 Okay, so good afternoon to everybody. Thank you very much for being so punctual. We know it's not easy to leave restaurants and be here on time, so congratulations. And we are then going on to the second lap in our summer school today after a wonderful session, which we would like to thank for this morning. It's an enormous source of inspiration. We are also very happy that it was very interactive, so Albert's words really worked wonderfully. Very, very good interaction. We hope it continues this afternoon. Be sure it will. So it is a pleasure and an honor for our summer school and in the name of the directors and the organizers, I would like to welcome Professor Fred Genesey with us. Professor Fred Genesey has a chair at McGill University in psychology, as you probably very well known. He has specialized in child second language acquisition and bilingualism throughout his entire career. Very well known internationally, published internationally. There is nothing I will say here that you may not know. Perhaps on a more personal note for those of us who were interested in the field of bilingual child language acquisition, he can be said to be a founding father in the field, inspiring enormous amounts of students and researchers and doing very, very serious work in this field together with other colleagues in Canada. To be brief, perhaps the two main areas in which he has worked more consistently are the areas of second language acquisition in early childhood and bilingualism within educational context and the very special educational context that take place in Canada. With immersion programs where he has looked at the acquisition of lexis, the acquisition of syntax, from theoretical perspective, that is generative grammar perspective, but also other perspectives, but always trying to find to study the situation in education and to find solutions, if one can talk about solutions, or at least paths of ways to deal with the educational issues that come up in our diverse society, with immigration kids, with diverse levels in classrooms, heterogeneous profiles in students and in pupils, right? So this is where he did most work on having an impact in the immersion programs in Canada, together with of course other other colleagues there. The other area where he has been a pioneer and has led pioneer work is in the very interesting area of adopted children and how they develop linguistically and how they, of course, adjust in the educational settings in which they happen to be living after adoption. A very, very interesting area that for adopting parents is always crucial and very few universities in the world have actually undertaken this kind of work. So with no more to say, I pass the floor on to Professor Genesee. We are very happy that you could accept our invitation and thank you very much. Well, thank you Carmen. Well, it is a pleasure to be here and to talk about this particular topic. It's going to be quite different in focus from the earlier one, but you're very fortunate over the week you're going to get a real broad view on different aspects of bilingualism. As Carmen said, one of the areas that I worked on is acquisition of languages in the preschool year. My interest, I should preface this by saying one of the things that unites my various interests in this general topics is looking at children's capacity for language learning. And this will come out very clearly in this talk, but I was motivated initially by a point of view which was very common at the time that I started research, which was in the last century, was that there are costs to being bilingual. There are costs to raising children bilingually and there are costs to children being educated bilingually. And that was a very prevalent view at the time and it really sparked my interest in actually raising questions about how children actually acquire two languages, whether it's in a school setting or in a preschool setting. And I was very lucky to be in Montreal. Montreal is a lot like Marseille-Bona. It has a strong minority language, if you like, French, but it is also facing a very strong majority language, English. And the dynamics between those two languages are interesting politically, but they also create a perfect work environment for people who are interested in research on bilingualism. One of the things that is very nice about Montreal, and I think this is probably true about Barcelona, is that these are two languages which both enjoy very high status. I may not be true here, but in Quebec, even though French is the official language, numerically, English is the dominant language in the country and in the continent. And despite the politics, it has a high status among most people. So that's actually important to consider as you listen to me talk about this research and as we talk about that, because in those kinds of environments you're looking at children's capacity for learning language in an environment which ascribes high status to those two languages. And that has a really, really significant effect on children, their parents, the educational system. Many other areas of the world are bilingual or multilingual, but often there is much higher status assigned to one or several languages opposed to others. So to use Belgium, is anybody from Belgium here? Belgium is also a bi-trimadial, but there are deep divisions over the status of the various languages. And those kinds of effects filter down to families to school systems and can have a significant impact on the language learning environment that children are exposed to. When you do research on young language learners, and you'll see this as I talk about this research, you want to be very, very aware of the general context in which they're learning language because it can have a profound effect on their acquisition of that language. So I'm going to focus on research that has looked at children in the preschool years. But before I start, can I just ask a couple of questions about who you are? How many of you, you're all graduate students or most of you, is that correct? Yes. You're in, are you, how many of you in a program that's in education? A few of you. How many of you in a program of psychology? Most of you. Any linguists? Okay, okay. And you're all, and you're all, and what stage of your education are you at? Are you at first year of graduate studies? A couple? Second year? Third year? Fourth year? Fifth year? Okay, so you're spread out, but those that have been third and fourth year, you're well on your way to finishing, I guess. Okay. So what I want to do is look at bilingual acquisition. Well, as we've seen from the talk this morning, second language acquisition can occur at any time in one's life. And I'm going to focus on children who grow up with two languages. So that from birth or even prenatally, they in most cases, not all children I'm going to talk about are in that situation. But a lot of children who are raised bilingual will have been exposed to two languages before they're even born, but certainly get exposure to two languages or more as soon as they're born or soon after that. And those are the children I'm going to focus on. And there are reasons to treat this group, at least for research purposes as a distinct group, because it may well be, and this is an empirical question, as well as a theoretical one, that they learn language, second language in different ways than people who begin to acquire a second language, say from four or five or six years of age. And I'm going to show you some evidence at the very end, some fMRI data that shows you that these age differences sat in very, very quickly. But I'm going to focus on these children because when I started this research, there had been a little bit of research on dual language learning in the preschool years, but surprisingly little. I started doing this research in the late 1980s, early 1990s, and with the exception of work by Jurgen Meisel in Germany at the time, there was really virtually no contemporary Anne Voltaire and Kessner in Italy. There was virtually no or little research that looked at language acquisition by language children. Most of the research was focusing on monomengal children. And I got involved in this field because I was asked to write a chapter in a book called Language Development in Exceptional Circumstances. And they asked me if I would do a book on children who were raised by Dinkley, and I had never done anything on this. I don't know why I was asked to do this chapter, but I agreed to do it because I thought this would be interesting to do. So I did my, it was relatively easy to do a review of research at that time because there was very little research on it. And the prevailing view at the time, which is a starting point for this story I'm going to tell you, was that when children acquire two languages at birth, they go through a stage when they have only one, really one system that's made up of both the languages that they're learning. And it's only with time that they actually separate their languages into two linguistic systems. And when I reviewed the research, I didn't see evidence. I didn't think the evidence actually told that story at all, even though that's what the researchers were saying. So I said that in this chapter and I sent the chapter in. I didn't know what the other chapters in the book were until I got a copy of the book. And there was my chapter, I think it was like chapter seven. And the other chapters of the book were things like language development in deaf children, language development in children with Down syndrome, language development in children with alcohol syndrome. So clearly the view was that exceptional circumstances meant children who were at risk for language or general developmental disorders. And at that time, there was this view that learning two languages was similar to that. So I think the editors reluctantly included my chapter, because in the preface there was some remark about Genesis chapter, that really was sort of swimming upstream on this point of view. So I just want to introduce that notion to you because even though we probably don't feel that way about why they listen, certainly if you're in a country like Spain or a city like Barcelona, it's everywhere, but it still can be widely prevalent within the community at large. So I want to start off with the question, but why would you actually want to study dual language acquisition? And this is sort of probably obvious for this audience, but I think when you're as graduate students, you always want to ask yourself, why are you doing what you're doing? I just remember I had a graduate course that I took from my family in Donald Ham, who's a very famous psychologist, arguably the first cognitive psychologist, and he said, if it's not worth doing, it's not worth doing well, because often we have a tendency to think that if it's well done, it's worth doing, and he sort of inculcated this idea that the first question to ask yourself is, why are you studying this? So one of the main reasons to look at bilingual acquisition is to understand this form of language acquisition. It's a very widely prevalent form of language development. It's been argued, although we have no real evidence of this, but that there may be more children raised by or multilingually than monolingually. In fact, can you think of any countries that are actually monolingual? Can anybody think of one language? Can you think of any country in the world where you would say, pretty much this is a monolingual country? I don't mean officially, I mean in terms of people who actually have it. Yeah, officially. There's a lot, I guess. Yeah, but in reality. Can anybody? Good point. That was not the one I had in mind. Pardon? Liechtenstein. Liechtenstein? You think so? Okay, maybe Liechtenstein. Yeah, there are a lot of the, the example I thought of was Japan, because I can't, is anybody from that area? Are there significant language or dialect variations? Yeah, okay. But it's relatively homogenous, but yeah, so that's the reason I ask that is because it gives us some evidence that in fact there may be something to this notion that more children are raised in a higher, multilingual community state. And so just by sheer numbers alone, they warrant some kind of investigation. But underlying that in those researchers, the reason for looking at these kinds of learners is that if we have any claim to working on a theory of language acquisition, if we're only looking at monolingual children and to a large extent until the last 20 years, looking at monolingual English speaking children, then this theory doesn't have much generalizability. So looking at these children is essential if we want to build up the facts about language learning in multiple kinds of circumstances so that we can develop a theory that is generalizable to different kinds of language learning. And I'm not talking about language learning that's unusual. This is really, it's not exceptional. Just play the title of that book. It's really very common if your answers are to be believed. Also, and I won't get into this, but looking at dual language learners, bilingual kids, is a way of testing out theories of monolingual acquisition because if the theory is valid, then it should explain and apply to children who are acquiring more than one language. And if not, then maybe it needs to be modified. I once described some of the data that I had on these bilingual children looking at grammatical development. And it was a case of children learning English and French and they were learning grammatical structures that were incompatible with one another in the two languages. And so I showed it to this linguist colleague of mine who was a generative linguist, Chomsky, who had a very strong theory of language and language learning. And he sort of dismissed these data, even though I thought they were very relevant, because it was about a constraint that was contradictory to languages. And he just didn't want to discuss it because from his point of view, the variation you see in these children is only variation. It really isn't substantive enough to really have theoretical implications. Now, I'm not saying all generative linguists would think that, but I think that a good theory has to apply to not only monolingual children and not only to English, but all kinds of language learners. Now the other point, and this is something motivates me to quite an extent in all my work, is that given that these children are very prevalent in our community and more and more so in Europe and in North America because of enormous levels of integration, we need a good empirical evidence to provide professionals who work with children who are bi or multilingual. And here I'm referring to educators, physicians, clinicians, anybody who comes into contact with a family where the child might be in the process of learning a second language really needs to have good empirical evidence about how these children develop, not just linguistically but in general, so that they can make proper recommendations. Physicians, for example, are probably the gatekeepers for a lot of clinical services for children and they're probably providing advice to parents about what to do with their child, even in matters related to language, even though they may know nothing about language. I mean, in Canada often parents would be advised by clinicians, doctors, physicians, that they suspect the child has some kind of developmental disorder down syndrome or a specific language improvement and if they're raising the child in Spanish or Arabic that maybe they should use only English or only French because exposing the child to two languages with these kinds of difficulties early on in the development is going to add challenges for this child and he or she doesn't need more. So if you have any kind of clinical interest at all, keep in mind that the impact of this kind of research is really quite broad because it's not just professionals who work with language issues but it's also physicians as well and certainly educators. And of course, there's the issue of policymakers and those of you who live in the arsenal know that there's major implications of this research with respect to policy and practice, although I'm not always convinced as researchers we actually have much of an impact on policy. I remember once we were we applied to do research in the school in my community on looking at the role of phonological awareness in reading that position in second language learners. And this was a school, this may not be very meaningful to some of you, but this was a school that was very heavily enthusiastically about whole language, the idea that you should teach reading using a meaning-based approach and you should avoid teaching phonics and decoding and all this stuff. It was a very top-down approach to teaching reading and it was very popular at the time. And even though we had done a lot of research in the school board for many years, they refused to let us do the research because they felt that this approach that we were examining, which talked about phonics, was not compatible with their policy which really advocated whole language. And I remember saying to the school administrator, I wish I had that much influence over school policy because usually we do this research and it has no effect whatsoever on policy language. But in the long run our research can have an effect and so if you have interest in the applied area, these are audiences that you want to think about. So I won't go over this because I've kind of covered this now already, but there are a lot of us, the one thing that's exciting about doing this kind of research is that it does matter actually. People have lots of questions about these issues. So even if you're a really committed theoretical researchers, a lot of what you may do is going to be of interest to people outside the research community, whether you want it to or not. And I like to present this email. This was from somebody some time ago and I particularly like this email. First of all because it comes from a father. It's very rare that I get emails from fathers. It's usually mothers or women, very seldom do you get men, a lot's changing, asking about their children's language development. But this was from my father. And just read through this. This is only about a third of his list that I've included. And of course, for a researcher, this last bit is wonderful because without questions we couldn't do research. But these are questions that parents ask themselves, but these are also issues that educators are often concerned about. We had a lively discussion at lunch about some of these things that despite the fact that we may think that there's no problem raising children by labeling or educating them by labeling, lots of people are worried. Parents are often worried about these kinds of things. So that's, your research can be important. Here's one from a school psychologist. And this is in Montreal where it's hard to, in the English speaking community, it's hard to find a modeled English child. And yet you still find these concerns. So here's a school psychologist working in the school system. She's facing a child who is thought to have a specific language impairment. And she doesn't know what to do. I mean, that's a serious problem because this is somebody who is a gatekeeper for this child and she doesn't know what to do. And, you know, she's honest and says that what I, what everybody around me is saying is that learning another language is putting too much pressure on a child who already has trouble learning the language. And then here's an interesting thing from the LA Times, Los Angeles Times, that's really interesting because this was sent to me by somebody in the US and they sent it to me because the headline is actually quite positive, which is not always the case in the US media, but then read the rest of it and you'll see that again there's concerns. So they're basically saying it's great when you're bilingual, but it's challenging to become bilingual and there may be reasons to think that the process is really difficult. So that's a lot of what I'm going to talk to you about is, is this difficult and how do kids actually do it? So anybody want to share any comments about this? Do any of you have, if you had children, any of you have concerns about raising them bilingualy? Right, right, that's a very common concern. Yeah, yes. And particularly if you take data from the urban areas like Delhi and Mumbai, where several families are there, where India need for it to strike into the community. Right. So our children, they go and there are several models, two sometimes seven languages, and at the age of three there are seven languages, they're very smart. If father does say for example, a speaker from Marathi and mother is a speaker of Hindi, and both of them have short language as they can. So children are very competent in doing conversation from the fathers, family, members, they don't come from Marathi, the mothers come from there, they come from Hindi. Correct. Well, and so here we have two different contexts which have quite different views. The U.S., where often there's a lot of concern, I think in a lot of English speaking communities, because English is such a dominant language I think people who speak English almost think that, well, it's normal to speak English, you don't need these other languages if you wait long enough and everybody is speaking English and you don't have to worry about it. But they also problematize bilingualism in ways that you don't find in other communities. But it also illustrates why it's important when you're reading research about the children learning language, or if you're doing this, that you pay careful attention to the context in which they're doing that. Because there may be differences and I'll show some of those differences to you in the results you get. And some of those differences may be related to contextual factors. And what's interesting to ask is how does the social context influence the way schools or parents interact with children and how does that affect their language development? So it's highly variable. These attitudes are not necessarily the main way people think about it. It varies a lot and certainly this group here has probably more open views about this than other groups. But it is a concern among other people. So what I want to do is I want to start you with a little bit of history. In case it shows my age, I always think history, I agree with Juven. It's important to know what happened in the early stages of these kinds of this kind of research. So I'm going to look at research that had looked at code mixing and you'll understand why. Then I'm going to review research on, that's attempted to look at the process of language acquisition in bilingual children. Some of the best research on this has actually been done in Spain and some of it at this university. So I'm a little bit anxious about getting it right. If you're studying this topic here and you see any mistakes I've made in presenting this, keep it to yourself and tell me later. Because there's a lot of this research, it's not the kind of research I do. I focus on this three to five, two to five year group and this is really a very young group. You'll see it's remarkable the research techniques that people have devised to look at very young children. How many of you are working with preverbal children? A couple of you, two or three. Okay and it's really quite remarkable. So keep an eye on what I say and tell me if I've gotten any of it wrong as we as we go along. And then I want to end off with talking about what the long-term outcomes for language learners are and this is really an issue of looking at the critical period hypothesis, but from the other end of the critical period. You know most people define the critical period as being around 12, 13, 14 years of age. There's been very little research that's looked at how early there might be a critical period. And I want to review some really interesting research that's been done in Sweden and some research we've done on internationally adopted children that looks at children who begin to acquire a second language very early, 12 months of age for example or three years of age, and looks at their long-term outcomes to see if there's similarities or differences in what they look like in comparison to monolinguals and what those differences mean. Because I think one of the issues that permeates all of this research at least in my mind is what does it mean to be a native speaker? And in so much of the research on language acquisition, whether it's in school settings or preschool settings, the gold standard that people use to evaluate language development in children is typically developing monolingual children. And that raises a whole lot of questions that we really need to answer because then when you start to look at second language language or bilinguals you're starting to make comparisons that may not be fair or may be inappropriate and they're probably not very informative with respect to theoretical issues. But we'll come to that at the end. And also this is research that we've done using neural imaging techniques so and it's very interesting research and I wanted to share that with you. Okay, so this is a starting point for this round of my talk. This is one of the very first studies. How many of you have heard of Werner Leopold? Yes, you should have. If you're studying bilingualism and bilingual acquisition. This is the second study as far as I know that ever looked at bilingual children and you can see it was done in 1949 a long time ago and he actually worked in the US. He was German and he studied his two children who were learning English from his wife in German from him or vice versa. I forget how it went. Does anybody remember? It was German and English in any case and he did what is arguably the most extensive single case study that anybody's ever done. It was written up in nine volumes. I didn't read all nine. I read a summary of it. And this is one of the things that he says and I just want you to read that Hildegard was his daughter because it's relevant to the research I'm going to describe. Okay, so what he is seeing in his daughter in the very very first stages of her acquisition of German and English which was occurring from birth in the home. This is a very common situation. How many of you have children now who are you're raising bilingualy? And in that situation is the spouse speaks one language and you speak the other one? Is that the case with all of you? Okay that's very common. It's not the only situation. You may have a pair of child care services which are in another language. But this is a very common situation and what he noticed was that he claimed that when she spoke she really used words from either language whenever she spoke and it was she did this rather indiscriminately and the reason she did this was because in fact she there was only one language from a neurocognitive point of view. So in his point of view in her mind there was a language a hybrid language made up of English and German so when she spoke English or German words came out and it was only later that two speech what he calls two speech systems emerged. Here's another quote that was widely influential at the time and basically they're saying the same thing that in the first stage the child has one lexical system which includes words from both languages and the implication is that there's no grammatical system and that only in the second stage the child has two different lexicons or vocabularies but applies the same syntactic rules to both languages. Does everybody know what morphosyntax means? Who does not know it because I'm going to use these terms a lot. It's just a fancy word for referring to grammar and the syntax is the word order of grammar and morphology is the little pieces of language that ties everything together word endings functional words and so on. So if you don't understand any of this as I go along please raise your hand because if you don't understand other people or not. So it was only in the third stage they argued that the child speaks two languages but separated lexicons and separated grammatical systems. So you see this thing and if you looked at more contemporary research you would see the same notion that children were learning a unified system and the evidence for that was that they were code mixing and this business of code mixing is very interesting because lots of people who are not bilingual think that code mixing is a bad thing to do. Educators who are this probably is not very common here because I would suspect that do most Catalan speakers also speak Spanish? Oh okay seeing Quebec a lot of French speakers would not speak English but a lot of times monolinguals think that mixing languages is a bad thing and that it it implicates that you have you're making mistakes okay so it's just a bad thing it's not seen as something that's normal or good and there's two kinds of code mixing one is this kind where you're mixing within an utterance and one is where you mix between utterances so in this case I'm not going to talk about these in any detail but the mother is showing the child a picture book and asking the child what's this and the child says cheval which is finished for horse and then the mother says what's that one and the child says doggie so this is a case of mixing okay not within an utterance but between utterances it's these cases that are interesting because these are the ones that are most caused for concern because people even researchers used to think that you shouldn't do this English words should occur in English sentences French words should occur in French sentences so there was a lot of research actually that used code mixing as an example of this unified stage of development but there are a number of problems I'm going to I'm going to go into some of this methodological stuff because you're graduate students and you need to know some of these things but you find in the early research on bilingualism that the methodology is that often quite can be quite weak because people don't define their bilinguals very well they don't have a good conceptualization of what language they're studying and so on so it's really important when you read this stuff when you do your own research that you really pay careful attention to what you're doing now one of the problems was that a lot of the early evidence like Voltaire and Tescher and the one from Werner these were case studies and while case studies have a very very very important role to play in language acquisition research and I'm going to show you some case studies you have to be careful about generalizing from a case study to a whole population or you have to be careful to generalize about a whole linguistic system they also when you go through these reports they will often give you examples of the child code mixing or code switching it's often called code switching in the adult literature the child acquisition research tends to use the word mixing just to try to be more neutral but they tend to look at some examples of code mixing and they haven't looked at the non-mix how often does a child not mix because if a child only mixes once or twice during an hour conversation with an adult what's that telling you quite different from if the child is mixing all the time they also didn't look at the the real context for the mixing in any depth so they often did not look at whether they were mixing into the dominant language or mixing into the non-dominant language and in a related vein they didn't look at the general context so that was a starting point for some research that I did and I probably won't get through all of these studies but I'm going to discuss a couple studies I did on code mixing early on this is not recent research but really I think started other research to disputed this argument that these kids are confused so I want to give you just the context because I said context is important these are kids who are learning French and English at home this is not the best combination to look at nor is Spanish and Catalan for that because these are very similar languages you might not think French and English are very similar but they really are there are very few differences in word order between English and French there's a lot of cognitive vocabulary the morphosyntactic rules differ a bit but not nearly as much as other language pairs but that's the pair we had to work with we're looking at children in the one and two word stage of development how many of you don't know what one word stage means or two word stage everybody knows that okay one word when children acquire language at least children who know languages like English and Spanish when they start to talk they usually produce a single word at a time and then they start to produce two words at a time and then they go on to produce more complex utterances so people in the field refer to this as the one word stage the progressive two word stage and then they start to produce more complex utterances in other words children don't start off producing multi-word sentences they go through the symbol the symbol trajectory as I said already this the context is very particular so we're looking at a community where both English and French are widely used and highly valued French is actually the dominant language in Quebec 85 percent of the population speaks French many of francophones don't speak English 15 percent of the community is either English speaking or what we call elephones speak another language so it's a minority language in Quebec but it's obviously a majority language in the country so I'm going to this all when I go through this research one of the things that I want you to keep in mind is that this research on these very early stages of development and the way this research has been done are really done in order to look at children's capacity for dual language learning and they're not intended to tell you what what what is a typical pattern for children in this community or for children who are raised generally okay and this will become a bit clearer as I go on but the argument here is that look if you think that children who are raised bilingualy are confused initially and they mix their languages because they're actually non-cognitively mixed up then if you do research even with single case studies that shows that that's not what these children are doing that is is telling you well if these children are able to separate the two languages at this stage then in principle any child could but it's not saying that all children do the same thing do you see the point I'm making so you have to be really mindful of the fact that this is not research that's attempting to be to describe typical patterns in all children because the patterns that children might typically show would require larger samples and it would also require that you take context into account more carefully all right so as we go through this I'll try and remind you of that so the very first study we did just to give you an idea how we did this this is a very simple it's embarrassingly but it had the research that hadn't done at that time didn't do that we used this research this the first time I should say yeah this is the first time I'd ever worked with such young children I always worked with school-aged children I was quite terrified actually because with very young children you can't control very much they just do what they're going to do so what we did these children were in the one and two word stage of development so they were you know between 12 months of age actually more like nine months of age and a bit older and what we did is we simply observed them with their parents in the home playing with toys that they would normally play with and the reason we did that is we wanted it to be as familiar to them as possible and we wanted them to be as verbal as possible so we we didn't feel we could do anything that was too unnatural or too experimental now what we did which was really very simple but which really had not been done is we systematically looked at these children using their languages with the mother on one occasion and it all just turned out that in this particular study all of the mothers spoke English and in another occasion all they spoke with their father and all of the fathers spoke French so that makes it easy to describe the results it's not that way and all families in Quebec that just by chance was what happened in this case and then and the idea behind this was if these children really have a fused linguistic or unified linguistic system then the distribution of English words and French words should be random and it should be equally random whether they're speaking with their mother and with their father now these were contexts where the mother most often used French English and the father most often used French now parents this is what's called the one parent one language rule most bilingual parents know this rule and most parents I would hesitate to guess think that it's important to use this rule so the child isn't confused how many so how those of you are raising your children by language do you follow this rule no you don't did you did yes but one the very often one fence yes it's almost impossible to keep to adhere to it a hundred percent out who else is raising children by language what do you do and you don't speak Japanese Anglophone fathers of Japanese wives very seldom speak Japanese right so that was one of the concerns that we have is how do we ensure that our child maintains what language is right and that's enough before she went to school right she was very much she spoke a lot of English right after she came to school did she go to school in Japanese yeah okay that's the other reason for doing this but a lot of time parents do this because they think that that the idea behind this is that if the mother speaks one language and the father speaks the other language then it'll be clear to the child that they're learning two languages and it'll be clear that this is one language and this is the other one the fear is that if both parents use both languages the child won't be able to separate the two languages and that's a nobody's really looked at this entirely carefully but that's an interesting issue but there's a lot of evidence to think that children have no trouble detecting that there's two languages even though you may be mixing them anyway we so that's why we did it is that clear so what we simply did is we recorded the kids for about 20 25 minutes 30 minutes we then uh we were in the room videotaping we didn't interact with the kids we then took the audio tapes and the videotapes we transcribed them we did we transcribed about if i remember correctly about five to ten minutes of these videotapes how long do you think it would take to transcribe five to ten minutes of an interval how many four hours ten or eleven at least because these kids are very hard to understand parents think they understand but as a researcher you don't that's very very time consuming and then because we wanted to really see how well these whether these kids were really confused we decided we're going to look at them uh when they're with the mother and father together because if they really use it if they're really confused then having the mother and father in the same room should really confuse them right because everything's mixed up at that point so is that more or less very straightforward nothing very sophisticated technologically great because it was very simple okay and there's the rationale so here's the data we have six children uh so this is a this is better than a case study but this is hardly the sample sizes that you heard about this morning these are very modest sample sizes but it was an improvement over what had already been done because we have six case studies and this is looking at the children's using English so as we transcribed it we tagged as best we could whether the word was an English word a French word or something we didn't understand so this is uh these are sort of surnames for the kids red is using uh English because remember the mothers are all English speaking so red means using English with the mother and blue means using uh French with the father did i get that right and you'll see that in every case but Jean that the children are using more English with the mother than with the father okay uh which is hard to reconcile with the notion that they're confused because the mother's language habitual language with the child is English and they're using more English with her than they are with the father this is interesting for number of reasons first of all when you do this kind of research with young children you always want to be sensitive to individual differences because there are huge individual differences this kid Jean used English in French equally with with the mother and father used English equally with the mother and father and it's this child is interesting because he was probably the most proficient given his age most proficient of all of the children we looked at so if confusion is a sign of immaturity in the language he wasn't he was the most advanced admittedly not very advanced because he's pretty young but also his parents from what we could tell were the family that code mixed the most within the family so it was a form of communication in this family which was more prevalent than in the other families and the point going back to a point the human made is that you know looking at exposure and things like that it's really really important if you can to look at exposure so not only how much time the kids spend in each language but who's using which language but it's very hard to do because we could have done that during these sessions but that doesn't necessarily mean that this is what the parents do when they're with the kids alone without the experimenters sitting there but if you can do that kind of research it's really important to try and do it because then you can correlate the patterns that children use in the family with what you're seeing in the lab and is there any relationship so there's no evidence here that this child is confused here is the use of French alone remember French is the father's language again in every case you see the child is using more French with the father than with the mother some cases very little French in fact even though that is the father's native language or dominant language and again gene is doesn't care too much here's the when the this is the most stringent test of it here's the kids use of English when the parents are together playing with the child again there's more English which is read or the mother's language with the parents even in gene's case and there's more use of French with the father English from French and then when we look at French there's more French with the father in every case so interestingly when the parents are together gene is actually differentiating more than when they were separate right so it's hard to reconcile those results with the notion that these kids have a single system at this stage of development we then just to look at this a little bit in more depth I sort of had got this idea what do you this is interesting how how much control do these kids actually have over their language because one of the things that is of concern about children raised bilingually is do they really are they really communicate or are they compromised and if you're really bilingually competent then you you are able to adjust your language use with people that you're just meeting you know any form of community competence means that you're you're competent using the language with anybody basically it doesn't mean that you can communicate confidently the people you've been living with for a year or two that's not really confidence because maybe what you've done with your parents is you've just learned by rote to use this language with the mother and that line with the father so what we wanted to do is look at how children would use language with people they didn't know and we wanted to see whether they were whether they were sensitive to the language that was being used with them and it's therefore whether the the input that this this stranger gave them would influence the way the children themselves used each language because another another thing about community competence is you tend to adjust your language use to the people you're speaking with and by looking at whether they were sensitive to the input it would tell us that I reinforce our notion that the code mixing that they're engaged in is not driven by internal competence issues but is driven by a more social and external kinds of factors okay so again it's the children are two and a half years of age so they're well into the two-word and multi-word stage of development these stages of development by the way overlap it's not like on Tuesday it's one word and then Wednesday it's two or three words and then if they overlap a lot so these children are predominantly you can use more than two words in a single audience the children met with these people they were strangers they were they were confederates of ours they were people we trained to interact with these children they had never met these people before so they had no experience with them and they so whatever language they were using with this adult they had never met had nothing to do with their past experience it was really it was really a reflection of their perception of the situation at the time and their ability to deal with that situation now we had we trained the speakers to use the the child's weaker language initially because we wanted to force the situation where the child was compelled to use their weaker language because you might expect and this is what research shows is that if children are using their weak language they're prone to use the other language because they know it so we thought well by having an interact with somebody who's using a weaker language we're testing the limits of their ability to control that tendency okay and they met with this stranger on three occasions on three separate days in the first day this the stranger code makes 15 percent of the time so the stranger say let's the week say the weaker language is French say that's the father's language and therefore the child's weaker these children are usually stronger in the like the mother's language because despite all these liberal language fathers mothers still be most caregiving and children tend to be stronger in the language that they're hearing the most naturally so they're using the the strangers are using the child's weaker language 15 percent of the time but otherwise is using I'm sorry is using the stronger language 15 percent of the time but otherwise is using the child's weaker language 85 percent of the time and then on the second occasion the child is hearing the stranger use the stronger language 35 percent of the time so the point here is not particular percentages the point here is that the child is hearing somebody using more or less of the other language and we wanted to see whether the children would adjust and change their code mixing accordingly and then we did a third experiment where they shifted back down again because we were surprised at the results okay we did this by the way you know typically in these situations the the child is there on the on the rug in the living room with the experimenter of the parent and we have somebody sitting in the corner video recording the interaction and the video recorder in this case was trained to count how often the stranger used each language and in this case for example every now and then the video recorder would indicate to the stranger it's a lot to increase or decrease their code mixing so that we could hit this 50 percent target this is very hard if you try to do this on your own it's very hard to control your own code mixing to reach a percentage so the experimenters guiding the the inter the conversational partner of the child to reach this target is that here so it's a it's kind of a semi-experimental procedure so here's the results this is again we have more than one child there's what three six children again different children this is the high rate of mix low rate by the interlocutor the high rate by the interlocutor and the going back to the low rate so you can see that in every case the except for this case that on the first occasion there's code mixing that's very little and then on the second occasion when the the blue line when the stranger enhances her code mixing the child also enhances it so in every case that happens so the child is actually responding in accordance it's code mixing in accordance with the rate of mixing of the stranger right the reason we did we went back to a low condition again is we thought well maybe what's happening here is that on the first occasion the stranger is really unfamiliar to the child on the second occasion the stranger is more familiar and therefore they're just code mixing more because it's easier for the children to code mix because when they code mix they're able they're using their stronger language so we thought well let's test this by exposing the to the child to the stranger again but this time the stranger lowers the rates of code mixing and we're forcing the child to lower his or and see if they can do it and they did again except with the exception of this one child so you always have to be sensitive to individual variation and this tells you kids are you know these young kids respond in different ways but again what we're seeing is that there's a very there's very sensitive fine tuning of what these kids are doing this is when you combine all of the children okay i'm not sure what the difference is i forgot um now we had a very well i had a very uh i was very enthused about these results when we got them and i was thinking gee this is interesting how the kids how are the kids doing this this is rather complex so they kind of got a mental calculator where they're they're keeping track of the adults rates of code mixing and they're some or other calibrating their rates of code mixing accordingly but it turned out to be much more mundane than that because when we did a detailed analysis of what the kids were doing what we found was that they would be using their weaker language let's say it's english the adults using english most of the time and then the minute that whenever the adult used the other language the term immediately after that the child used the other language so the reason why their rates of code mixing were very similar to the rates that the experimenters were using is because they were tracking but on a case on a turn-by-turn basis what the adult was doing and if the and if the adult used the other language in a sense that said to the child it's okay now to use your other language and that's what they did it's very pragmatic but it works because that's the nature of social interaction yes oh right these we didn't look at that but I have another study I could show you that actually looks at repair strategies so I probably won't have time to get into it but the kids are there I kids are very sensitive if there's a breakdown communication so and if the breakdown in communication I mean there's lots of breakdowns in communication between children this age and adults I mean if you've ever talked to a two year old or two and a half year old there's lots of reasons why you can't understand what they're saying either mumble they mispronounce a word they used a long word they're off topic but in this case bilingual kids can create a breakdown in communication because they use a language that the other person doesn't know I mean but it's interesting when you look at it for most children who are being raised in a monolingual environment the adult is more competent than the child so if there's ever breakdowns in communication the adult can repair the breakdown or knows how to repair the breakdown bilingual children are in the interesting situation that if they're talking to a monolingual the monolingual is less competent than they are right so we did a study where we looked at children's ability to repair breakdowns in communication when they used the wrong language with them and they were they were if they said something to the interlocutor in the wrong language and the interlocutor said what didn't say say that in the other language just said what they switched to the other language again these are two and a half year old children so all of these studies taken together really indicate how competent these children are so there's again there's no evidence for this theory of code mixing is a reflection of incompetence so then the question is why do they code mix and the obvious reason the obvious reason is that they code mix because they don't know the words in the language that they're using I mean all of you probably figured this out but again nobody had really looked at that so here's a study with the initial group of children and we simply looked at how often they code mixed when they use their weaker language most bilingual kids are stronger in one language than the other and as I said it's usually related to the amount of exposure to that language so here's the six kids five kids who used and we looked at whether they were code mixing from one others to the other as a function of whether they were using their dominant or non-dominant and what you can see is that all of the children except these two three of the five children are using they're they're doing more code mixing when they're using their non-dominant language so they're talking away in say French and they come across a word that they don't know in French but they know it in English so they use it in English right makes sense it's a it's a you know monolingual children have a phenomenon over generalization so monolingual children if they're talking about something there's a famous example of the mother out walking her baby and the postman goes by and the postman sticks his head in the bag and says okay baby and the baby says daddy right and the mother's very embarrassed because this is not the daddy the daddy's at work and the the point here is that the child doesn't have is is wanting to refer to the postman but the only word he really knows that refers to an adult male his daddy and that's the word he uses so it's a form of over generalization well in a sense what kids are doing here is a little bit like that it's a strategy for filling in gaps in your lexicon when you have to talk about something and you don't know the word and these bilingual kids have the advantage that they can use the word from the other language and in most cases it works because in their in their families and in many cases in their community the people around them know both languages and so they'll understand what's being said again individual differences so uh this fellow gene uh didn't show that pattern but as i mentioned before he was actually equally competent both languages so it was not an issue for him william we don't know what's going on yes they don't need to switch right right but for them or they couldn't switch because they don't know the same in the case of the bilingual often that's the case at least most of the research shows that if you look at bilingual older bilingual children and you look at their vocabulary in each language separately and you can't compare them through age matched peers yet often the bilingual kids have smaller vocabularies in each language yes so let me turn this around and see if i'm saying that the challenge here is not really the so much a challenge for the child but it is for the parents because they have to create an environment which makes sure that the child learns all of the words that they need to know and therefore by having the parent use each language mother french father english you have a better chance that the child will get exposed to all of the words that are relevant to their environment is that yeah right well that's why when you look at these kids you you have to be careful that you're looking at the environment in which they're growing up because if they're i'll give give you a bit of this later on but if they're growing up in an environment where the distribution in in terms of how much of each language they're hearing is balanced then it's reasonable to think that their vocabulary in each language is going to be pretty comparable whether it's the same as a monolingual children is is enough question but if they're getting a much more exposure from the mother than the father then their knowledge of french uh the mother's vocabulary language in that the cat language is really much higher than that of the father yeah well it depends on whether you think that their parents are justified because of the confusion issue or the parents are justified in terms of the size of the tabular these kids are not confused they just don't know the words in both languages yeah so in that sense that's right i think the one parent one language rule makes a lot of sense but it's good for the parents because it gives them a way of structuring the environment so that the children do get good exposure to both languages is that well that's right now at this stage of development kids vocabulary are very limited because their their lives are very limited uh it becomes even a more acute issue as the kids get older because then their environments get much more complex but then what happens is we're getting sort of a little bit away from these data but you know you shouldn't we shouldn't expect that bilingual kids are going to have the same vocabulary in both languages and can you think why social pressure right so if you're they don't necessarily you shouldn't expect them to have the same vocabulary because the world in this the home language is quite different from the world in the school language or their friends language so they're only learning the words they need that and so that so there may be an imbalance in the vocabulary because their world is more complex than that of a modeling but as a juvenile if you look at the children's uh combined vocabulary in other words you give them credit for every word they know for every concept regardless of what language it's in it's called conceptual vocabulary you find that they are on par with modeling your children in some cases better because they know the names of some objects in both languages so that's important the fact that they have an unequal distribution in the two languages is is at this stage due to limited memory probably but also due to the fact that their world in each language is subdivided they might even with the mother the father in the home what the mother talks about with the child and what the father talks about the child are not going to be the same thing so they might know certain things about the home from the mother that differ from what they know from the father right the question is this gender maybe have an impact we've kind of looked at that informally i've never we've never found in other words is the mother going to influence the daughter more than the son and buddies first we've never seen that the mother has an impact on both children because the mother spends most time with the child and also mothers are more child centered than fathers i mean there's that's quite a bit of research on this that mothers are more likely to talk about things that the children want to talk about sound familiar but the father is more likely to talk about things that the father wants to talk about anyway there's more research to be done on that here's another here's another study that looked at why do they code mix we just looked at two kids again again this is about what can happen not what does happen and we wanted to look at banning your kids often have a translation equivalent so they know the names of an object in both languages so they know that this is bottled in english and mutay in french okay so we wanted to see if they had more translation equivalents how much they code mixed whether they had the translation equivalent or not so the argument is they're going to code mix when they don't have a translation equivalent so we did a diary study with parents where the parents recorded what the kids were saying the language that they said it in and every time they code mixed they were asked did the child know this word and they have a language and what you see is that for wane 90 of the time that he code mixed had no translation equivalent for that word as far as we could tell for phoenix these were very young children i haven't got the ages here but they were you know they were really in the one word early two word stage so they're really around 12 months of age more or less he code mixed about 65 percent of the time when he didn't have the translation equivalents now what's important i mean this is all very simple stuff and probably stuff that you especially do got by the children you you've seen but what's important about this when you put it all together is that by looking at this phenomenon more carefully the results are are indicating competence not incompetence remember we started off with this this unitary language system hypothesis where people were saying that this code mixing is really a sign of incompetence insofar as they did not have separate systems well this research is saying not only they have separate systems and i'll show you other events of this at an even earlier stage but they have a lot of control over the way they use these languages so the code mixing is actually a sign of competence not incompetence because it isn't random and they can control it to quite a degree i won't get into this this is a really interesting study but i've kind of given you the results on this okay so from so this is i'm going to just jump to the next section but this is research that really was fairly was fairly prevalent early on when people started to look at bilingual acquisition and this is in like the 1990s and then and it really sort of dealt with this issue of whether people think that there is a single system or not and it basically says they don't they have they seem to have these two systems and it's it's evident at least as early as the time when children start to produce single words so any comments or questions on and again this doesn't this what you have to interpret this very carefully this really speaks to the issue of children's what children can do children can keep their languages separate if the environment indicates that there are separate languages if there are lots of communities where the language that is commonly spoken is a hybrid of two languages so if you look at those children you're not going to necessarily see that there's evidence of separate language systems but in this community and in most communities where there's separate languages that's what kids are exposed to and they know this they don't know it in a conscious sense but they know it intuitively and that that's indicated by these kinds of results is that is that okay okay now i'm going to i'm going to step back and look at research this is not research that i've really done this is research that has taken off from this early research and it's been looking at more at the product more at the early stages of new language acquisition and it's kind of trying to look at the process of early language acquisition beginning in the pre-verbal stages of development looking at language development even before kids are verbally competent all the research i talked about up until now is really restricted to kids who can actually use language whereas some of this research is looking at children who are pre-verbal and it's the same it's the same sort of issues as why do why do you do this well there's lots of practical reasons but the theoretical ones are quite compelling and a lot of the research here has really been designed to ask the question how do children actually do this how is it that they can acquire two languages at the same time that a lot of children are only acquiring one do they go through the same developmental stages as monolingual children again the framework here is always with respect to monolinguals and i want to raise that issue in a moment but do they go through the same milestones do they go through these milestones at more or less the same age of development age of development or do they are they slowed down by this process you know it used to be that a lot of the really early research on bilingualism and some of the kind of cognitive research that human was referring to the cognitive stuff some of this was done in the 60s and 70s walley nambert and paul collars in this group a lot of the thinking at that time was based on the very simple hypothesis that learning two languages either a good thing or a bad thing and it was good because it was bad because learning two things is twice as hard as learning one that makes sense it's harder to do two things in one but there was also the other point of view that learning two things is good for you because you exercise your muscles but that's that's as simple as the theorizing was to a large extent it was a little bit more complicated but not much more but the whole field has moved into a stage of development of its own where the questions are getting much more sophisticated and much more theory-driven so the questions here are well if they're able to do what i've shown you they can do in these conversational studies how do they actually do that how do they acquire two grammatical systems how do they acquire two lexical systems do they do it in the same at the same ages because that would mean that they can do the same thing in half the time that monolinguals need to do it okay one of the similarities in processes that underlie dual language development versus monolingual development are there differences and what are those differences related to and fundamentally there's the question is do models of monolingual acquisitions apply to dual language learners and if they don't maybe we need to change our models of monolingual acquisition if we have any pretense to developing a theory of language acquisition in general and ultimately it seems to me it raises the question that i mentioned before is what does it mean to be a competent language user so we always use you know where we say well the child is good it speaks like a native speaker and we invariably mean like a monolingual native speaker because these children are native speakers right they don't and the canadian census form now when you're asked to give your native language you're allowed to enter more than one language i don't know whether that's true here but it should be because there are a lot of people who do not have one native language but we still really implicitly if not explicitly use monolinguals as the frame of reference to determine what a native speaker is and it seems to me as all of this research especially the neural cognitive stuff as it moves forward really calls for us to really broaden our conceptualization of what it is we're studying and how we categorize these phenomena any comments so i'm going to break this up into kind of two segments before i get into the one on the very last one and i'm going to look at what i would call foundational steps this is where i'm sort of on shaking ground because this is not my research this is research that's being done by other people but i'm going to look at two domains of development that have been done with these children one is looking at language discrimination and speech perception and the other is looking at early word learning and vocabulary development now the reason i call these foundational is that you have to put your sort of head in the head of a five-day-old child and so you're wanting to learn language and you're a dual language learner so you have to learn two languages and the question is how do you break into those two languages this is true for monolinguals when you acquire one language and you're you're born you can before you and you hear all this noise around you how do you figure out what's noise and what's language and how do you then start to break the language up into bits and pieces which you can then combine into words and grammatical sentences right so it's like listening to a foreign language for us how do you deal with this incredibly complex acoustic system okay in the case of these children it's a very same challenge but in a sense it's much harder because they're getting two sets of noise linguistic noise and they have to sort them out right so a lot of this research in my opinion is really looking at these foundational steps and again what it doesn't do it's important to understand what it doesn't do as well as what it does do it doesn't necessarily tell you despite what I probably described this presentation what it doesn't do it doesn't necessarily talk about the influences on these children's language development I thought I was going to do that and then I realized there was just too much work here what it's really looking at is children's capacity to sort out this very very rich linguistic input that they're getting as soon as they're born okay it seems to me it's interesting I was on a I was recently on a panel for a thing called the national academies of sciences engineering and medicine in the united states so this is a federally funded organization which is commissioned on occasion to do a review of research on a specific topic this is something that's been going on in medicine for a long time where and I think this happens in other countries they want a panel of experts to sit down and sort out the research on treatment for breast cancer or treatment for prostate cancer or treatment for toothache because they want to make decisions and policies that are evidence based but the evidence is overwhelmingly complex so they get a bunch of people together and they say look would you work for us for two years and we won't pay you and we'll give you a lot of work to do but this is what we want you to do and people for some reason rather I still don't make sense I agree to do this it was one of the most horrible experiences of my life but I learned a lot but it was a lot of work and we were asked to do one on how to improve the educational success of English learners in the U.S. the kids who come to school and don't speak English are referred to English learners the really children or Spanish speaking for the most part I and the term here would be English no I'm not English learners obviously immigrants Newcomers okay yes sometimes they referred to English in additional language in the English speaking world there's lots of terms I was on this panel and I was working with the preschool age group and even though preschool is not relevant directly relevant to educational issues per se when you think of it looking at these children in the preschool years is critical to looking at their subsequent development in rather school right and in the case of dual language learners kind of a neutral term it's critical that we look at this research in order to understand the capacity their ability to learn two languages and and and whether or not they do or not may not be their fault it may be due to circumstances that they're going up in so I was on that subgroup of that so we looked at a lot of this research of these children from about zero to three we looked at the research that was looking at these foundational steps all right and then that leads into research that looks like kids morpheus and tactic development which because once you've got these foundational steps under control then you start to build more complex grammar and this will become clearer as I go forward so you know one of the things that emerge from all this is is that when you look at this research no matter where it's done a lot of it's been done here a lot of it's been done in Canada a lot of it's been done in the US the point of comparison is invariably monolinguals and I don't know about here but in in Canada in the US this is a very controversial issue because people think that it's unfair to compare dual-language learners with monolinguals because they're different I don't know what's your view on this is this an issue here it's a big issue for us because there are a lot well you have a lot of immigrant kids too but you know you have cities in Canada like like piranha where 60 percent of the kids in school do not speak english when they come to school or Vancouver that's a good point I mean using so using the monolingual as the framework references problematic for that reason that who are these monolinguals there's more and more research on variation in monolingual children not as much as you might imagine actually but I think that's going to happen more and more but what about comparing these immigrant kids let's say immigrant kids for the sake of argument although a lot of these kids will not be immigrants as such because they may have been born here but they speak Berber or they speak Turkish because they're in a community in Barcelona where that's the dominant language in in in Canada these kids can grow up speaking Arabic because in their immediate family community that's the only language to hear so we're not immigrants but then we also have a huge immigrant population as you do and the kids are their parents have immigrated and they speak only that language and we invariably when they get when we're seeing them in the preschool years or certainly at the school years we're comparing their performance in the majority language with monolinguals yes right yes that's what that that's how i would answer did you want to comment on this i mean when we want to make sure that there is no delay or some language impairment and so on we know how to check that for monolingual children but we do not have norms for bilinguals and i mean as far as we know that development is different it's not only like to them so you know like the state of development that's possibly quite qualitative by as long as you don't have a bilingual understanding right yes we'll come back to that yes I think there's a great difference especially in preschool because kids here for example they come to school and they have a Catalan mother or Catalan but they can give a speech and so what quite often teachers think they're intelligent because they they have a outward outcome compared to others maybe they're they're shy even maybe in their context they're much shy that in the Catalan context they're shy the immigrant kids or the even Catalan it's so much related to them yes but i'm not sure what are you saying that this is a fair comparison or not i mean they compare even if they're not conscious about it but should but should we be doing that special because we don't actually have monolinguals right everyone speaks that's right that's true it's very i mean right no one finds that no that's true to our knowledge so when we have immigrant kids coming to school just like in class country and we observe how they acquire Catalan we compare them with other Catalan bilinguals there are no vast monolinguals there are no Catalan monolinguals it's interesting and then it becomes an odd issue it's interesting because it becomes an odd issue but is it really the same as comparing yeah no that's i hadn't thought of that that's true but is it really the same as saying taking somebody to Canada say Montreal who's from an Arabic speaking community because it's true in the English speaking community in Montreal all everybody the kids all now speak French and English when we do research on monolingual English kids you can't do it in Montreal because they all speak French not like native speakers but so but is that the same as so comparing them to English speakers in that context i'm not sure it's the same as comparing them to Catalan speakers too because both languages are official here they they get a lot of right yeah no it is yes right right that's right but i but i but i agree with Albert the fundamental issue to me because i had to fight over this people didn't want this it didn't want these studies being described in terms of how monolinguals are doing because they thought that's really supporting this monolingual bias and i think at a theoretical level that's true i think that there's a there's a it's a it's a bit of a dead end i think at this point in our in research to take the monolingual as the primary source of comparison but i also think that it's useful to do it to the extent that you're asking well is the process the same because we know a lot about monolingual acquisition and it would be interesting to know if bilinguals can do the same thing as monolinguals in the same way because then that's very informative right and if there are differences and this is really the crux of the issue the problem is if these immigrant kids are performing different from the monolinguals then that difference is usually interpreted as a deficit that yeah it's bad they're not the vocabulary smaller than the monolinguals they're they don't have the same grammatical complexity as monolinguals and that's always almost in deregulation say always often interpreted as something that's negative but let's take but maybe what you're right but what if you take the vocabulary example so say you find that these kids have a smaller vocabulary in English than monolingual kids why should they have to say vocabulary in English too right because the world is different and the language they're they're well in each language it might be smaller together it's actually larger yeah right right but that is it then then you interpret then it's how matter how you interpret that is that because they're they don't have the competence to do it or because the environment that they're living in right but often people right but people don't often take that next step that's the problem they say there's a deficit in their vocabulary or they're not doing as well in something else therefore there's a problem rather than saying the child needs more enrichment because this is important for them in both languages I agree with you that if these children are in a monolingual or a Catalan Spanish bilingual environment they need to be able to measure up and it's a matter of how you interpret those differences it seems to me I also think theoretically it's even less important to compare them to monolinguals at one level because if they're truly different then why would we expect them to look like monolinguals all the time no I mean we have to allow these theoretically for the possibility that kids who learn two languages from birth are actually different and that those differences are not necessarily bad I think in the real world it's often the case that it's not useful but in a theoretical way don't we have to open up to the possibility that this morning we saw that we are different right and I'll show you some data at the end right yeah anyway I just want to throw that out there because I because some of these some of this evidence has caused me to really reflect a lot on okay what do we do with this research how do we interpret this because more and more there's differences emerging and therefore the question is especially when you do research how many of you are kind of oriented towards more applied as opposed to academic research or not sure yet better start thinking about it you guys okay I mean when you work in applied settings educational settings clinical settings then this is a challenge because you you you don't want to be in a situation where you think you're saying things about these children which are actually pejorative or negative or you know that aren't fair I don't have any answers myself but when you tell parents oh either vocabulary is lower in English math everybody says oh gee we're going to stop using Polish at home because that's why they're not doing as well in English they would probably have smaller vocabularies in English even if only English was being it who knows anyway I just want to throw that out there because it seems to me in terms of the meta picture the field of language acquisition is now starting to turn maybe turn a corner where our notion of what it is that we're studying is changing and you can decide okay so I'm going to start off by looking at language discrimination speech perception and I'll I sort of what I'm going to do is show you and this is really a kind of a fast and simple overview of this research and there's probably a lot of questions that could be raised about what this really means or where to go from here so my interpretation may be a little bit overly general but I think it's fair enough given the state of the research but if you disagree let me know so as I said one of the challenges for children learning language I'm talking about monolingual children is how to sort out language from all of the other noise that they're hearing because you can imagine when children are born it's really noisy out there right and they're hearing language they're hearing cars or hearing the tv they're hearing the radio they're hearing all sorts of stuff so in order to learn language they really need to be able to sort out the signal from the noise as it were all right and research with monolingual children we've known this for some time shows that newborn monolingual children show within really within the days or hours of being born if you give them the option of listening to different languages and looking at which language they orient towards they prefer to listen to the mother's language and the mother's voice over another woman or another language are you all familiar with this research it's quite compelling so there's this attentional bias that children bring to learning language within hours of being born um sorry friend oh maybe it's from Canada we think 19 is it's about 19 also i'm generating a lot of energy up here so i'm getting hot and also a joc this is joc mailers research shows that monolingual neonates they can discriminate it's so they've been born they're really young you give them two unfamiliar languages and you do whatever you need to do to see what can they discriminate between these two languages or does it all just sound like one big language and the research shows that they can discriminate between unfamiliar languages which belong to different rhythmic classes languages have different rhythmic structures and there's a limited number of structures there's two very common ones three very common ones and if they're and if they're distinct if they belong to different rhythmic classes neonates can discriminate between these languages quite early in life so this is from a very general perspective this indicates that even before they're born in the womb children are processing language in a very very sophisticated way so they become in the first case of the first study they're becoming familiar with the rhythmic structure of the language the mother is using because they refer the mother's language over the father using the same language presumably because in the womb they're getting feedback about the mother's breathing the mother's body movements and so on and infants like familiar things so when you give them that choice when they're born they show preference for the mother's language and they show that they can discriminate between unfamiliar languages so this is really really important because it focuses their attention on the acoustic information that tells them what how the language that they're going to learn is structured okay so the question is what about bilingual learners because they've been exposed and this is the case of children that are raised from the very beginning in two languages do they show the same biases do they show the same tendency because if they do then this is really useful because it means that as soon as they're born these children and often potentially all children then have the capacity to attend to what's important in their environment when it comes to constructing this new language so this is a group of researchers there's two really big groups of researchers in this field some of them are here in Barcelona and the other group is Janet Worker's lab in University of British Columbia there are others Patricia Kool John Mailers to some extent anyway you're going to hear a lot about this this group as well so they were looking at uh they looked at English and I'm going to always give kind of tell you what these participants looked like because it's important to know these things in this research so these kids were young birth to five days of age these are really young children they were either English modeling of speakers this was in Vancouver English is the dominant language but lots of other languages and the other kids spoke either spoke Tagalog and English or they were learning Tagalog and English okay and these two languages have different stress patterns which we don't have to get into rhythmic classes one stress times and one is syllable time so they belong to different rhythmic classes so they're good to look at from that perspective I'll also describe some of these techniques to you because some of the because I don't work with some of these were interesting to me but also you may find them interesting and the shows you have sophisticated some of this research has become even with children this young so they recorded sentences um that were spoken in either English or Tagalog Tagalog Tagalog Tagalog okay good I can't I can't get it wrong anyone's acceptable that's what you're bilingual you opened all sorts of lessons um and these these sentences were matched for things like page duration and the number of syllables and they were low pass filtered just to preserve the arithmeticity they wanted to preserve only the acoustic information which which primarily distinguishes languages but also this is information that uh pre prenatal infants would actually hear fetuses actually hear this information in utero and they and they played them uh each of these sentence types uh for as long as and this was a non-nucleic sucking response so the children while they're listening to these uh sounds are sucking on a nipple that is attached to a device that measures the rate and the pressure of their sucking and this is really an indication of their attention because when children and this happens fairly quickly with young children infants their sucking rate decreases and the pressure decreases so they're familiarized with these stimuli until they uh as long as they maintain the 80 percent rate of sucking okay and when it falls below that then they go into the test phase so they become familiarized uh with the language these sentences and then you want to see whether they detect a shift in the language so uh in the test phase they hear alternating um uh samples of language between English and Tagalog so for uh in this 10 minute period they would alternate between sentences in English and in Tagalog and they determine how uh uh how much they actually responded to each of the of the shifts from English to Tagalog and back okay so it's a way of looking at whether the children are detecting the differences between the two languages um and this is one of the early techniques that was devised to look at preverbal infants so you're looking really at a level of language learning in children who are not able to actually express what they know about language. Let me see if i can interpret these carefully correctly for you. Here are the results um these are the English monolinguals so this is their uh their preference so what if this is no preference between English and Tagalog okay during the test phase so the best place to start is with the English monolinguals so their rate of uh responding is primarily when they hear the English stimuli so there's a lot of kids in this range this is the their rate of responding to the English stimuli during the test phase in the case of the the Tagalog monolingual children they're um they're showing no they're this this doesn't look this way to me but this is what the statistics show there's no preference there's no differential preference between the English and the Tagalog speech samples they're showing preference for both so they're so which is kind of what you'd expect given that that they have been uh exposed to both English and Tagalog pre prenatally and also for the few days after birth before they were tested uh interestingly the uh they had a bilingual Chinese group which also showed no preference so uh they're arguing that this is this effect is not simply because they were bilingual um but it's because they were they were uh they were exposed to Tagalog and English because they're showing this the split they're not showing a preference for Tagalog and and English because they're they haven't been exposed to two of them okay it's a control group but the important point here is that the monolinguals show preference for the language that they've heard and the bilinguals show in a sense of preference for the two languages they've been exposed to okay so this is important because uh this shows the same kind of biases that you find in monolingual children and it's a bias that really gets a leg out on the language acquisition process um so they showed this uh language uh there's a there's a follow-up study um that showed that kids can discriminate between these languages uh as well as monolingual children so it wasn't that they weren't showing a bias between Tagalog and English because they couldn't discriminate it's because they were showing this bias for both of them because they uh had been exposed to both of them okay and also if you look at neonates these kids they can distinguish between rhythmically similar languages at the same time as uh as monolingual children this is when monolingual children do it all right so the value of this is that um at this point we're not seeing differences between these bilingual kids and the monolingual children at least in this simple paradigm they're showing this own preference to the two languages that are in their environment and that are important to learn and they also show this I didn't go into that study but they show the ability to discriminate between uh rhythmically similar languages at the same age as monolinguals so the fact that they have this more challenging input environment hasn't actually altered these fundamental processes now um so that's that's a couple of ways simple foundational ways in which these dual language learners are similar to monolinguals but when you read this literature what you also see is that while that some of the the fundamentals are the same there are also differences between these kids so this goes kind of back to our earlier discussion that that sometimes the differences are interesting and and they're there and the issue is how do you interpret the differences and so I'll give you an example of a difference that emerges and this was an interesting one uh research and again on monolingual children has shown that monolingual children very young children can discriminate languages based on visual cues alone they don't know that so they'll show they'll show kids people speaking in different languages English or French or whatever but they mute them so they and then they they test them to see if they can discriminate between the visual cues do they know that somebody speaking English uh is now speaking French do they detect that they don't know that there's a shift from English or French but they know that there's a shift in how the mouth is moving okay and monolingual children do this and if and you must do this in in Barcelona this is a common bus time activity in Montreal you're sitting in a bus and you see people and you say these are interesting people what are they speaking and I mean this happens a lot to me in a bus I'm sort of thinking does this look like French does this look like English and it's based simply on the visual cues okay so monolingual children do this and the question is do bilingual children do this as well all right so this is a study that was published again in Janet Worker's lab and they looked at monolingual English children and they looked at bilingual French English children and these were children were fit four six and eight months of age so again here's the the procedure in the familiarization phase the children are listening to silent video clips of speakers producing a sentence in either English or in French they don't hear what's being said but they see the speaker producing English or they hear the speaker producing French okay and they in this case they're looking at they're using what's called the head turn preference procedure they're looking at the kid's visual gaze so there's a visual there's the visual in front of the child you're typically what happens is there's a bullseye that turns and that gets the kid's attention they're sitting in the chair in the mother's lap a mother is in this case probably blindfolded but also has headphones on so they can't influence what the child is doing they get the child's attention by moving the bullseye and then they will start with the visual of the person speaking English or French and they see how long they attend to that stimuli and they they they discontinue the familiarization phase when the child's looking time falls below 60 percent okay and then they do the other language so they're familiarized with the two languages in that way and then in the test phase the child either here here's the same speaker speaking the same language or the same speaker speaking a different language and they they look at the child's looking in order to see whether they notice the difference okay here's the results monolingual infants I always have it takes me a while to figure out these graphs so this is the monolingual infants and this is looking at the monolingual bilingual bilingual infants together and this is rather complicated what I want you to really focus on is over here because this is really that brings it all together during the test phase the pink data are the children at eight months of age the blue data are the children at six months of age so what you see is that what's critical in this is that when they're six months of age both the bilinguals and the monolinguals will during the test trial their attention is towards the discrepant trials they're detecting the difference all right when they're young and monolinguals do this up to a certain point usually six months of age and then they stop using the visual cues presumably because there's enough acoustic information or they just I don't think there's been an explanation for this but monolinguals don't continue to use this visual cue beyond six months of age probably because they don't need to because all they ever hear is English whereas and if you look at the eight month data which is in pink the bilinguals who are the dotted pink line they continue to use the visual cues and the monolinguals at eight months don't so what these data are showing is that the bilingual children in contrast to the monolingual children use these visual cues beyond the age when the monolinguals discontinued using it now the art so the argument they propose which distinguishes them from bilinguals is that this is an adaptive process for these kids because they're constantly having to distinguish between the languages that they're hearing and visual information in addition to the acoustic information and also presumably the speaker helps them to do that for the monolinguals the argument I guess is they don't need that information because all they ever hear is English right so I've I've cast this study in a slightly different way maybe than the experimenters themselves would but I'm sort of presenting this to you as an example and there will be others of how on the one hand while we're finding similarities between the bilingual kids and the monolingual kids there are also differences and most of these differences really speak to the adaptability or the flexibility of the dual language learners in adopting strategies or using strategies that are relevant to the complexities of the situation they're facing okay I mean in and out itself it's you know it doesn't mean a whole lot but within the context of the challenge of learning two languages it seems to me that's what it's telling us okay was there a comment over there no just the face so it would be me speaking English on one occasion or a me speaking French on another occasion so they're used to seeing me speaking English and then in the test trial they would they would see me they're not hearing it but they're seeing me speaking French by my lips and they they were tending because it was discrepant from what they'd seen before some of this stuff is quite clever actually I can say that because I didn't do it now this business of going back one study of distinguishing these prosodic these languages on the base of prosody is a worth commenting commenting on a little bit further and I won't get into this in detail because it gets rather complex but as I said languages have different prosodic structure rhythmic structures where there's a strong weak or a weak strong pattern in the in the rhythm of the language and it turns out that these these rhythmic patterns languages simply speaking can be divided into those that have the strong pattern versus a weak patterns strong versus weak and weak versus strong and that pattern in the rhythm of the language which is the overall shape of the sound of the language correlates with a group with the grammar of the language so languages are divided up not only by the rhythmic patterns they exhibit but also in terms of the grammatical structure so there are languages like English where you get there's the the function words of the language words like to and the and a and e and on these are these are words that are have no referential meaning you can't point to anything that tells you what the word that means whereas the clicker is a content word because you can point to somebody something that's a clicker right so in in general content words like this tend to get more stress and they tend to be longer in terms of their their duration of the signal and they also are more frequent in the language whereas these function words these sort of grammatical morphine kind of words they tend to be less frequent in the language they tend to be low stress and they tend to be a short duration and that's not a hundred percent the case but it's true of many languages certainly in English and French and English and Spanish and French the content words tend to be multi-syllabic they tend to get stress and they tend to be relatively frequent whereas the other words these function words have these other characteristics so what you can see I hope is that the the stress pattern of the language where you have strong weak the languages which have a strong pattern tend to be languages where the content words precede the function words so does anybody speak a language like that it's it took me a long time as a young student to figure this out in English and Spanish and French the function words precede the content words so we say to the store in the box a box a store right but there are other languages where the function words follow the content words so the pattern what so what this what this actually means is that the grammatical pattern of function plus content word maps on to the weak strong pattern of the rhythm of the language whereas the pattern strong weak maps on to the structure of content words and function words so what what's important about this is that when kids are learning there's a lot of things like this in language acquisition that kids will use information at a simple level the rhythmic information which they which they are exposed to in utero they use that information to try to figure out the grammatical structure of the language and it's called bootstrapping because they're using information from one domain of language to bootstrap them into a more complex level of the language all right so the reason i mentioned that is going back to this study the fact that these children are able to use the rhythmic characteristics of the input language or languages in the case of dual language is important not only because it tells us they go through the stage of development at the same time as monolingual children so there's no cost to being bilingual in that sense but they're also use they're able to use that information to actually begin to bootstrap into the grammar of the two languages so if they're learning two languages which have different grammatical more for syntactic syntactic patterns they're already beginning to learn that when they're learning the rhythmic characteristics of the language does that make sense so it's doubly important in a sense it shows us not only is the timing the same but in reaching this milestone at this level they're also beginning to work towards the next milestone which is to figure out how are the words organized in the language so there's more detail than i gave there but that's the general point i guess we're supposed to have a break is that right okay so we'll come back to this business of foundational things after we have a coffee so any comments or questions okay so yes so the question is somebody up there controlling this Albert okay the question is this business of children preferring to listen to the mother's language over a foreign language but also there's been research that's looked at their discrimination so it's not a matter of preference but they do they notice a difference between language a and language b both of which are unfamiliar okay and both monolingual and bilingual children make that discrimination fairly early on i think it's around four months of age if not earlier and there's no difference in when they do that if you give them language if they're if they're from different rhythmic classes if they're from the same rhythmic class like german and dutch or something they don't discriminate them at the same time they it takes a month or two longer before they can do it okay so it makes sense because if if what they're really responding to is the overall prosody of the language then if they have different stress patterns weak strong strong weak then that's fairly clear but it is interesting how that they can do this very early in development but if they both have strong weak or weak strong then it takes them longer to pick out the differences any other questions comments okay so here's another so we we see now that they can they prefer one the language that's addressed to them over other languages we see that they can discriminate these languages at the same time as monolinguals and we see that they use visual cues longer than monolinguals to as an additional presumably as additional source of information to detect which language is being used another major milestone in language development in monolinguals it's been widely studied is their perceptual discrimination in other words or phonetic representation when the best way to understand the challenge for children is to imagine that you're hearing a language that you've never heard and it all just sounds like garbled you can't differentiate one sound from the other you don't know when a word starts and when it stops and therefore it becomes essentially impossible to learn the language you know the notion that children can learn language from tv it's probably an overstatement because if all they hear is the language and there's and they don't associate any of that language with meaningful input it's probably virtually impossible for the children really to learn language from tv if it's just continuous speech because they can't pick out the individual sounds of the language so what's been widely documented in the case of monolingual children is that between about birth and six to eight months of age children can discriminate many contrasts in the language many phonetic contrasts in the language regardless of their language experiences so let me give you an example in english the sound r and l are they are different sounds they have different acoustic properties and they're also linguistically important because if you substitute the two sounds in a in a common frame it changes the meaning so rot and lot are different words and those two sounds have what's called a phonemic difference okay does ever do i need to explain this does everybody know the difference between phonemic and finish everybody know what a phonemic difference is okay so what happens is when kids are born up to about six to eight months of age they distinguish these kinds of contrasts for a wide set of sounds regardless of their specific language experiences and then around six to eight months of age and certainly by the end of the first year of life they can discriminate the sounds that are phonemic in their native language the input language but they have difficulty or cannot discriminate sounds that are no longer phonemic does that make sense so japanese children can distinguish r from l up to six or eight months of age but after eight months of age you give them that contrast they don't discriminate the difference because in japanese changing r and l sounds don't make a difference okay so this in itself is very interesting from a dual language learning point of view because what it means is that when children are born they're born with the capacity to distinguish a lot of the acoustic differences or contrasts phonetic contrasts in the in any language so there's no so there's no limitation on their language ability due to that discriminability ability does that make sense but the question arises is what happens with respect to this what's called perceptual attunement that occurs between six and eight months of age do dual language learners show the same trajectory in this development so there's been some research done in this is 2008 by burns analog again a group from janet workers group and this is the contrast between ba and pa so i won't get into this in detail but the difference between the ba sound and the pa sound is a matter of when the the voice onset time the difference between the when the vocal folds open and the release of the air and i won't get into it but if you're a french speaker if the if there's a delay in this release that occurs around here on this continuum then you no longer hear ba you now hear pa in english uh the you you you the distinction between ba and pa occurs here all right so this is this range of sounds of sounds that are ambiguous with respect to whether they're a french pa or an english ba so and and this is a function of the vot so the question is when do children learn this distinction and because it's different in english and french the question is if you have english french children do they show the same pattern of distinction for this as monolingual children you might think that would be different again on the simple logic that this is a more complex problem space for them they're getting two sets of sounds and maybe it's going to take longer for them to sort it out and this is really important because they they have to begin to form an inventory repertoire of what are the sounds that are important for learning learning the words of english because these sounds are important in learning individual words in these two languages and it's important for sounding like a native speaker so this was a study that was done with monolingual english speaking children and with bilingual french english children again here's these are kids who are six to eight months of age 10 to 12 months of age and 14 to 20 months of age and this is six to eight months of age is the age around when it's still a universal ability that children have that's independent of language experience and then here these two ages when it is actually becomes more language specific and what this is really telling us is this is we're now seeing the influence of language experience on these kids perceptual abilities so the technique here is that they they hear in the in the familiarization phase excuse me they hear this the sound that occurs at this point along the vot continue so it's ambiguous as to whether it's a french pa or an english ba okay so for the for the english speaking children in other words this is not clearly one or the other or for the english speaking children this is going to be here heard as a a ba as opposed to a pa for the bilingual kids it could be here heard as either one depending upon whether they think it's english or french now the bottom line in all this is when they are going to cut cut to the quick on this because this is a rather complex but the the bottom line on this is that in this phase of six to eight months of age when it's a kind of universal pattern the bilinguals and the english kids and the bilingual children are showing the same pattern so they they they're disciplined i'm sorry this is going to get confusing they dis a bit they when they hear this sound as a test when they hear this they hear it as a ba but when they hear this they don't hear it as a pa so the french boundary is sufficient for both groups to distinguish between these two sounds as different phonics the argument is that this is a salient for some reason rather they don't know why this is a salient difference in vot and that both groups can hear it whether they're bilingual or monomagnol so it's really the the the universal stage of their development when they when we get into the 10 to 12 months of age the question is what kind of what kind of pattern do they show and they're they're they're showing recovery to both the the the ba pa sound and the ba sound the bottom line here is that the bilingual kids are showing the same attunement pattern as the monomagnol kids so despite the fact that they're hearing two sets of input with two different vot times for this discrimination the bilingual kids are able to do it just in the same way at both ages as the as the monomagnol kids so it's not slowing them down so that their ability to to create these phonemic inventories is basically the same for the two groups okay now that it's not necessarily the case when for all contrasts so this is a study that was done here and it's looking at the distinction between uh a katalon sounds so this is a contrast between two sounds that are um phonemically different in katalon but not in spanish because the spanish sound falls in the middle and it incorporates those two if you present bilingual kids with this contrast what you find is that they actually fail to make this discrimination as early as the monomagnol kids and it takes them a month or two longer to make this discrimination and that's arguably because the spanish sound that they're learning incorporates this distinction and it takes them longer to separate them so the point the reason why I present this study is that um that there are differences here so if you've got contrast phonemic contrast in the language which are phonetically fairly straightforward then bilingual children are able to do this at more or less the same age as the monomagnol children however if they're learning two languages such as spanish and katalon where there are contrasts which are ambiguous so it's a contrast that's important in one language but not in the other language then it may take them longer to sort it out because it is ambiguous um so what's important about this is that there are again there are differences between the bilinguals and the monomagnols but in this particular case the difference doesn't reflect a sort of lack of capacity on their part per se but it reflects some of the complexities of learning two languages which may have in this case acoustic properties that are ambiguous in the two languages right so i because i i i want to be careful in emphasizing that dual language acquisition may be very similar in some respects as a monolingual acquisition but there's also respects in which it's different and and it's a matter of how you interpret these differences that is important so in this case the interpretation is that this is a hard contrast to make because it's important in one language and not in the other and it takes them longer to recognize when it's important in the other language so i won't go into that because so this is just a case where they they need more time to learn this stuff okay so up to this point we've seen that kids show this bias to to listen to certain languages they show the ability to discriminate various languages they started to create these phonemic inventories in their two languages and there's important ways in which they're the same so the next stage in language development is to start to learn individual words because it's not enough just to learn what the rhythm of the language is and what the sounds are to make up the language you want to start putting those sounds together to create words now in saying that this is the next stage i don't mean to imply that language acquisition is a simple process of discrimination phonetic representation and word learning and then grammar these things often are co-occurring but their progress is emerging earlier in some cases than in others so the as i've mentioned a couple times now one of the challenges in learning a language is how does not only form say what this the sounds of the language are but what constitutes a word in that language so this is a study that again was done here that is looking at children's ability to segment words from continuous speech and this is a very difficult process because there's no clear boundaries in continuous speech between when a word begins and when that word ends so when children are hearing the input languages they're really hearing continuous speech which really means is that there's no there's no clear it's like beads on a string and you don't know where the where a bead indicates the beginning and end of the word okay you have to figure that out using other properties of the input so this was a study that was done fairly recently six and eight month old Spanish Catalan bilinguals and monolingual infants of each language so in the in the familiarization phase the children are given sentences with pseudo words in either Spanish or Catalan so so these this is continuous speech but embedded in this is words that they've never heard before and in the test phase they simply wanted to see did the did the bilingual where the bilingual children able to recognize in the test phase the pseudo words that have been presented in continuous speech during the familiarization phase because that would give us indication that they're able to segment words from continuous speech in the same way as monolingual children and that they were all able to do it at six months of age so again they were better in eight months of age but they were all able to do it at six months so again it's really rather dramatic that these milestones are emerging in pretty much the same time even though these kids are getting much more input complex input at least in two languages so looking at word learning is also about learning the meanings of words not just kind of produce words but you have to know what the south the words mean and one way in which is sought the children learn words this is a big topic of much discussion research is that you associate the meanings of new words with new objects in the environment so in an experimental paradigm if you give children the new words to learn they're likely to associate those new words with objects that they've never heard before and this is just called associated learning excuse me because you associate the new word with a new object it's a simple matter of association there's other theories of word learning but this is one so in this study which is called the the switch task again it's by the group from vancouver they use what's called the switch task there's a habituation phase where they hear two pseudo words lift and meme so these are words that differ from one another with respect to every sound in the word the initial the medium and the final sounds and these words are associated with novel objects so a so one might be associated with this kind of thing and the other word might be associated with another object so there's an association between the new word and a novel object in the in the test phase they're even they're either presented with the same trials where the novel word is associated with the same novel object as during the test phase or they're presented with a switch trial where they're presented with an object which has the other object's name so you're switching labels for objects and that's different from what they heard during the test phase so and the the logic of these studies is if they look longer at the switch trials versus the same trials then it's indicating that they've learned the association well precisely what the the challenge in a lot of this research is knowing exactly how we interpret these results but at a minimum what you can say is that if there's deferential looking towards one condition as opposed to the other condition then minimally you can say that they've detected a difference between the familiarization phase and the test phase now the the the hypothesis that they were testing out here is that bilinguals might be slower at this task because they uh they must they have to learn more words and more sounds so there's more associations to learn and given that we know that the vocabulary size in each language is smaller maybe there that's indicating that they're just not as good as this on the other hand it could be that they're equally good as monolinguals because overall their conceptual vocabulary is the same so overall the vocabulary learning is just as extensive as that of monolinguals in which case you might not expect a difference but if it's really their experience in each language that matters you might expect them to be inferior and what the result showed is that neither group showed learning at 12 months they weren't able to expect this difference at 12 months but both groups showed it at 14 months again no difference so these research these studies are interesting especially if you're a graduate student and you're trying to do publications it's generally hard to publish null results so one of the controversies that Juman was referring to in this cognitive advantage field is that the research was coming out of Bialystok's lab and other labs that was reporting this cognitive advantage and the people there are a lot of people doing similar research who could not find these advantages and when they would submit it for publication they were arguing that there was not a bilingual bias there was a publication bias and that for some reason or other the journals were only favoring the studies that got the positive effects I don't know why they but the problem is in all research that's true it's very hard to publish the results of your research if you get null effects because there's a lot of reasons why you might get no difference between one group or another could be you did a bad study you're re-selected the stimuli or something wrong but this is I mentioned this this is relevant to you as students as a null effects are interesting if they tell you something so if you expect that there might be a difference so in this case you expect the bilinguals might do less well and you don't get that difference then it's important and you can get a published so if you're doing something where you think that there might be null effects you have to have a good argument for thinking why you might get an effect so that the null effects are informative now I'm just going to extend this study so we see that they can do this as associate of learning paradigm as well as modeling yields but it's not true in all cases so the the study I just talked to you about used pairs of words that had multiple points of difference but in in in this other study by chris fennel and his colleagues they looked at the ability to form associations between novel words and objects using minimal pairs doesn't even know a minimal pair these are words that differ in only one sound so bed and death everything's different except the initial sound whereas the other one it varied in all three sounds and in this study again there's bilinguals and there's monolinguals and they tested them at 14 17 and 20 months of age and what's important here is that the monolate the monolinguals were able to associate these novel words with novel objects at 17 months of age but the bilinguals were not there's no difference here so but by eight by 20 months of age bilinguals were able to do this so again there are differences where the where the learning conditions are more complex and where there's not enough information in a sense then it may take bilinguals longer to do it so again there's this interesting contrast between what's similar and what's different now there's an interesting sequel to this story and that is that on the one hand the argument could be that this is a more complex task and the bilinguals need more time to do it because it's complex it could be that they take them longer because they had reduced exposure in each language but there's another explanation and this was motivated the study by Fennel Byers-Highline and another group where they argued that maybe the testing conditions were not appropriate that the stimuli that the but the stimuli that were presented to both groups of learners were produced by monolinguals but the language input that bilinguals get is often produced by bilinguals and so the kind of if you like phonetic landscape that bilingual children are getting is not the same as what monolingual children get so bilingual children might be hearing novel sounds from people who are bilingual but they may also be hearing novel sounds from people in a language that they're not totally proficient in so when they go to produce the English words and if they're francophone then they may not sound the same way as they would if they were produced by an anglophone so that the the input is more variable and therefore it takes them longer to stabilize the sounds that are phonemic in their language so they did an interesting study where they they used the same paradigm as before they gave them this associative learning task and but this time instead of presenting them stimulus stimuli that were produced by monolinguals they gave them stimuli that were produced by French English bilinguals and you see some of these sounds might that might be phonetically different in the two languages even though to our ear they sound the same but they may have different voice onset time distinctions in them and what they found was that bilinguals could do this at 17 months of age if they were using stimuli spoken by the by by bilinguals and the monolinguals could not do it at 17 months of age when they listened to the stimuli produced by the bilinguals so that's really interesting I mean it it indicates that when you're doing this kind of testing with kids you have to be very sensitive to the kinds of environment in which these kids are learning the phenomenon that you're studying because the input that they're getting or the exposure that they're having is not simply the sum of two monolingual inputs there actually is a difference in the input that they're getting and it's actually in a sense of testimonial to how sensitive they are to the linguistic landscape or the acoustic landscape because it's the difference between being able to do it and not being able to do it and what's interesting here is that this study was replicated by another group in another lab so it's always important that this stuff get replicated now one final piece here and then I'll move on to the final set and again this is an area that shows that bilinguals are adaptive and really creative in the kinds of strategies that they use to acquire the language and so there is everybody know what this phenomenon of a mutual exclusivity is who does not know what this is okay I'll just do a very simple thing and most of you know so mutual exclusivity is postulated to be a strategy that monolinguals use to acquire the meanings of new words and it runs basically that if you hear an unfamiliar word you are associated with an object or a concept for which you do not have a label yet so you don't assign a new name to something that already has a name so if you know that that's called glass then I say there you're not going to associate that word or I use a pseudo word a pseudo sound in English you're not going to give it two names because it already has a name and this has been documented in to occur at monolingual children when you look at bilingual children it does not apply as what as much then you do it to some extent but they're more than willing to learn more than one label for the same object because in when you're bilingual this is what language is all about but the same object can have more than one label I mean you wouldn't think anything of this but monolingual children actually avoid doing this so it's a way in which bilingual children differ from monolingual children but it's evidence that they are adaptive of the kinds of strategies they use in order to deal with the realities of dual language learning okay so just to sum up and then I'll move on to the final section so there's clearly similarities in what goes on with these kids which is really really important to to know in the in a sense in the abstract in the sense that given the right learning conditions dual language learners show similarities and some really basically fundamental aspects of language development so in terms of language discrimination preference phonological fine-tuning associate of word learning these are there's this is not everything there is to learn but these are some key ones but at the same time there are these differences and to a large extent the differences reflect differences in the learning environment differences in the complexity of learning two different languages which often have characteristics which are either different or in fact may be ambiguous and just differences in how the two languages work so any comments on that I think that what has not happened in this area of research which probably needs to be done is there's been very little research looking at the factors that influence the factors that influence these developmental patterns so in contrast to some other areas of language acquisition research people have not looked at SES very much they haven't looked at the long the language differences very much they haven't looked at inconsistencies of the language input so I think there's much more room for looking at a variation in these patterns and at the moment my my perception of this research is that a lot of it has been done in order to address this issue of whether children have the capacity for dual-language learning or whether it exceeds their capacity and therefore should be evidence in a sense deviations from this pattern but so far the evidence is saying look there are many many similarities and there are certain differences but these differences don't reflect a lack of capacity they really reflect adaptability or flexible now one of them just really jump ahead and describe two studies one that's we've done in one that was done in sweden that looks at this business of dual-language learning into the future of these learners and this is really um this is really done in the context of the critical caring hypothesis and I won't go into this detail but this is all the notion that was promulgated some time ago by this guy in his geese right Conrad Lorenz who noticed that at a certain stage of development these young geese would imprint on certain objects in most cases it's the mother goose but they would also print on a moving block of wood if that block of wood was presented to them at a certain critical period in their development right now in most cases the environment of animals lines up with things that are functional for them so for these geese it's functional to imprint on the mother goose because it's the mother that's going to take care of them and keep them out of harm's way and so on but this is a picture of Conrad Lorenz and these geese have imprinted on him because he was the first moving object that they encountered at this point in their early point in their development okay so this this business of periods in development when you're particularly sensitive to certain kinds of stimuli is very important because it means that the behaviors that you learn or the cognitive alterations that you go through at that point become hard-wired and that's really important for learning because it means that it's going to be hard to change the that higher wiring by later experiences you don't want a system an organic system like this to be changing whenever the environment changes you want it to be adaptable to those changes but you don't want it to really totally restructure itself so in the case of language learning it's useful for kids to become fine tuned to the the phonetics contrast in their native language because then they can use that inventory of sounds to build words and grammar and so on you don't want them to be changing that inventory whenever they're incidentally exposed to different sounds in different languages okay now this was a this was a concept that was applied by wilder penfield from the mongrel neurological institute it to the case of first and second language learning and and the other fellow of course many of you know about is eric menenberg who argued that language is also subject to these critical period effects and it was generally thought that these effects occurred around 12, 13, 14, 15 years of age and the idea was that up until that stage in development your neuro cognitive system is sufficiently plastic that certain kinds of learning experiences will change the structure of your brain of the knowledge that you're coding in your brain and then after that period it's hard those kinds of in fact those kinds of inputs will not change learning in any significant way okay so most people yeah well that's what i'm going to look at that's the real question here is can you acquire a second language completely or like a native monolingual speaker at the same extent if it begins after this critical period such as the person you knew and i'm going to show you some data from sweden that suggests that it's hardly ever exactly the same but it depends on how you look at it so that's the issue and then i'll talk about a study that we did with internationally adopted children which also addresses this issue so everybody's kind of familiar with that now what people have done in this history this is an incredibly complex area of research from a common sense point of view and from a general empirical point of view there's no doubt that the older you are when you begin to acquire a second language the less likely you will achieve a level of performance that is comparable to that of native speakers and that's pretty well attested but what we but what what has seldom been looked at is how early these these critical period or age effects occur and this research that i'm going to present to you that was done in sweden in our own research really looks at the lower limits of this and raises i think something interesting question and for our purposes what it does is it raises the question is can children who learn additional languages that don't occur when that doesn't occur from birth can they ever be identical to monolinguals yes uh right you're saying that i'm contradicting myself and you're right in a way um but i think you'll see i think actually it does matter despite what i said before and in response to your earlier comment it does matter in the lives of many children whether they're different from monolinguals so because in the school system for example in clinical settings differences from monolinguals often lead to decisions about what to do with these children so uh so i don't think that those differences are irrelevant i think they are what i think is the question is how do we interpret those differences and what do they mean from a theoretical point of view so that so this is sort of asked asks a question that in some sense is contradicting what i said earlier but i think it leads to some interesting results which helps us to shed light on language acquisition that are very very much competent in the life of the star learning right so so competent that if they are not monolinguals well let's come back to that when you see the data and we'll see whether it's a question that does make that comparison provide any additional insights and then we can say well then why bother i mean in general i have kind of agreement i don't care i mean i get into this all a lot of time with parents when they're trying to decide whether to send their kids to a dual language program a bilingual program or some sort and they're they're worried that if they study through the medium of another language that their english vocabulary is going to suffer a little and my attitude is okay because in the process they acquire another language and that opens up far more doors than having a smaller vocabulary closes this is this but but these findings from this research raise a lot of questions in my mind about what to do about these differences because they're there and i'm not quite sure what to do it myself i'll throw them out and you can tell me what you do yes and i think that we always think about continuous bilingual development watch this hadn't been sick at this disruptive and you lose your class right i think that's quite different so some of the research i'll present deals with that loss in the first language yeah and i want to tell you that one thing not enough because through the country because i haven't found the person who first of all who learned the second language was quite late something got to work and became better than those the native speakers the language and the example is i like to think to my colleague is Adelaide Makhit who is a russian writer immigrated to France against drug learned territory in russia but immigrated to France around these countries and he was the youngest person in the academy in his writing right much to the horror of francophones well they accepted him so i didn't know but for instance if you mean pronunciation yes he definitely has a foreign accent when he speaks french but when he suggests a pronunciation but the quality of the language and the writing is better than most of his books well in english people use the example of Joseph Conrad whose whose native language is polish but then he wrote in english and he is brilliant absolutely brilliant yeah and but your point is an important one depends on what you look at and one of these studies makes that point so whether it matters or not go back to elbert's point also depends on well what are we looking at against a living way and also south indians because south indians are better than russian languages and they are so i don't know how people believe in this issue because it's a truth to me it is and i also love to speak the way of the english language right no it is a very i mean it's it varies by context in montreal for example because english and french even though french is the official language but because english is so important in the general sense that you often in the in public domains and on the media you often have people who say in the english media they speak english as a second language and it's very clear that it's their second language they're not the best second language speakers but people they don't care they just listen you know it could be a weather report could be a news report nobody cares but in other areas i've been in if the speakers on the media don't speak flawlessly like a native speaker they don't believe what the person is saying because they think if you can't speak english properly then i don't i don't believe what you're saying or they they see them in a more negative light so these things do matter but but in montreal everybody at some point has to speak a second language so we've become very tolerant of what it means to be a second language in monolingual environments i think they're often less tolerant because it's not an important issue in their community or they think it's not so this is one how many people know this this is one of my favorite studies abraham abrahamson and kenneth hilton stamm at the university of sweden in the Stockholm or university of Stockholm in sweden and this is a that's what happens when you're bilingual so this was quite a complex study and he really he really set out to look at the critical period of hypothesis in the traditional way by looking at groups of people who had begun to acquire swedish as a second language at different ages so he had and but he did some really really really interesting methodological things and that's why i like the study so he had 195 speakers of swedish as an l2 and he went through an extensive screening process to identify these people so in order to be considered for this study they ran ads and did all sorts of stuff throughout Stockholm for a long time and they only identified people initially for initial inclusion if they self-identified as having native life competence in swedish and then they actually uh had subjected them to a variety of tests including judgments by native speakers of swedish and and said and they had a highly proficient and less proficient l2 speakers and they said to the judges is this a native speaker or not and so on so they they only worked with the people who were self-identified and other identified as very very proficient in in swedish so right away we've got people like a number of you referring to people who were very very competent in the second language the first language of this group was Spanish which is useful to have that control most of them on average were 19 years of age at the time of testing so they had they were not children they had lived in sweden on average for 10 plus years so they had lots of exposure and they spoke the swedish to this characteristic of Stockholm all right then he broke this group of 195 people into subgroups based on the age of onset so this is also an interesting methodological issue i don't know if any of you have tried to work with this construct the baby onset but it's very can be very hard to know when people begin to learn a second language because it just doesn't one day you wake up and you're starting to learn swedish eventually it's often a gradual process and don't know for sure anyway he subdivided them into people who had acquired swedish as a second language before 12 years of age there were 107 in this group and 88 who began to learn swedish as an after 12 years of age they in other respects were quite comparable to one another now he what he wanted to do was he wanted to assess their competence in swedish in a number of different ways so it gets at the question that he raised well may whether or not you're as good as the native speaker may depend upon how you assess people and that's really what he set out to do all 95 of these people had been judged as i said by native speakers of swedish as having native like competence they were judged by 10 native speakers so this was a very rigorous assessment i won't get into the other stuff and then he assessed whether they had they had were perceived to have native like competence so they were all judged in general to be native like but then they gave them this more more carefully calibrated score whether they were perceived to be native like from zero to 10 and they were analog and those those that subgroup was then looked at in more detail depending upon how old they were whether they are less than five years in the country they began to learn the language at less than five years of age between seven six and seven eleven years of age between 12 and 17 years of age or older and then at 24 months 24 years of age okay getting tired sorry so here's some general statistics about these this isn't the most interesting thing but of the early learners 62% were perceived as L1 speakers and 6% were absolutely not so there was a there was a tendency for the younger learners to be perceived as native like but of the later learners and I forget offhand actually how he subdivided that I'm sorry only 6% were perceived as native speakers and 58% were perceived as L1 by one one judge or no judge so there was this clear age continuum that the younger was generally better than older okay now here's the data that are most interesting this he he subjected these these people who were judged to be overall native like to a very extensive battery of tests in fact there were tests there were 10 tests that they were subjected to and these were very very demanding tests of language proficiency in swedish they were things like they did a BLT thing if you're familiar with BLT it's looking at whether their voice on set time boundaries in swedish were the same as native speakers in swedish this is a very fine-tuned test but they were also given tests such as interpreting proverbs you know haste makes waste this is the kind of thing that is difficult to pick up as a as a native speaker they were also one of the ones that actually was particularly useful was they were asked to listen and understand swedish with noise in the background it's very hard for even proficient native speakers to comprehend the second language in a noisy environment so the point here was that all of these were tests that were difficult even for native speakers all right whereas in many studies of the critical pure hypothesis the way they assessed them is to use tests which many native speakers score perfect on on the assumption that native speakers should do well on these tests therefore those are the ones we're going to use to test that hypothesis but we all know that native speakers actually do well on everything so here's the so what you have here is these are the number of subjects sorry this is very complicated well it's quite simple in a little final analysis but he was scrupulous in presenting his data this is the age at which they had acquired in the second began to acquire swedish early childhood from zero to five years of age late childhood from six to 11 years of age adolescents early adulthood and later adult okay and all of these numbers here these are the bands that define these age ranges all of these these are numbers and the number of people who scored native like on seven of the tests or on six of the tests or on three of the tests who had acquired the language in early childhood is that can you get that it's important that you understand that took me a long time to get this so and then he does that for every age group all right so now what you can see clearly just from the step down is that the older the age of that onset of acquisition the fewer the people you have that scored native like on more than one or two tests right because further you get out along here the only people who are scoring native like are only scoring native like on zero one zero one or two tests okay yeah it is yeah that's right if it's not clear please tell me that so so these are the focus on this group to start these are all the people who learned Swedish as a second language starting between birth and five years of age right mainly three in five and this is the number of tests that they scored native like on one two so the people who there was one person there was one person who began learning Swedish as a second language before he was one year of age and he actually scored there were three of people like that who scored native like on ten of the tests yes you're you're getting ahead of me I think yeah but but but let's just look at the overall shape of it to start does everybody get that so as you the older you get the fewer and fewer people there are who scored native like on a lot of the tests and that's kind of what we all believe so that's not really surprising he's just presented it in my new detail and also using some very difficult tasks yes by self grading and then by these native speaker judges I can't I think he had a combination of those but I honestly can't remember in fact I think he used the scale that I use in another study where they were asked to judge they were given native speaker samples and these speaker samples and they were asked to judge them on the vocabulary grammatical competence comprehensibility and a whole bunch of stuff so native like was that they scored did I have that well if they if they scored like native speakers on on four out of the five then they were considered native like so there was a cutoff not all not all the native speakers scored at the top left right four to five scales so the scales were grammar vocabulary pronunciation comprehensibility and fluency maybe so if you scored uh and the scale it was karmic it say the scale was nine points if you scored eight or nine on four of those then you were considered native like it was very rigorous I mean this isn't really so these are people who like you were mentioning for all intents and purposes they sound like native speakers in swedish that's the important thing these are people who can pass as native speakers just based on what you're hearing right and what you can see is even using that very simple measure very few people who began to learn the second language after 12 years of age scored in that range if you take five as a cutoff there's a few but not very many whereas most of the kids who learned the people who learned it in early childhood are in the category of six or above so that's the general that's our general notion of the critical period yes right did everybody hear that one of the complications at this point is that the method or the mode of acquisition and learning would be quite different for these learners compared to me hang on to that because you'll see that that's true and it should favor the early learners but you'll see that's what's interesting about it but come back to that after I show you the next slide yes sorry in advance it may come up later on your presentation the basic argument from the people who do language acquisition and in adulthood against the critical period offices will be that the amount of input is never going to be the same a child right currently socializes with other kids and is highly dependent on an adult be it school of preschool environment whereas those who come to a very critical work may be limited to yes sir goodbye sir right and that undermines well right did everybody hear they come so that if you're young when you start to learn the language you've had lots of exposure whereas the older you get the less so I agree if you start to look at these people it's not very interesting but if you look at these groups in these groups some of these people have been exposed to swedish for 30 years there was a wide variation in the age at testing and even though they might have started late they were also quite old on so but again hang on to that because the interesting thing is in how poorly the young learners did so I agree with all these are very insightful comments you've been well trained as great but the interesting thing is what we're not expecting but it but if you still are bothered by that come back to it okay so that's what he called perceived nativeness so these are people and by that he means these are people who were scoring they were perceived as native-like speakers and then they their performance on these tests was looked at then he looked at what he calls scrutinized native likeness he retained all of the the subjects with a perceived nativeness score of sex or above which means they were perceived by the majority of the L1 speakers a swedish to be native-like speakers 81% were early learners and 19 related learners so he's whittled it down and he reduced it to 41 of the original 192 in order to control for factors like the range in the age of testing length of residence and their amount of exposure so he wanted to reduce some of the variation that you're pointing out is very problematic when you look at a sample that varies so much in age of acquisition and he had 15 swedish L1 speakers I mean it's really interesting if you're going to do this research to include native speakers of the language because it's surprising how often they don't score like native speakers so that's why that concept is also problematic and then he gave them this very expensive battery of tests I won't go into this but there was a grammatically judgment task you all know what that is I assume lexical interference I forget actually what that looks like semantic formulaic language idioms proverbs perception of word perception in babble babble noise speech perception these were all difficult tasks so here is the interesting slide for me so this is the this is among this very select group of participants how well they do it and all of these measures so we're not looking at the general shape we're looking at actual numbers so this the line I want you to focus on is really nine and ten so these numbers refer to in the number of individual not the number but the individuals who scored in the native speaker range which is not perfect but it's the same range as the native speaker scored on nine or ten of the tasks the group that's really interesting to me is the group that acquired starting to acquire swedish as a native language before five years of age only three of them out of 15 in this group scored in the l1 range on nine or ten of these does this surprise you what does this suggest to you are you shocked or are you just tired and you want to go home are you do you understand what I've done so what I've done is circle sorry so these are the people who scored uh let's look at this red sort the group within the range there were only three people out of 31 in the childhood group who scored native like on all 10 tasks three out of 15 in the really early group scored nine or ten so this is a lower selection criteria but in either case this is not a lot of people because this is the group that should be most likely to score native like so these kids so this is why we're saying your questions are good ones but uh the the thing is that these are the kids who would have had the most exposure to swedish they would have had exposure to swedish in schooling in their general lives in all sorts of ways they're still in many cases using the native language but it's a second language this is really shocking no no I'm sorry could you repeat no but we're not talking about excellent or not we're just saying they're they're only doing as well as the native speaker even though the native speakers are not excellent yeah yes right yes that's what that's what this was about so they included where did I have that slide yes they included 15 swedish everyone participants who are matched to the altitude speakers on age and education but I understand I mean it's so disturbing it's so disturbing you think oh there must be something wrong with what he's done now for somebody like me who's really been an advocate for early bilingualism and early immersion and all this stuff this is really shocking I'm you know what I'm going to do one more question because I want to talk about some of my own research because I think this is so who wants the last word before we move on I've talked about everybody else's research I wanted to say my own right right perfectly perfectly but if they're in a school system or then they're being tested in rigorous ways so yeah it's a question of what do we make of these results how do we interpret them one more and then I go move on but you can't get much better than this I mean before five years of age they were schooled in Sweden but let me do my stuff okay so we've I've for a number of years this is research that the I'm not well known for but we've been looking at children who are internationally adopted children from China does anybody know this research no this is really interesting right I got interested I got interested in these children because I wanted to look at what is the full what is the other limits on the capacity of young learners to acquire an additional language and these children are particularly interesting because they were born in China they were raised in China for you'll see for between 12 and 24 months of age they were then given up for adoption because of the country's one-child policy so the main reasons for adoption were not as it is in other countries not related to drug abuse in the families of the socioeconomic disadvantage inability to cope with they they didn't have those disadvantages conditions and in fact children who are given up for adoption in China are often raised and reasonably good or relatively good foster homes or institutions because the problem is if you've got children are raised in institutions which have poor quality then the children are going to suffer from developmental problems related to inadequate raising do you all know the Romanian adoption stuff I mean these were horrific conditions these children it was really quite good so for our purposes what's important is these two children were adopted by French-speaking parents in Quebec they usually were started to learn French between 12 and 24 months of age because that's when they were given up for for adoption and the acquisition of the ink Chinese stopped at that point so that's an interesting thing that I want to back to French is the only language that they learned subsequent to adoption and so they're essentially modeling your French speakers they have the so what the interesting issues is the normal neurocognitive substrates for L2 learning may have been altered because there's growing research that suggests that the first language is very strong offered very strong support for second language acquisition so the question is does losing a first language within the first 12 months of life have a consequence one way or the other for subsequent language learning I think if I so I'm going to ask most of you I've sort of given the way now how many of you think these children will look just like native speakers of French because they started to learn French for their child I should have asked this before Albert's one okay a number of you how many of you don't think so how many of you are not sure well okay we went in thinking they would look like native speakers right but the other issue the other thing that really drew us to this sample is not so much the attrition of the the home the birth language but I was really interested in these children because there's a delay in the acquisition of the first language and it's a very short delay but it's a delay nevertheless but it's a and it's a delay that has many advantages to it now often these children are thought to be at risk for development in general but language learning as well and this is this is a concern that motivated a lot of the research that was done before we started our research so one of the problems may be in general that the pre-adoption environment may be impoverished socially cognitively or linguistically we don't think that's true in these kids case but I'll show you why another risk factor made as I mentioned they lost the first language and maybe that undermines the foundations for subsequent language learning the other possibility which you would probably not predict is delayed onset of the second language is also a risk factor and part of the reason for thinking that is the study I just talked to you about right so we we set out to do that now at the same time these children have lots of advantages they're exposed only to the second language I'm like most second language these kids sometimes their new language is referred to as second first language I mean I think that's a really good term because it indicates that these kids are as close to a first language known as you can get when the first language is not a first language they also are raised in enriched learning environments because parents who adopt children are often fairly affluent because it's sensitive to adopt children they're also very conscientious parents because it's a lot of work to adopt children these children are also potentially at an advantage because they're from China the pre-adoption environments are relatively good and I'll show you evidence for that and they were mainly girls and arguably that's an advantage in the early stages at least and also they're within the classic critical period I mean I think anybody who thought about this for two seconds would think these kids have an advantage because they're so young now I'm not going to go into this in great detail but a lot of other research had looked at these kids but it was done mainly by clinicians and mainly in order to establish if these kids were at risk for their language development there's also been research on their social development but often what this research did was it looked at these children's performance on standardized tests in English in most cases to see whether these kids scored in the normal in the typical range on English and when that occurred the problem is that it did not take into account the fact that these children are actually raised in enriched language environments because these tests do not take into account that their parents are actually from upper middle class so from a theoretical point of view we're missing an important factor here right I mean from a practical point of view you might say who cares but from a language learning theory point of view you would say yourself yeah but are they learning as much as English or French in this context as children raised in similar environments or is there a limit and these are rather indirect measures of language proficiency now we also use them but that is a problem so so that we look at these children for many years actually we wanted to look is their acquisition of the adoption language like the L1 or is it like an L2 do they have the same level of competence as non-adopted children or do they show the early effects of delaying which is really what drove my interest in them and if there are early effects what age effects what causes those effects so we've done a lot of studies we initially did research when these kids I think my when they were four years four years two months old and then up all the way up to 13 years seven months of age and I'm going to describe so the very first one the kids were we did test them twice the first time they were 41 months of age to 56 months of age and then we did a follow-up the year later because we were so surprised at our results we wanted to test the generalizability what we found and they were matched to control kids there are 25 in each group 24 they were matched to the control group on age gender and importantly socioeconomic status so this is a really interesting issue because it raises a lot of problems and concerns but we felt like we needed to compare these children to children who are being raised in equally enriched environments if we wanted to really test the true limits of their ability to acquire this language right but it gets complicated all right so our results in general in comparison to the control they scored in the norm in the typical range actually on the test for age but they were significantly lower than the controlled children on expressive vocabulary expressive and receptive language this was the self the comprehensive English language test and the and sentence recall so i won't go into that in detail but they we were surprised we didn't expect this because all of the other researchers found that they were in the normal range and we did too but in comparison to this control group they were doing significantly poorer and i didn't want i didn't want to do anything with these data and unless we could establish that it wasn't just one off set so we brought all of these kids back one year later we have a new control room and we found the same results so then we published it because i think that's the other thing if you're doing this kind of research frankly you have to be really careful about the ethical implications of your findings and you have to be really sure that what you're getting is valid because in this you can imagine in this case if you're going to report these kinds of results if it's not valid then it's going to create all sorts of problems for the wrong reasons it still could have all sorts of adverse effects but if it's valid it's important to know so this this is just to show you the results of the specific tell if it appears in red then it means there's a significant difference and this is striking they're within the norms but they're not scoring as well as the controls so it means that they're not able to do something so they're so it's so what so that so what they look they're doing just as well you're going to regret that you ever said that they they look just like typically developing modeling with children but the so what is but hey there are environments where they're not doing as well as other children and one of the women who does a lot of the research on this in the u.s she said to me but this is really important because these children often go to schools this is in the u.s not okay they go to school where there's a lot of high achieving parents and children and if they don't do as well as the other children this has social repercussions for these children i hadn't thought of that but i thought it was an interesting point so we did another study when they were in school i'll show you the sample in a moment and i wanted i was particularly interested in doing this partly because people every time we tried to publish these results the researchers and the reviewers said but of course they're not doing as well as the controls they've had less exposure than the kids and we would say yes but they've had four or five years and this comes back to the early point if there really is this critical period advantage it shouldn't take more than five years to show itself that in my opinion but anyway in order to address that we decided we would do a follow-up when they were in elementary school and i was particularly interested to see whether the enrichment and language enrichment they would get in school would maybe boost their performance up because arguably in school they're getting a very enriched language environment right whereas in the home often these kids are single children because their parents haven't been able to have children of their own so maybe the language environment is not as enriched although it's probably richer than most kids get so we wanted to really test that out so all you need to know is that the age of testing now is nine to 12 years of age and they're in grades four to seven in our system their length of exposure to french had been 80 months so that's quite long right and again we were matched to control children in all the same ways here's the results here's the anything in red means there's a significant difference so so what's important about this is the test we gave them a we gave them a lot of control tests we never find any differences on cognitive ability or social emotional adjustment or attachment that's just to show that the receptive vocabulary was fine and just as good as controls reading comprehension was fine and so was word association but they're significantly lower on receptive grammar word definitions expressive vocabulary and recalling sentences and this is particularly interesting because they were also below the norms on sentence recall how many of you are in kind of clinical areas of language sentence recall task is a very simple test where you and if you're doing individual different stuff it's a great test the kids are simply asked to call verbatim sentences that you speak to them and they get more longer and more complex so you say the dog is sleeping children thinks the dog is sleeping the dog is sleeping inside the cat dog asleep it's not the same frame but the dog is sleeping inside the cat that has is on the mat so it gets more complex and you simply see how many of these sentences they get right it's used as an indicator of the possibility of specific language it correlates with a lot of things so it's a very useful omnivist measure so this was really interesting to us just to show you what their performance looks like in more detail work this is the percentage of children in the internationally adopted group who scored at the same level as the controls so this is between minus and plus one standard deviation so this is this is what you'd expect but this is the number of kids who are between minus one and minus two standard deviations below the control group and this is the number of kids who are percentage of kids who are more than two standard deviations below the number this is high very few are the much more better than the control now this is this is skewed because this is a very high performing control group but these kids have had the same kinds of experiences these kids more or less right okay so then we decided so the question is what's going on here why are these kids doing less well it could be that the pre-adoption environment is adverse for these kids we don't think so we didn't think so because their cognitive ability is high social emotional adjustment is high their medical condition is fine and they're scoring in the normal range on these tests so there's no indication that there really is some kind of developmental delay problem perhaps not enough exposure but by now they've had 80 months of exposure to french only in very enriched environments so it's still possible that exposure issues but gee this is really pushing the limits of how much exposure you need to be proficient in language could be schooling but we've tested them in schooling so that's not a fact so there's two factors one is attrition of the first language it could be because they've lost their first language they have a weak base for acquiring french as a second language but it could also be that they have a verbal memory difficulties as indicated by the performance on that test okay so and there's lots of research that shows that verbal memory working memory and short-term memory is related to language development in a lot of different domains in first language active issues second language I won't go into that it's a really interesting area of research so we did yet another study again we had high age children controlled children matched on scs and so forth and so on at this point they'd had nine and a half years exposure to french and they were in grades four to six none had repeated a grade so they're doing well in school and their schooling has been entirely in french here's the results here's the control measures no significant differences on this is a non-verbal measures of cognitive ability but in the wexler no differences here's the language abilities significantly they're within the norm on three of them but below the norm on two first time we've been below the norm on language does per se and they're significantly lower the control group on all of these here's the verbal here's the real interesting results the verbal memory scores of these children are within the norms in most cases but they're significantly lower than the controls but they're non-verbal memory is fine which is really amazing just to give you an indication of what's happening with their language here is a percentage of children i.a. children who scored more than one standard deviation below the control of children at seven years of age and then at 10 years of age what do you notice seven to ten years of age their performance is getting poorer not better that's really striking because you think the more exposure they have they should get better and children get better now again this is a very high performing control group so our theory was that verbal memory might be the underlying issue here so we did some regression analyses does everybody know what a regression amount it's really a fancy correlation that's what i told my statistics recent and they relaxed so when we went we put in a bunch of measures including short-term memory working memory and other uh another correlates like age of adoption length of exposure stuff like that and interestingly and we ran these correlation regression analyses for the i.a. children and the controlled children separately the for the i.a. children remember these children were almost 10 eight or nine years of age the short-term memory tasks were significant predictors of the performance on the measures that they scored significantly lower for the controlled children exposure or age was a significant this is what you'd expect for children who are learning the first language short-term memory working memory are significant predictors but only up to a certain point in the eight years of age and then age becomes more significant because it's not short-term memory doesn't influence language development what happens is language development improves short-term and working memory for reasons i don't believe but anyway this gave us some cause for concern right or at least pointed this into what might be the issues for these children now so the so the data point in this direction that either the attrition of the first language is a problem or delayed exposure to the altruism is the issue we don't know from these children because we can't separate delay in exposure to French and their attrition of the language right because the delay is caused by attrition to the sense right and that the and one or two of these is why they're doing poorly in verbal memory and our argument would be if you're exposed to a second language as much as little as 12 months of age then your verbal memory for the new language is not as good as that of children who have been exposed to that language from birth because the delay is influencing language development through its effects on verbal memory because because memory we're all born with the capacity to remember otherwise how could you learn but memory it has to be primed so it's memory for something memory for spatial information memory for acoustic information and for language so most theories of cognitive development when they talk about memory they distinguish between spatial memory visual spatial memory and acoustic or auditory data let me listen to information right so our theory was the delay that these children experience in their exposure to French has in a sense altered their verbal memory for French in comparison the long-lingual native speakers and that that sets off cascading effects on their opposition of French as a second language and these effects actually get largely because because weaknesses early on make larger problems later that's the hypothesis but maybe it's attrition of the first language because okay so here's where it gets really interesting and so we decided to have student Lara Pierce who wanted to do some brain research so we collaborated with Denise Klein from the Montreal Neurological Institute to look at do some scans with these children now at this stage there about 18 years of age there's we had been able to recruit 23 children these are new adolescents really we're 23 years of age and we also we also recruited 13 monolingual French speakers the same age and this is the really brilliant part of the study I can say this because it was Lara who did this decided on this group we had Chinese French bilingual children so these are children who were born in Montreal lived with Chinese speaking parents when Chinese is the first language and then began to learn French as a second language at the same age as the adopted children because it's very common in Montreal for immigrant children to begin to learn French very early on because they go to daycare centers or home care centers which are French so those are the three groups so the inclusion of this group allows us to disentangle I think I'm about allows us to disentangle attrition from delay because these two groups both experience delay but only the IA children experience attrition we're going so that was a really nice thing so we did two studies with them one was and this attracted a lot of attention we wanted to see if there were any remnants of Chinese in these children's brains to be across so we ran a study in a scanner where they simply had they heard pseudo words in Chinese that either were the same in tone or different in tone so as you probably know differences in tone are phonemic in Chinese is it somebody give me an example of a who's Chinese speaking here an example of a word where the tone changes and it changes the meaning of the word can you pronounce those no no use the variations in mom I mean that sounds the same to me I don't know about you but it's not it's the difference between saying mother and horse right because the because the a sound the tone changes the meaning of the word changes it doesn't matter in French or English what the tone is all right they just simply had to say same or different so we ran the kids and interestingly enough the performance of all three groups behaviorally was the same so the monolingual kids could discriminate just as well as the bilingual kids or the adopted kids there is no difference in performance and here's the performance of the bilingual kids I don't know why I'm going to put them all up should I allow people to leave now you can leave if you want I bet you want to see the day so here's the here's so this is a Chinese task the so and unfortunately people who do this kind of research has a very perverse habit of representing left on the right and right on the left sometimes but not always so in this case the left part of the brain is on the right side and the right part of the brain is on the left so in the in the monolinguals because it's a Chinese sound discrimination task they can do the task but the area of the brain that's doing the task is in the right hemisphere so they're using non-linguistic areas of the brain that are typically so associated with non-language process the bilingual children who speak Chinese as a first language they're using left hemisphere right so what do you think the i-h children are going to do are they going to look like the monolinguals because they're monolingual and French or they're going to look like the bilinguals because they used to speak they started to get Chinese but then they stopped how many think they look like monolingual French how many think they look like the bilinguals so if they look like the bilinguals what does that mean they've retained Chinese at least right and that's what happened so after and this was after 10 to 12 years of no exposure to Chinese there are traces of Chinese in the way they performed this task and this is an example of what uh you've been was saying that the behavioral results and the uh the neuroimaging results don't coincide because they could all do the task but they did it in different ways for the monolingual kids they're doing it as an acoustic discrimination task and for the monolingual bilinguals and the international doctors they're performing it as a linguistic task does that mean they're free-disposed to relearn Chinese well that would be an interesting thing to do that maybe they would learn Chinese at least Chinese phonology more readily than people would never be exposed to Chinese now if you want to leave please feel free because I've run out of time but I have one more study associated with this week and this is really the best study that you can leave you can leave but I don't know your name so I can thank you for that commenting on the question i think it were Korean adult teens uh adults who are the training to read their personal language and they were able to do that whereas the controls let me do the next piece because it's it is really interesting this just shows you the areas of overlap between the adopted children and the bilinguals here's another interesting result we did correlations between their involvement of this language area of the brain and their age at adoption so this is in months so this is the age at which they were adopted and this is the extent to which there was now how do you put this this is the extent to which there was involvement of the linguistic areas of the brain for processing this task so what you see is that the older they were at adoption the more their brain results were like that of the native speakers Chinese so exposure made a difference and really what implies is that the more ingrained so remember that I talked about this fine phonetic fine tuning thing so this is a period when fine tuning occurs so what this is suggesting is that even though they were not producing Chinese in many cases during this period of development their system was going from this language generals organization to specific to Chinese and it gets locked in and it gets and it gets more it's more native like the more exposure they have to Chinese now the next thing which is really the one that I wanted to do was to look at how they process French the second language because for me the question based on what we had seen in the behavioral results was well what are they doing in French I mean surely from a critical period point of view these kids should be using the same kind of strategies neurocognitive areas of the brain as monolingual because that's what the critical period hypothesis says the younger you are and especially if you're within that critical period you're doing it during a period of robotic critical plasticity and you should be able to engage the same area of the brain as if you were a monolingual speaker native speaker so they they did this in this case same group of girls they stayed on and we did this task it was an in-back task so they were given words in French that and they had to identify if they had heard the word in the preceding trial in the trial one two behind that and so on so it's a working many tasks for French sounds because you have to remember the sound and you have to remember when it occurred we'll just say that for a moment here the results here's the monolingual French speakers so this is the real reference point and there and here left is left just to confuse you so they're using left areas of the brain as you would imagine I haven't got into the specific areas that they engage but here they are here's the bilinguals so they're engaging left left hemisphere regions also some right hemisphere regions so what do you think the adopted kids are going to do look like the monolinguals or the bilinguals bilinguals that's what happened so even though they have begun to learn French between 12 and 24 months of age even though it's the only language that they know is and even though there's no longer interference from Chinese they're actually using the same underlying I mean this is a bit of a speculation because there's a lot that we don't but just to play this story out they seem to be they they're processing this language like a second language and the argument would be that the retaining these traces of Chinese probably related to phonological memory when they come to learn French then they have to build French on a system that's being set up to learn Chinese so that's why they look like the bilinguals right so we did a connectivity analysis this is the kind of analysis that that allows you to look at what area the brains are interconnected during the processing of the in this task so we ran this for the French monolinguals and what then you have a seed area seed region so you pinpoint one area that you want to be focal point of attention and here was left insula because this is involved in working memory studies and wanted to see whether what other areas of the brain were activated at the same time that you get activation of the insula and it turns out that the other areas of the brain that were activated were areas of a circuit which are thought to be involved in phonological memory in monolinguals all right so that's kind of what you expect when we looked at the bilingual in the i a children there we couldn't find any significant connections but for the i a children there was there was connectivity with the right frontal pole and the right middle frontal gyrus these are not areas apparently that are typically associated with phonological memory in other words that the results for the monolinguals are as we had hypothesized engaging a classical phonological working memory system but the second language learners at least these kids who have lost their first language are not now the implication is that this is due to delay and not attrition because we're not getting differences between the bilinguals and the adoptives now one final thing that's really interesting is there was a they did a comparison where you look at the task when it's simple you just have to identify every time you hear a target sound and compare that to when they heard the sound two steps back so the idea here is that by looking at memory for sounds that occurred two steps back this is probably demanding because you have to hold the sound in memory and you have to track what's happening between and so forth and so on so the question is when you increase the cognitive load how do the monolinguals deal with this and what do the bilinguals do and the i a children for the french monolinguals there was very little engagement of other areas of the brain there was some but it was relatively minor in other words this was not a significant increase in cognitive load for these participants interestingly for the bilinguals and the internationally adopted children there was increased activity in regions which are associated with attention attention cognitive control so the argument is and we talked about this with humans talk is that because they retained aspects of the chinese both the bilinguals obviously adopted kids they can't use exactly the same strategies of the i a kids can't use exactly the same strategies so they they use these what we call compensatory strategies to do perform this task and those are regions that are not strictly language related but they're related to more general cognitive process now this is pretty so the argument is that delay in exposure to the altitude produces differences from the logical working memory and that influences their language development so the question then is so what i mean really these kids are acquiring a second language under ideal situations somewhat like hilton stanz but they're not you know admittedly this is esoteric testing but what does it tell us in terms of theories of language acquisition and for people like me it sort of shakes me up because i wanted them to look like the model in this sample well it could be yeah we don't know what the mechanism but there's something arguably the traces of chinese of chinese of somehow or other influencing processing of french it could be an ambition potential interference but it could be something else that those that capacity to remember for logical information is not entirely well suited for french because it's set up for chinese so they're using other mechanisms to get around well no but remember full logical fine tuning means that you narrow your perception of contrast so it should make french easier but it doesn't so it may not be the i think so i mean other people said how do we know these kids are not genetically different but i can imagine all of them and we've had repeat different samples and do we say that these chinese kids are genetically different i don't want to say that i don't want to say adopted kids well but these kids were given up for adoption because of the policies of the country and and it's usually related it's not that they're entirely their home environment so we're entirely perfect but it's taken far from extremely adverse effects on the child well there is a lot of research that shows that strong first language skills support the acquisition of the second in the other way but not necessarily the same take me back to this very very interesting you take me back to the Swedish research because this is really important for us to be talking about influence and medication because there's always this belief that the other children are mindful of the better they get but you take away that their mother that gets in so many of these cases yeah but you see this research is not actually implying attrition because my name uh yeah but it i think it's close it means that even earlier they're always going to be different and the fact that they're doing this as well behaviorally is as a typically developing means that there's a kind of neural plasticity that can be used to reach those levels of proficiency yes because it's not just it's late exposure of the L2 and it's not just the L1 efficient i'm thinking that it's a disruption of the learning process which is true of a typical secondary learners because remember the bilingual kids were typically in every other set they showed the same brain in the results so it's not these groups are really wasn't well but the Chinese bilingual kids are 13 years of age all of them are so they're pretty far along the way and they're living in Chinese families a lot of them were in China i mean the control group is a pretty typical group of secondary learners so i think that what this says is that all second language learners are going to show they're not going to be the same as monolinguals and we shouldn't expect them to there's nothing wrong with that that's just the way it is it's normal for these children to look different in certain respects because they're using different strategies related to neural cognitive development as a result of their language history so in other words if they do show difference although these kids are not showing differences with respect to test norms so it says that look these kids are very good at acquiring normal levels of performance but when you look at them more in detail they're doing it in a different way no that's the that's the sticking part who have learned but then they wouldn't be adopted just well that's pretty odd you know but then they're going to look like the bilingual kids then you've got bilingual kids but if it's the adoption most research shows that adoption is a positive experience for children they improve in everything usually anyway thank you for your patience something to think about