Scarica Clitic Left Dislocation in Italian Child Language - Prof. Belletti e più Appunti in PDF di Morfologia e Sintassi solo su Docsity!
HO4 Children’s production of CLLD
Marking topics in the Left periphery: a-Topics Evidence from language development and crosslinguistic comparison with remarks on passive and properties of subjects 09.12. a-Topics in children’s production Results from acquisition presented in Belletti & Manetti (2018/2019) from elicited production have indicated that young Italian speaking children make a wide use of a-marking of a left dislocated direct object in the Spec/TopP position in the Left Periphery of the clause: the a-Topic They tend to produce ClLD structures as in (1)a, instead of the standard (1)b: (1) a. Al coccodrillo, l’elefante lo lava to the crocodrile the elephant him.Cl washes ‘The crocodrile, the elephant is washing him.’ b. Il coccodrillo, l’elefante lo lava the crocodrile the elephant him.Cl washes ‘The crocodrile, the elephant is washing him.’ a-Topics a-marking of the object Topic → a-Topic a DPO DPS Cl V (possibly, a multiple constructions) Other order possible: DPS a DPO ClV (a multiple constructions) (the different order does not appear to play a role, hence collapsed in the results, unless otherwise specified) The position in the left periphery becomes available when children are able to have access to this position (which is also a question on children’s discourse ability). Clearly, in CLLD, the preposed object il leone is a topic given by the discourse and in order to use this kind of structure we need to recognize the topic relation. At age of 4, children are able to master the position of topic in the left periphery, so they seem to have access to the role and use of clitics and to the functioning of the left periphery. DO > PDP - DOM (or- Spanish, et- Hebrew, pe-Romanian) The a-marking seems to be a property of the left periphery and has nothing to do with the topic construction.
Outline and general questions How is CLLD elicited? The acquisition of the Left periphery: One of the elicitation conditions (two topic condition, see following slides) leads to overt expression of the left dislocated topic, which in turns gives us a clear indication on the acquisition of the left periphery by young developing children. The same elicitation design leads to production of passive: difference adults vs children, both in the selection of CLLD vs passive and in the type of passive selected (the latter difference also found in the production of passive object relatives/PORs, cfr. later ho). Adults venire passive; children si-causative passive (only) Focus on the expression of the object topic as a-Topic and crosslinguistic considerations a-Topics and featural Relativized Minimality/fRM, in production, then confirmed by results from comprehension. Il leone, l’elefante lo bagna. → Here we have an overt expression of the object topic: CLLD (2 topic- contrastive topic) L’elefante lo bagna. → pronoun sentences (1 topic) In the pronoun the topic is not pronounced overtly while in CLLD structures the topic is overtly expressed. Way to elicit CLLD (and passive) – Experiment 1 Belletti & Manetti (2018) 1 TOPIC PATIENT S-CL-V (covert topic) → pronoun structure Q: What happens to my friend, the cat?? (1) La mucca la lecca e il riccio l’accarezza. ‘The cow her.CL licks and the hedgehog her.CL caresses’ Here, it is natural to answer without overtly pronounce the preposed subject. 2 TOPIC PATIENTS (contrastive topic) → ClLD O-S-CL-V (overt topic) Q: What happens to my friends, the dog and the bear? (2) Il cane il gatto lo bagna, e l’orso il coniglio lo veste ‘The dog the cat him.CL washes, the bear the rabbit him.CL dresses’ Here, it is more natural to answer without overtly pronounce the preposed subject.
Children’s and adult’s answers Q: Che cosa succede al mio amico, il re? ‘What happens to the king?’ CLLD (with or without overt the topic) ((a) Il re), la mucca lo lecca (The king), the cow him.CL licks Most typical answer by: CHILDREN
PASSIVE
(Il re) viene/è leccato dalla mucca (The king) comes/is licked by the cow Most typical answer by: ADULTS Structures expressing a topic patient vary depending on the experimental setting and the informational context (see Volpato et al. 2015; Manetti 2013; Pivi & Del Puppo 2013). Here example from the one topic condition With respect to the types of overt topics in children’s ClLD: results overlapping SO and OS interpretation, therefore we deduct that there is no significant difference between the two. Zooming on the types of overt Topics in children’s CLLD CLLDs with a - Topics vs. simple O/Topics in DP^1 DP^2 CL-V (collapsing SO and OS interpretations; orders: (S) pre-posed O (S) V): Children produced 88% a-Topics (n= 44) vs. 12% simple O/Topics (n= 6) (with S overt lexical noun phrase)
In OS sentences, the subject is either in its initial position or has been moved to the left periphery but we don’t know whether this is a multiple Topic structure. On the other hand, in front of So sentences we are certainly in a multiple Topic construction. The contrast is between partial and complete crossing of the subject: SO sentences display a partial crossing of the subject and might have some advantage because of the partial crossing while OS sentences display a complete crossing of the subject and this might be the cause of their being more difficult to be computed or understood. Children’s productions of a-Topics When the subject is lexical and preverbal, in 88% of children’s ClLDs the left dislocated object is realized as an a-Topic and not as a Simple Topic. In the remaining 12% of the cases the object was a Simple Topic The questions that felicitously elicited ClLD structures had 2 topics: children were required to overtly pronounce both (one of the two is the subject and the other is the object) in order to be informative, and they were. Two raised by the results A number of new questions are raised by the results on CLLD and a- Topics. The big general question is: Why? Let us split it in different sub-questions, and answer them in turn:
- Do the children tested speak a Differential Object Marking/DOM language?
- If not, how can we characterize their use of a marker of this type? I. Is this marker frequent with topics in standard Italian?
a-Marking in the left periphery: Comparative considerations The fact that a-marking of direct objects is a left peripheral phenomenon characteristically affecting left dislocated direct object topics is not an isolated fact concerning Italian speaking children during their development. Rather, children’s behaviour is robustly attested both cross-linguistically and diachronically. As it is often the case in acquisition, during their development children may typically adopt grammatical options that are manifested in the grammar of some existing language. 10.12. a-Marking of object Top in the LP: a limited option in adult (standard, non southern) Italian
- With Object Experiencer of psych-verbs (worry class): (5) (?)A Gianni, questi argomenti non l’hanno convinto to Gianni, these arguments him-CL have not convinced (6) A Gianni, la gente non lo conosce to Gianni, people him-CL do not know Cfr.: (5’) Questi argomenti non hanno convinto (a) Gianni these arguments have not convinced Gianni (6’) La gente non conosce (*a) Gianni people do not know Gianni (Belletti & Rizzi 1988, fn. 27, example due to P.Benincà) Verbs like preoccupare allow for the appearance fo a with preposed objects. Questo fatto preoccupa Gianni → interesting -assignment with psych-verbs Theme (S) Experiencer (O) By left dislocating the experiencer, this can be introduced by preposition a. Remember that if we are dealing with a- marked objects that are not located in the LP, we are dealing with a DOM language. In 5 ’ and 6’ there is no difference between psych-verbs and non psych-verbs in the fact that a - marking can only affect the object in the LP (and certain objects better than others, such as objects of psych-verbs). This stands good in Sardinian people because Sardinian dialect is a DOM language. Right dislocation would sound better, i.e. “la gente non la conosce, a Gianni”
- With pronominal objects: 1 st and 2nd person (singular) (left) dislocated pronouns require preposition a in Top position. (cfr. Renzi 1988) (7) A me/?Me non mi si inganna to me/me one does not me-CL cheat (Berretta 1989) These preposed objects are all definite (pronouns are definite) and that is because topics are linked to a given information. Sometimes, also non-experiencer pronominal objects and also of 3 rd^ person pronominal objects require the preposition a in topic position: (8) a. A te /?te ti licenziano di sicuro → non-psych verb with 2 nd^ p.s. pronominal object to you/you they you-CL fire for sure b ?A lui /✓lui lo rispettano tutti to him/him they him-CL respect all c A noi sul lavoro non ci assume più nessuno → non psych-verb, 1 st^ p.p. pron. object to us on work nobody us-CL hire anymore The nature of a-Topics: A comparative detour in Spanish (and Sicilian) Leonetti (2004) As argued in Leonetti (2004) there is a very tight relation between Topicality and a-marking in the overall Spanish DOM phenomenology, which clearly emerges in CLLD: A left dislocated direct object is obligatorily a-marked Sicilian has a similar behavior according to the description in Iemmolo (2010). Specifically a-marking is systematic on left dislocated objects. a-Marking is obligatory with topicalized direct objects in ClLD in Spanish, also when the verb allows for both a-marked and un-marked direct objects in simple SVO sentences in which the direct object is not topicalized. Consider the following examples crucially relevant in this respect: a-Topics in Spanish Leonetti (2004; cfr. also Laca 1987) Simple declarative: (9) a. Ya conocía (a) muchos estudiantes already (I) knew many students b. Habían incluido (a) dos catedráticos en la lista (they) had included two professors in the list Indefinte objects: optimality is a - marked in SVO languages
a-Topics in Balearic Catalan DP-lex in object position DP-lex in CLLD topic position The role of affectedness in Spanish DOM A somehow brings some interpretation with it, which has to do with affectedness in Spanish: the DO seems to be directly affected by the event described by the verb. Physical Psychological Torrego 1999, Heusinger 2008 Indefinite objects are optionally a- marked in object position usually, but not when they are physically affected by the action of the verb (not only we have to look at the nature of the object, but also to the type of action described by the verb). The same applies in case of psychological affectedness (comparison between sentences b.)
(Psychologically) affected object (experiencer) and empathy toward it Hypothesis:
- a-marking of Topics is the manifestation of a property located in the Left Periphery
- a-marking adds some further feature to the interpretation of the topic. Assume, also based on the crosslinguistic evidence reviewed, that a-Topic is not just a topic, with the specificity and givenness that the topic interpretation carries along, but it is an affected topic (linked to the benefactive interpretation induced by preposition a. (For time reasons, no detailed discussion now of the different status of English ‘to’ and the consequences for a-Topics and causatives, on which see also Kayne 2004, Belletti 2019). A - marking is indeed a property of the LP regarding topics but it also adds a further feature to the interpretation of the topic: it contributes to the interpretation of the topic as affected topic (liked to a benefactive interpretation in Italian, which is also introduced by a preposition a ). Benefactive (additional dative) argument of the transitive verb: Detour: On the interpretation induced by a Preposition a is the same preposition expressing the benefactive (additional dative) argument of transitive verbs. Ho comprato un vestito ai bambini // Farò una festa ai bambini (I) have bought a dress to the kids // (I) will do a party to the kids The benefactive interpretation expresses a type of affectedness similar to the one assumed here for the a-marker of topics (and indeed the preposition is the same). Cfr. Kayne (2004) on the peculiar benefactive interpretation also induced by à in French. The same preposition is also found in fare a/faire à causatives. And the expression of an affected interpretation of the causee has often been recognized in this type of causatives, since Guasti’s (1993) analysis. Cfr. Manzini & Franco (2016) on the the role of preposition a expressing a kind of possession relation; Anderson (1987), Kayne (1993), Guasti and Moro (2001) and more recently, Collins (2018), Wood and Zanuttini (2018) for related proposals.
(Psychologically) affected object (experiencer) and empathy toward it The a-marking may also express an empathic point of view: a kind of speaker’s perspective of involvement (empathy in the sense of Kuno 1987) so that the speaker (psychologically/ emotionally) participates in the event that affects the topic argument (and through empathy he/she may also feel affected by it). Possibly sharing similarity with a peripheral Sentence Final Particle/SFP like ne in Japanese that signals empathic involvement shared by speaker and hearer, according to Endo’s (2018) recent discussion. So, empathy is something that happens in languages. From here, we can assume that there is an affectedness feature, [+a] feature, that can be associated with the left peripheral topic head in the way described. The experimental condition favored the relevant (affected) interpretation of the Topic. The assumed analysis a-Topic attracted into the Left periphery With a further moved/cliticized into a higher head, thus re-establishing the pre-positional order. Cfr. Kayne 2004, Prepositions as Probes. We can assume that this preposition is cliticisized to the left-dislocated object (similar particle) Kayne ( 2004 ) The left-dislocated object is attracted to the specifier- TopP position
We can assume that a diachronic path from LP a - topics to DOM exists, and that perhaps Italian is undergoing some changes in this direction: LP a - Topic → low vP-peripheral a - Topic → small v 1 st^ step 2 nd^ step 3 rd^ step From Left peripheral a-Topics to DOM A possible path is also suggested by the crosslinguistic diachronic evidence reviewed:
- Left peripheral a-Topic >> Low vP-peripheral a-Topic >> small v Belletti ( 2018 ) Intermediate step: if the [+ a ] feature is not associated to the high topic, but to the low one, then the result will be clause internal (because of subsequent verb raising), but the interpretation would still be linked to a topic “Grammaticalization”: this option is generalized, all objects become differentially marked. 1 st^ step 2 nd^ step 3 rd^ step
a-Marking of the topic as a way to modulate intervention Use of a-Topics can be analyzed in terms of the grammatical principle featural Relativized Minimality/fRM (Rizzi 19 90, 2004, Starke 2001, along the lines proposed in Belletti and Manetti 2018/2019, Belletti 2018a, b, within the approach developed in Friedmann, Belletti & Rizzi (2009) and Belletti et al. (2012): Main insight: a-Topics allowed children to deal with the intervention configuration arising in the ClLD structure by resorting to a featural intersection relation between the two DPs, a relation that children are known to master well at this age. Intervention and lexical restriction Pre-posing a lexical direct object DPO over a lexical DPS instantiates the intervention configuration: The hardest intervention situation whose computation is notoriously difficult for young children is the one where both DPO and DPS are lexically restricted (FBR, 2009 and crosslinguistic literature on A’-dependencies such as Relative Clauses and Wh-questions). if we assume: a-DPo^ DPs^ Cl V +Top +NP +a + NP +u a-Topics and intervention locality More precisely, the realization of the pre-posed object as an a-Topic opens up the possibility of better coping with intervention locality, RM in its featural formulation/fRM by modulating the long distance relation, as pre-posing of lexical DPO over lexical DPS, in either order (DPS DPO or DPODPS), instantiates the intervention situation that young children are unable to master: INCLUSION INTERSECTION
Recall: Relations w.r.t. relevant features between Target and Intervener expressed in set theoretic terms Types of ClLDs and the intervention configuration A case of grammatical creativity against a poor input Children’s extended use of a-Topic thus results in a case of grammatical creativity against an otherwise poor input. It is a rule-governed creativity: Despite poor and limited evidence, children have picked up a grammatical option that has allowed them to cope with an otherwise too complex structure for them to compute, due to intervention. Thus, children’s linguistic behavior is not at random: it takes place within a defined grammatical space. In the case under study, it occurs under the pressure of locality. Further indications in the same direction come from the following:
The crucial problem is the +NP of the two elements so what if S is not +NP and so it is a pronoun? The prediction is that if we have a pronoun (aDP (+NP +Top) + pro… <-> ) = disjunction in the Y position? If there is a pronominal subject, which is possible in Italian, it will likely be a null subject, then we expect that the structure will be ok meaning that the structure will be ok also without a so with a simple topic structure. And indeed, the results show that the realization of a-topic depend on the nature and position of the subject. No intervention with post-verbal and null subject Spanish is a NSL too, it is similar to Italian in the fact that they both allow for null subjects but it is different from Italian because it tolerates more overt pronominal subjects here we see the difference between the two (SP: DOM vs. IT: non-DOM) may live. Picture: Cat washing dog → postverbal subject Picture: boy combing king → null subject Recall that in the experimental conditions the subject was expected to be: o Overt in order for the answer to be completely informative (in postverbal) o Preverbal as the subject was not new info Whence the relatively limited number of these productions whose presence is all the more interesting, since children have shown to be generally discourse appropriate, also in the use of subjects. The locality pressure is the source of the not fully felicitous answers.
The previous experiments tested production, now we test comprehension. The answers to these questions we found very few lexical subjects and overwhelmingly present null subjects but interestingly most subjects were null but plural. Number Mismatch between subject and object (Experiment 2) Number Mismatch condition. ◦ Singular object and plural subjects (the opposite of the previous experiment) ◦ 36 children (from 4 to 6 y.o.) and 20 adults One patient topic vs. Two patient topics (13) Il coniglioO^ i gattiS^ lo toccano The rabbit the cats him.Cl touch (14) La formicaO^ le raneS^ la coprono The ant the frogs her.Cl cover If we create a number mismatch condition: the preposed object is singular and the subject is plural. In comprehension and in extended mismatch condition it turned out that the number of mismatch condition ClLD is not understood. This is interesting because the different roles of these two features are different in the comprehension of relative clauses. Results on the use of lexical vs. null subject in Exp Most salient feature of these results: In Experiment 2 most subjects in ClLD were null In Experiment 2 most subjects in ClLD were null and plural We can compare the results with the ones of the previous experiment discovering that here we have an overwhelming production of null and plural subjects. In experiment 1, in the match condition, we observe a seldom use of null subjects in general (and of the few pros^1 few were prop^1 ) interestingly because the subject children were shown was singular, so what is this plural?
