidence that statistical distributions provide a reliable and usable cue to discovering syntactic structure.
Gabriele Wagner Research has shown that in L2 English, progressive marking begins To what extent is the infant's mind prepared to acquire the with activities and only later extends to accomplishments or semantics of natural language? This study examines infant achievements. Shirai and Andersen (1995) proposed that this pattern representations of a key semantic concept, telicity. Telicity is a emerges because activities represent the prototype for the progressive, feature of predicates: Telic predicates describe events with inherent but very little research has focused on how learners extend beyond the endings (I ran to the store); Atelic predicates describe events prototype. Achievements such as 'die' are interesting in that they without such endings (I ran). However, events in the world have no interact differently with markers of progressive aspect across languages. aspectual value: speakers predicate choices reflect their construal of While achievements are compatible with the progressive in English and events. Can 12-month-olds adopt Telic and Atelic construals of Korean, this combination is ruled out in Chinese. We investigated the directed motion events? A habituation looking-time method was interpretation of progressive achievements in L2 English by advanced used, following Woodward (1998) and Lakusta et al. (2007). Three Chinese and Korean learners to examine whether there is evidence of types of manners were tested (hopping, gliding, scooting) that varied universal difficulty, as the prototype hypothesis suggests, or whether in how goal-directed they appeared to adults. The results suggest similarity between the L1/L2 can facilitate acquisition, as a transfer that infants can adopt a Telic construal of directed motion events, model would predict. Results show better performance for the Korean through some but not all manners of motion; their goal-directedness learners, supporting transfer, but native-like attainment is not reached was largely predictive of infants' taking Telic construals. in either group. Hacohen Semantics: I report experimental data on the acquisition of compositional-telicity in 23 typically-developing Hebrew-speaking children (aged 7;9-14;6) and 9 Other adults. Using a TVJ, participants were presented with video-clips of incomplete events and had to judge whether the accompanying (a)telic predicate matched the event. Results show that adults rarely accept telic predicates as true descriptions of incomplete events (13%), while freely accepting atelic predicates as descriptions of the same events (89%). Data also reveal that both NP-type and definiteness play a crucial role in the derivation of Hebrew compositional telicity. Surprisingly, children were non-adultlike even at 14;6. Testing definiteness and mass/count independently, I find adultlike knowledge of definiteness, but not of the mass/count distinction. Based on these data, I argue that knowledge of definiteness and the mass/count distinction are both necessary conditions for the acquisition of compositional-telicity. Furthermore, the order of acquisition is 1) definiteness; 2) mass/count; 3) telicity.
Ross Hunter This study investigates how children (age four-five) interpret We investigate children's learning of determiner meanings, a topic at presuppositional information encoded in prosodic contrastive the intersection of two domains of research: children's word learning stress and, in particular, whether these children can calculate (focused on open-class items such as nouns and verbs) and children's implicatures generated by contrastive stress. During a story-telling understanding of quantifiers (largely focused on scopal interactions). activity, children had to interpret contrastive stress cues in order to Natural language determiners are all conservative, despite other provide felicitous answers to questions about the story. Through logically possible determiner meanings. The arbitrary nature of this examining both children's success rates and their error types, we generalization suggests that it may derive from constraints on found that as children learn to interpret contrastive stress, they determiner learning. We present an experiment comparing the recognize its presuppositional character before learning how to learnability of two novel determiners: one conservative, and one correctly identify a presupposition. Interference from syntax was nonconservative. We taught each child one of these two novel one factor confounding children's ability to identify the correct determiners, using scenarios illustrated on cards. Children succeeded in presupposition. Nonetheless, children generally gave answers that learning the novel conservative determiner, but not the novel kept some information constant, suggesting that they understood nonconservative determiner, suggesting that the generalization that some element to be presupposed. Hence we propose that children natural language determiners are conservative may derive from a learn the general meaning of the contrastive stress prosodic tune restricted set of hypotheses children consider for determiner meanings. before they learn how to correctly apply it to the semantico- pragmatic information in a particular discourse.
Syrett We ask whether children are aware of the semantic constraints of certain syntactic environments (1-2), and recruit this knowledge Semantics: when extending word meaning to a quantity or predicate interpretation. Scope 1. Partitive: [two / *happy] of the children 2. a. exactly [two / *happy] children b. very [*two / happy] children 30 children (M=3;9) heard subsets of two red objects described as 'pim of the objects.' At test, they were shown two blue and three red objects and asked to locate (a) 'pim of the objects' or (b) 'pim objects.' Children who heard (a) assigned a quantity interpretation (77%), in contrast to (b) (38%). Robust performance on a What's- on-this-Card task coupled with other children's (M=3;9; n=12) ability to assign a predicate interpretation with 'very' (90%) demonstrates children know about these constraints, but mapping to number word meaning is fragile. Results from adults and from older children with 'exactly' provide further support. !"#"$"%&"'( ! "#$%&'!()'!*%'!+)'!,!"-&.$/'!0)!1223)!04$#&%&5'!6$#7%&5!$&8!9-8:;$#%<-&'!0)! ! ! ! ! ! ! ! ?#$@%4#'!,!A)!B$/&4#!C48)D'!!"#$%"&'()"$*+,*$",'",&"*%#+&"$$(,-)!E%;;78$;4'!*FG!0$.#4&H4! ! ! ! ! ! ! I#;J$:9)! ! ! K$<4#7-&'!A)!L)'!0%M4#74854'!()!K)'!,!B-.;$&8!")!NOOP)!"Q%;8#4&R7!H-96#4Q4&7%-&!->! ! ! ! ! ! ! $9J%5:-:7!74&<4&H47!.% Morphosyntax/ Root infinitives Lee, O. This study presents experimental evidence against the claim that the Korean disjunction -(i)na under negation receives only the conjunctive interpretation (Szabolcsi, 2002). Previous studies have found that English children, English adults, and Japanese children prefer the conjunctive interpretation of the disjunctive operator under negation (Goro & Akiba, 2004). The present study finds that Korean children readily accept the disjunctive interpretation, provided appropriate contextual support for the disjunctive interpretation is available. ! ! ! "#$%&!'(!)%*+*,&-!./0-1+,*02!"#345!6/#77#,*+#%!#2-!126/#77#,*+#%!8#/,*+*8%&!#6/&&7&2,! ! 1.%23& 4%..#),&'5.##$#/,& 6#"'2+,&7%& ! Moscati YVJ&#/!0%-3!BYbZ!j!Yb''E! Zl!BZmeE! 'ZZl!BemeE! ! ^k'Z! ! Past-participle agreement is obligatory in Italian when the direct UVJ&#/!0%-3!BUbY!j!Ub''E! WY('l!BUYmUQE! 'R(QlBRmUQE! ! object is realized by means of a clitic pronoun: in this paper we will ^kYe! ! present new empirical data showing that the acquisition of object-Past eVJ&#/!0%-3!Beb'!j!eb'ZE! WZl!BSZmRhE! YZl!B'hmRhE! ! Participle agreement in Italian is delayed w.r.t. other forms of ^kY'! ! "0,#%!! RWl!BQYm''WE! YYl!BYSm''WE! ! agreement (i.e. determiner-noun, subject-verb). Data were collected ^khh! ! using an elicited production task, testing 55 monolingual Italian- ! speaking children between age 2;0-4;10. The results show that Italian ! children systematically allow the default past participial form even in .*+,1/&35!,9&!%*,,%&!8*6!:#39&3!,9&!+0:! ! those cases where this form is not allowed in Standard Italian. We );8&/*7&2,&/5!<=03>!!9#!!?#,,0!!*%!!!!!7#*#%*20!!!#%%#!!!!!!71++#@A!!!!!!!!!BC8&+*?*+!D1&3,*02E! consider this result as evidence for the presence of a developmental !!!!!!!!!!!!!!!!!!!!!!!!!F9#,!9#3!-02&!,9&!!!%*,,%&!8*6!!!,0!,9&!!!+0:!"#$@! stage where past-participle agreement is optional. This phenomenon is ! =9*%-5!!!!!!!!!!!!!G%>!9#!!!!!!!%#H#,%&& & & & & &&&B+%*,*+I-&?#1%,!70/890%06JE! reminiscent of other romance languages, as French, and it can be ! !!!!!!!!!!!!+%!9#3!!!!!!:9#39&-!$'()!!!! explained by assuming that initially children adopt a non local ! mechanism to check agreement (Guasti & Rizzi, 2002) instead of the ! adult spec-head local configuration (Sportiche, 1996). *#+#),#-&.#"#.#/)#(0& K2,*21++*L! M(! N! O(! P*%%&/(! B'QRSE(! Perovic Pratt This study compares the knowledge of finiteness in 10 children This study used a comprehension task, as opposed to naturalistic with Autism Spectrum Disorders (ASD) (mean CA=11;2; non- data, to better understand child Spanish speakers' interpretations of verbal IQ=73.9; vocabulary SS=69.4) and eight children with verbal tense morphology. Twenty-three Spanish-speaking children Williams syndrome (WS), matched on CA and non-verbal IQ. (mean age = 3;10) from Mexico City were introduced to a Winnie- Using two elicitation probes from the TEGI (Rice & Wexler, 2001), the-Pooh puppet, who presented each child with a series of three we found considerable difficulties with both present and past tense, pictures of himself. A modal, present progressive, and perfective of regular and irregular verbs, in children with ASD only. These representation of each action was included. Children were asked to patterns confirm the reports of impaired knowledge of finiteness in point to the picture that corresponded to the tense and aspect of the ASD and its spared knowledge in WS, suggesting that ASD is a risk spoken stimuli. Children were scored on their ability to select the factor for impairment in finiteness, but this impairment does not corresponding picture. Upon passing training and filler tasks, each seem to be associated with the level of non-verbal cognitive child completed 6 experimental trials. Children were largely able to impairment. The results will be discussed in the light of deficiencies successfully interpret tense and aspect information as it was on more complex aspects of grammar (passives and raising) in WS conveyed in the adult-like cues. As such, we can assume that and ASD, which suggest that grammar in WS may be subject to children at this age have a near adult-like understanding of tense fractionation not previously reported in the literature. and aspect as they are marked in Spanish morphosyntax. Roeper Sundara How does the child manage the interface between language, context and This study investigated position effects in children's cognition in order to advance his (language particular) grammar? We comprehension and production of 3rd person singular -s. In will provide a panorama view of acquisition through snapshots of how Experiments 1 and 2, we tested 27-month-olds' listening the child moves from lexical to productive rules using theoretical preference for grammatical and ungrammatical 3-word sentences principles like asymmetric merge, recursion, and movement. while watching a cartoon of an action, thereby providing a Pragmatics, implicatures, and the emergence of quantification all play a referential context. Children were better at detecting the absence role in the child's world, a world full of half sentences (ellipsis), attitudes of 3rd person singular -s in sentence-final position compared to (implicatures) and hidden propositions (entailments). Strong innate sentence-medial position. In Experiment 3, the production of assumptions about the interfaces between grammar and mind allow the 3rd-person singular -s by the same children in Experiment 2 was child to convert what seems like excessive, confusing information into examined using an elicited imitation task. Children who listened sharp steps along the acquisition path. These questions lead us to the longer to ungrammatical sentences in Experiment 2 (medial recognition of new forms of language disorders (revealed through the position) had higher overall production scores, suggesting that DELV test (Seymour, Roeper, and deVilliers (2005)) and they can help their lexical/grammatical representations were more robust. The us articulate the challenge of preserving a sense of children's dignity in results will be discussed in the context of (a) input frequency (b) the exploratory phase of scientific inquiry which, perhaps unavoidably, the perception of grammatical morphemes in non-referential begins with simplistic versions of the human mind (as discussed in my tasks, (c) comprehension results for plurals, and (d) implications book "The Prism of Grammar" (2007)). for the development of grammatical morphology more generally. L2A Rothman This study examines the adult L2 acquisition of properties related to the Spanish DP by two sets of intermediate and advanced groups compared against a native Spanish control: L1 German and Italian learners of L2 Spanish. In doing so, we test the predictions of Representational Deficit accounts (RDA) (e.g. Hawkins & Chan 1997; Hawkins & Hattori 2006) against Full Accessibility accounts (FAA) (e.g. Schwartz & Sprouse 1996; White 2003). Experimentally, we investigate L2 knowledge of properties of noun raising as well as related syntax-semantics properties, namely, the available readings for adjectives depending on their position with respect to the head noun. Our results are consistent with recent studies of English learners of L2 French (Anderson 2001, 2007) and L2 Spanish (Judy et al. 2008), all of which support Full Access to UG in adulthood. White, L. In the English existential there-insertion construction, indefinite/ Duffield weak expressions are permitted, while definite/strong expressions In this talk, we report on continuing research investigating the are excluded (e.g. 'There is a/*the fly in my soup'). Languages processing of English VP-ellipsis and VP-anaphora constructions by lacking articles nevertheless show some definiteness restrictions. native English and Dutch L2 speakers. Using a more time-sensitive Russian (no articles) and Turkish (no definite article) prohibit method (eye-tracking during reading) three main findings emerge. strong determiners in positive existentials, while showing no First, processing both VP-ellipsis and VP-anaphora is easier following definite/indefinite contrast in negative existentials. We report on syntactically parallel versus non-parallel antecedents, a result that an experiment involving Russian- and Turkish-speaking English supports the semantic account of ellipsis (Merchant, 2001) over the L2ers. We hypothesize that learners will initially assume that traditional structural account proposed by Hankamer & Sag (1976), English is like the L1, permitting definite expressions in negative Sag & Hankamer (1984). Second, differences in the timing of the existentials, while more proficient L2ers will arrive at the effects suggest that the process linking the elided clause to its appropriate L2 restrictions. The task involves contextualized antecedent in VP-ellipsis is different from that involved in VP- acceptability judgments. Results show that advanced L2ers from anaphora resolution. Finally, the results of the L2 learners suggest that both L1s respond like native speakers, accepting only indefinite/ competence differences rather than performance factors underlie the weak DPs in both affirmative and negative existentials. Lower observed differences between the judgments of Dutch learners versus proficiency L2ers accept certain strong DPs in negative existentials, native English speakers observed in earlier studies. suggesting transfer. The results suggest that subtle definiteness restrictions are acquirable, regardless of L1/L2 differences. Wilson Kras The Interface Hypothesis (Sorace and Filiaci, 2006) proposes that This paper investigates whether the lexicon-syntax interface, an while L2 learners are able to acquire narrow syntactic properties of interface between two domains internal to the language faculty, can the L2, phenomena at the interface with other cognitive domains be fully acquired in the L2. The phenomenon under scrutiny is are present difficulties. We discuss two visual world experiments auxiliary selection with intransitive verbs in Italian, a syntactic which investigate L2 processing at the syntax-discourse interface, phenomenon determined by aspectual/thematic factors, i.e. telicity by examining the antecedent preferences for two different and agentivity of the verb/predicate (Sorace 2000). In the paper, we pronominal forms in L2 German: personal pronouns and anaphoric report the findings from an experimental study in which adult near- demonstrative pronouns. These anaphors are subject to different native speakers of Italian whose native language, Croatian, does not constraints on antecedent preference in native German speakers. have auxiliary selection, and native Italian speakers performed a self- We examine three constraints on anaphor resolution: grammatical paced and a speeded version of an acceptability judgement task. The role, topicality and thematic role of potential antecedents for these results of the analyses of three types of data (relative acceptability anaphors in L2 German, and show that L2 learners are not able to judgements, absolute acceptability judgements, response times) use these constraints in online processing in the same way as native reveal that near-natives have acquired both the syntactic and the German speakers. We conclude that L2 learners show a difficulty lexical-semantic aspect of auxiliary selection with intransitive verbs. integrating different sources of information in real-time Such findings provide support for a hypothesis predicting complete comprehension, supporting the Interface Hypothesis. acquisition of the lexicon-syntax interface in the L2. Lozano Processing The aim of this presentation is to show that both English learners of Romance and Romance learners of English are sensitive to L2 discourse status (focus) in subject inversion constructions (VS), but they show residual yet persistent problems when encoding information status syntactically. Both our experimental and corpus data support the claim that deficits at the syntax discourse interface are syntactic in nature. Additionally, an important conclusion of our work is that (i) unaccusativity is a necessary but not a sufficient condition for the acceptability and/or the production of Verb Subject structures in non-native grammars and that (ii) properties at both the syntax-discourse and syntax-phonology interfaces, which have been shown to be relevant for a variety of word order phenomena in native English, also play a crucial role in L2 constituent ordering. Skarabela Gagliardi This study compares priming effects in two types of possessives in The study of filler-gap dependencies has held a central position in English: the prenominal s-possessive ('the doctor's mother') and the syntactic theory and models of sentence processing, due to their postnominal of-possessive ('the mother of the doctor'). Participants unbounded character and the demands this unboundedness places on were 30 monolingual English-speaking 4-year-olds. In the test on-line sentence understanding. While much work has examined the phase, half of the children heard a prime containing an s-possessive acquisition of wh-questions and relativization, very little work has asked (the s-group), and the other half heard a prime containing an of- whether the patterns of acquisition for these dependencies are parallel, as possessive (the of-group). Significantly more s-possessives were research in syntactic theory would predict. Earlier work has found that produced by the s-group (65%) than the of-group (4%) and subject wh-questions are understood by both 15- and 20-month-old significantly more of-possessives were produced by the of-group infants, and that object wh-questions are understood only by 20-month- (53%) than the s-group (23%). The priming effect was stronger for olds. We demonstrate, however, that both 15- and 20-month old infants the of-possessives than the s-possessives. Furthermore, the effect can understand wh-movement and relativization involving both subjects persisted into the post-test in the of-group, but not in the s-group. and objects. Moreover, we show that differences in reaction times We will discuss these findings in light of current debates about the between the two ages are found in both wh-questions and relativization, nature of children's syntactic representations and the role of implicating a common processing mechanism for these dependencies. frequency versus rules in language development. Reichle This study examined the event-related potentials (ERPs) indexed with the processing of syntactic focus structure anomalies in L1 and L2 speakers of French and found that they are, for L1 subjects, similar to those seen for prosodic anomalies. Subjects were visually presented with questions in French, followed by responses containing felicitous or infelicitous focus structures. N400 and P600 effects were seen in L1 subjects for the infelicitous condition, which agrees with results for prosodic mismatch and focus structure revision from previous studies (Hruska & Alter, 2004). In low-proficiency L2 subjects, the same stimuli elicited a widely distributed positivity from 400-800 ms, which was interpreted as a P3b component similar to the one seen by Magne et al. (2005). In high-proficiency L2 subjects there was marginal evidence of an N400 followed by a P600, suggesting that an increase in proficiency leads to processing signatures more similar to those of L1 subjects.