Hello! My name is Tihana Kraš and in this presentation I will discuss the development of null and overt subject pronouns in multilingual speakers. Some languages, such as Italian, Spanish, Croatian, Polish, Greek, Hebrew and Arabic are null-subject languages. In these languages, in the pre-verbal position of finite clauses, subject pronouns can be both omitted or expressed. If they're omitted, they're called null-subject pronouns, and if they are expressed, they're called overt subject pronouns. Both options are illustrated in Italian on the slide. If after the sentence "Emma ha imparato a nuotare", meaning "Emma has learned how to swim", we say "È felice", meaning "She's happy", we are using a null pronoun at the beginning of the sentence, and if we say "Lei è felice", which also means "She's happy", we are using an overt pronoun. In non-null-subject languages, such as English, French and German, subject pronouns always need to be expressed, so only overt subject pronouns are possible. The two types of pronouns in null-subject languages have different functions in discourse. To explain what they are, I need to introduce the concept of the antecedent. The antecedent is the noun or the noun phrase that a pronoun refers to. Pronouns agree in person, number and gender with their antecedents. Going back to the functions of null and overt subject pronouns in discourse, null pronouns typically refer to topical antecedents (those that are in the focus of attention at the moment of speaking), while overt pronouns typically refer to non-topical antecedents (those that are not in the focus of attention). This can also be Illustrated in the example on the slide. If after the sentence "Emma ha imparato a nuotare", we say "È felice", so we use the null pronoun, we are probably saying that Emma herself is happy because she has learned how to swim. On the other hand, if we say "Lei è felice", so we use the overt pronoun, we are most likely saying that it is someone else who is happy because Emma has learned how to swim. This other referent can be identified from a broader context. When it comes to contexts within a sentence, it has been proposed that Italian subject pronouns tend to refer to different antecedents. More precisely, it has been proposed that the null pronoun tends to refer to an antecedent in the subject position because subjects are the prototypical topics of a sentence, while the overt pronoun tends to refer to an antecedent in a non-subject position. In the example on the slide, this would mean that in the sentence, "Marina ha salutato Nina mentre correva", meaning "Marina greeted Nina while she was running", it is most likely Marina who was running because the subordinate clause contains a null pronoun, and Marina is the subject of the main clause. On the other hand, in the sentence "Marina ha salutato Nina mentre lei correva", it is probably Nina who was running because the subordinate clause contains an overt pronoun, and Nina is the object of the main clause. This proposal has received support in empirical studies. Native speakers of Italian indeed identify the pronoun antecedents in line with it. The same proposal has been shown to apply to Croatian. In the sentences with a null pronoun, such as "Marina je pozdravila Ninu dok je trčala", meaning "Marina greeted Nina while she was running", Croatian native speakers, in most cases, identify Marina as the one who was running, while in the sentences with an overt pronoun, such as "Marina je pozdravila Ninu dok je ona trčala", they predominantly identify Nina as the one who was running. The comprehension and production of null and overt subject pronouns lies at interface of syntax and discourse-pragmatics. Syntax determines whether both null and overt subject pronouns are possible in a language. On the other hand, it is discourse-pragmatics that determines when each of the two forms are used, that is their distribution. In line with this, we can say that the acquisition of subject pronouns contains two components. The first component consists in learning whether null subject pronouns are present or absent in a language, while the second component consists in acquiring the rules that govern the distribution of null and overt subject pronouns. When it comes to the acquisition of subject pronouns in multilingual speakers, it has been shown that the first of these two components is relatively unproblematic, while the second one can sometimes pose difficulties. Let's take a closer look at Italian, a language that has been extensively studied in relation to this topic. Studies on Italian have shown that highly proficient English-Italian and Spanish-Italian bilinguals produce and accept overt subject pronouns referring to inappropriate topical antecedents to a significantly greater extent than monolingual speakers. It can be said that they overgeneralise overt subject pronouns to context in which null subject pronouns are required. On the other hand, they produce and interpret null subject pronouns appropriately. Which bilinguals were tested in these studies? The bilinguals tested include English-Italian and Spanish-Italian simultaneous bilingual children, adult near-native speakers of Italian whose native language was English and adult near-native speakers of English who were undergoing attrition in their native language, Italian. A lot of the evidence that has led to these generalisations has been obtained by a picture selection task originally used by Tsimpli and colleagues in 2004. In this task, participants were presented with a sentence and a set of three pictures. They were asked to select all the pictures that matched the sentence in meaning. The sentence consisted of two clauses: one main and one subordinate. The main clause contained two noun phrases, in the role of the subject and the object, while the subordinate clause contained an overt or a null subject pronoun that matched the two noun phrases in number and gender. The subordinate clause either preceded or followed the main clause. By selecting the pictures, participants were in fact identifying possible antecedents of the subject pronoun used in the subordinate clause. For example, if the sentence was "Il papà saluta il figlio mentre va in bicicletta", meaning "Dad greets the son while he goes on a bicycle", the participants had to decide who was going on a bicycle: Dad, as shown on picture one, the son, is represented by picture two, or another person, not mentioned in the sentence, as picture three depicts. If they chose picture one, they identified the subject of the main clause as the antecedent of the pronoun. If they chose picture two, the object, also called the complement, and if they chose picture three, an extralinguistic referent, not mentioned in the sentence. Sentences such as this one, with a null pronoun, did not pose any difficulties to the bilingual speakers, but those with the overt one did. These difficulties can be seen in the results of Sorace and Filiaci (2006), in which the previously described picture selection task was used to test near-native speakers of Italian whose native language was English, as well as native speakers of Italian. The figure on the left shows the participants' responses to the sentences of the type "Il papà saluta il figlio mentre lui va in bicicletta", meaning "Dad greets the son while he goes on a bicycle". In this sentence, the overt pronoun follows its potential antecedents. In the figure, the lighter bars show the native speakers' responses, and the darker bars, the near-native speakers' responses. We can see that both participant groups preferred the complement as the pronoun antecedent. However, the near-native speakers chose the subject as the pronoun antecedent more often than the native speakers. This difference has proven to be significant in a statistical analysis. The figure on the right shows the participants' responses to the sentences of the type "Mentre lui va in bicicletta, il papà saluta il figlio", meaning "While he goes on a bicycle, Dad greets the son". In this sentence, the overt pronoun precedes its potential antecedents. We can see that the responses of the two groups of participants differed to a greater extent in this figure than in the previous one. The two participant groups preferred a different antecedent for the overt pronoun: the native speakers preferred the extralinguistic referent, while the non-native speakers, preferred the subject. In addition, the near-native speakers chose the subject as the pronoun antecedent significantly more often than the native speakers. Therefore, in both cases, the near-native speakers accepted overt pronouns referring to inappropriate topical antecedents, that is the subject of the main clause, to significantly greater extent than the native speakers. However, not all multilinguals were shown to have problems acquiring the rules that govern the distribution of null and over subject pronouns. In a series of studies on Italian, highly proficient Croatian-Italian bilinguals were shown not to accept more Italian overt subject pronouns referring to inappropriate topical antecedents then the age-matched Italian monolinguals. This applies to adult L2 Learners, child L2 learners and simultaneous bilinguals. What is more, child L2 learners and simultaneous bilinguals were shown to accept fewer Italian overt subject pronouns referring to topical antecedents then the age-matched Italian monolinguals. Importantly, Croatian-Italian bilinguals were shown to interpret Italian null subject pronouns appropriately, similarly to English-Italian and Spanish-Italian bilinguals. The absence of difficulties with the interpretation of overt subject pronouns can be seen in the results of Kraš (2008), in which near-native speakers of Italian whose native language was Croatian were tested together with native speakers of Italian. In this study, a modified version of the task used by Tsimpli and colleagues in 2004 was used. In this version, the participant could choose only one picture for each sentence. Looking at the results for the same type of sentences as shown previously, that is sentences with the overt pronoun, we can see that neither in the sentences in which the pronoun follows its potential antecedents, shown on the left, nor in the sentences in which the pronoun precedes its potential antecedents, shown on the right, the near-native speakers chose the subject as the pronoun antecedent more often than the native speakers. To summarise the findings on Italian, evidence from a number of studies points to the target-like interpretation of overt subject pronouns in Croatian-Italian bilinguals in contrast to the non-targetlike interpretation of overt subject pronouns in English-Italian and Spanish-Italian bilinguals. What could explain these differences? I believe that these differences could be explained by crosslinguistic influence, that is by the influence of one language on another in the mind of a bilingual speaker. Let's compare the four languages involved in the relevant studies. In the table, we can see that Italian and Croatian use identical means to refer to topical and non-topical antecedents, that is null and overt subject pronouns respectively. On the other hand, Spanish uses both null and overt subject pronouns to refer to topical antecedents and only overt subject pronouns to refer to non-topical antecedents. Finally, as a non-null-subject language, English uses overt subject pronouns to refer to both topical and non-topical antecedents. I believe that the fact that English uses exclusively overt subject pronouns and Spanish both null and overt subject pronouns to refer to topical antecedents could influence the way English-iIalian and Spanish-Italian bilinguals use overt subject pronouns in Italian. More precisely, I believe that under the influence of their other language, these bilinguals use overt subject pronouns in Italian to refer not only to non-topical, but also to topical antecedents. On the other hand, Croatian-Italian bilinguals use only null subject pronouns in Italian to refer to topical antecedents because this is what they do in Croatian as well. More evidence coming from studies focusing on other null-subject languages and directly comparing bilinguals speaking different language combinations is needed to see if the bilinguals' difficulties with overt subject pronouns are indeed caused by crosslinguistic influence. To summarise, in null-subject languages, subject pronouns can be null or overt. Null subject pronouns are used to refer to topical antecedents and overt subject pronouns to refer to non-topical antecedents. Acquiring the rules that govern the distribution of null and overt subject pronouns can be difficult for multilinguals. More precisely, some highly proficient bilinguals have been shown to overgeneralise overt subject pronouns to contexts in which null subject pronouns are required. On the other hand, they use null subject pronouns appropriately. The bilinguals' difficulties with the use of overt subject pronouns might be caused by crosslinguistic influence. What teachers should remember from this presentation is that the distribution of null and overt subject pronouns in null-subject languages may pose difficulties to bilingual speakers. The properties of the bilingual's other language seem to play a role in the presence or absence of these difficulties. For this reason, I believe that raising awareness of the way in which null and overt subject pronouns are used in the target language and comparing it to the way in which subject pronouns are used in the bilingual's other language may facilitate acquisition. Thank you for your attention and to the EU for funding the preparation of this presentation. If you would like to learn more about the project within which this presentation was developed, please visit this website.