 Be a disaster. Thanks Maria. Welcome everyone. I'm very pleased to be opening this event today for everyone here. My name is Professor Marko Thomas. I'm the Pro Vice Chancellor of the University of Greenwich and I look after the Faculty of Liberal Arts and Sciences in which this important research centre, Creole, resides. I'd like to just begin by just thanking Maria and Nancy for their customary, typical, excellent organisation that they always have for these events. And it's great to be kicking off the new academic season with this event. And even though we are meeting on teams and people are starting to feel that perhaps we're getting a bit tired of teams and Zoom now and we can start doing more things face to face, it's interesting that we've got 42 people already on this call and I don't think we'd be able to have got 42 people from the number of different locations that we now have where we're running this event in Greenwich. Of course we do want to get back to organising things face to face in the future and probably having a sort of hybrid mix of these things, although sometimes it feels as if the technology hasn't quite caught up with the ambition. I think today's topic of language in autism is really, really important. I myself am very interested just in the whole meta language of autism and designated disabilities themselves, which is often quite troubling. I think huge assumptions can be made in relation to the linguistic development of ASC children where the impact of deficits and social interactions can be assumed to underscore cognitive development too. And this really provokes questions in relation to the relationship between language and autism more generally. Alongside this we've seen a huge amount of greater public awareness around neurodiversity and a growing willingness to move away from neurotypical registers that might lead to homogenised views of behaviours that are deemed acceptable or credible and legitimate in wider society. So I think this topic is really, really welcomed. As I said it's incredibly important and quite timely as we're all eager to discover much more about this important and much overlooked area of research. So a huge welcome from the Faculty of Liberal Arts and Sciences and I'll hand over now to our Director of Research Professor Chris Bailey. Thank you Mark and again welcome everyone to this opening research talk under the under the CRL initiative and welcome to our colleagues from Amsterdam. We're certainly looking forward to your talk. CRL has from my perspective as Director of Research and Enterprise has been one of the fastest growing sensors we have within the faculty. Really you know responding to these societal challenges and you can see the slide here with regards how language and linguistics is relevant for health, culture, creativity and inclusive society and digital industry and space. At my area is in computing and engineering so I'm very interested in how artificial intelligence also interacts with linguistics and you know we hear a lot about computational linguistics and then I start thinking about Chomsky and these sorts of things. It's a very fascinating area but I think what really is impressed me is how we're bringing into key areas for the university but how CRL is really sort of exceeding in research, knowledge exchange and building partnerships and these free areas really underpin the university's future research strategy. So great to see the agenda for today and I look forward to seeing future successes with CRL going forward. Again welcome everyone. Thank you Chris. Thank you. So what you I hope you can see now is a slide that captures the strands of research that we have been carrying out in CRL. Sorry, sorry. Yeah okay maybe it was an echo. It is echo, sorry. Yeah so these are the main strands that we have been developing at CRL and now I'd like to invite the colleagues who are the leads of each individual strands to say a few words about what is coming up in the year. I want to thank you also the advisory board members who are presenting the call I've seen before because they have been supporting the center from the very beginning and they have insisted that we gave an idea about all the research that is covered under the umbrella of CRL. So for some reason I cannot see all the list of participants but if Rosanna, if you are there, would you like to say something about the relationships between CRL and the Institute for Life Course Development? Hi, hello. I hope you can see me. Hello Maria, hello everybody. I am Rosanna Pacella. I am the director of the Institute for Life Course Development and I'm really here in support of the very, very strong CRL-ILD partnership that we have built over the last two years. I think everyone can agree that it has fostered interdisciplinary research and collaboration across the faculties and really added value to both faculties so very, very proud to be here and you know just in terms of some recent developments to say that the Royal Barrow of Greenwich we have a very strong partnership with children's services in the Royal Barrow of Greenwich and they have been in touch because they would like to focus on speech language and communication needs in children and especially wanting us to look at the long-term consequences so looking at health, employment, education and even criminal justice, youth offending outcomes that are linked to language needs across the life course and to also look at the costs of language needs to the community. So very exciting work and yeah very excited to be here and to be part of this CRL-ILD collaboration. Well thank you Rosanna, thank you because the ILD has been an asset for CRL, everything we've been building up together so thank you and yes what you've mentioned is something very important that is coming up in the agenda so thanks, thank you very much for your time all the time and also here today Rosanna. Thank you. Very good so let's carry on with the colleague who belongs to CRL and ILD Anna Samara who is the lead of the of the reproducibility club which is a very important initiative, a national initiative and she is the lead that represents the university at UK RN at the UK Reserved Reproducibility Network and that is part of the very important movement of open science so Anna would you like to say a few words about the work that is coming in the year? Yes of course, hi thank you Maria, hello everyone, it's a pleasure to be back here at the launch of the CRL activities for 2021-2022 and I'm here in my capacity as one of the representatives of the Science Practice Hub that lies within CRL and one of the key activities within the Science Practice Hub is the reproducibility general club that is also aligned with my role as local network lead for the UK RN as Maria already mentioned so just a couple of brief updates on things that happened over the summer but also on things that we are planning for the near future. So first of all I represented the University of Greenwich in my capacity as local network lead in a summer retreat that took place in July in Cumberland Lodge where we held very interesting discussions with other representatives from other UK-based universities on issues of reproducible science and there is many interesting opportunities for collaboration across institutions that we will be exploring in the coming few months. I also wanted to mention that Greenwich University participated in evidence that the UK RN submitted to a science and technology committee inquiry so there was a science and technology committee inquiry on reproducible research and the issues of problematic bad science practices and now the University of Greenwich we also submitted a response and some evidence to this parliamentary inquiry. So these are things that happened over the summer and we're currently putting together a schedule for the new reproducibility series for the new academic year probably with an emphasis on external speakers coming and giving us information about open science practices and techniques that we can use to become better scientists really. So hopefully in the next few months we will have a schedule finalized we will probably be launching the first talk closer to the end of November so please watch this space for many more interesting talks this year. Thank you. Wow amazing Nana very very good work thank you very much thank you and looking forward to it looking forward to everything and thanks for your leadership on this thank you. So let's move on to to Katarina Katarina Stenke who is the lead of the narrative seminars because we've been speaking about all this corner here of science practice and open open science and then we also trail also does research on on this trend on the pillar that has to do with creativity and and heritage as well literature oriented. So Katarina hi I know that you guys have a program ready for the year so can you say a few words about it or all the members yes yes we can. Fantastic well I'm I'm really delighted to be here speaking as a representative of the co-organised wonderfully successful in fact narrative seminar series that we started running last year there's myself along with associate professor Justine Bailey senior lecturer Emily Critchley and now most recently Annika Centeno all from the subject groupings of writing literature and language so we're definitely on the humanities scale of things and we're very proud of what we bring to parallel and you know very rich and active research cultures have informed our seminars but we are also interdisciplinary and we've really taken our leads this year to advise yet another successful program of seminars around the theme of narratives we've been paying attention to this interdisciplinary dimension but I'm just kind of very briefly talk about what we have planned and I will also post in the chat a flyer for our first seminar of the year which is happening in the 27th so our program this year aims to reflect the interest of reaching prel as well as responding to key themes from university strategic goals we'll be looking at building on our successes from last year reaching an increasingly wide audience we already had some fantastic sessions last year with speakers and attendees from multiple countries and the different continents the term one we have themed around the topic of transnational narratives so we're looking at narratives that describe or respond to the challenges and possibilities of transnational crossings which seems like such an important topic for us in this present moment we'll be looking first of all next week at 19th century transnational narratives we're going back to the past to see what we can learn from it and we've got some fantastic speakers at the University of Newcastle and the University of Ghent so that's our first first panel and we've also got a wonderful session organized by Professor Justine Bailey where we'll have colleagues from the University of Liverpool speaking about a really kind of noxious kind of narrative that has had to be rehabilitated that is the narratives of enslavement okay amazing thank you I'll speak very briefly about what we've got for term two to term three and then I'll move on to colleagues but it's worth saying that term two we'll be looking at narratives of sustainability and we have a series of panels which combine creative writers with activists and academics within the field of environmental studies and in term three we've got narratives across disciplines where we're specifically focusing on how to use this kind of concept or this idea or research tool of narratives in different disciplinary contexts and we'll be in dialogue there with a research group from criminology gender deviance in society who would so we're in short very very excited to be running this series and we really value and benefit from our association with Braille and the students and especially with Fourier Arcade now specifically working on it. Thank you thank you Katarina very good we're looking forward to it and and yes please paste put in the chat the program for next next Wednesday the one after right so yeah thank you okay so I'd like to invite the last two colleagues I'd like to hear from Neil Sonder, San Cici Laval they they have a very active participation within the center and they will have things to tell us about what is coming up so Neil go ahead thanks Maria thanks everyone very briefly I'm Neil I'm a lecturer senior lecturer in the School of Mathematics and Computer Science and my particular area is in pure mathematics but Crale is a wonderful research institute where ideas from mathematics and computer science can feed into the interdisciplinary research but also just locally in our school we've recently had some new acquisitions and expertise in AI and machine learning and Crale is just the perfect place for us to be exploring those research interests in a wider context as well so things that are coming up later in term one I'm organizing a discussion with a lecturer in artificial intelligence to ask what it really means for a computer to understand and and dig deep down and into how language in very much the Chomskyan view is there to help comprehension as there as much as it is to communicate with with other agents if one thinks about neural nets that are fundamental to AI one thinks about automata which can be deterministic or indeterministic and lying behind the mechanics there is some sort of formal language and the properties of the neural net can be expressed in terms of these formal languages so so we're hoping to have some more seminars in term two on that but I'll be advertising those as sooner to the time okay that's lovely we are looking forward to those Neil thank you very much thank you and last but not least before we move on to the talk Cecil I know you are here I saw you earlier on thank you very much for being on the call thank you Maria thank you everyone so I am Cecil Laval I'm an associate professor in second language acquisition and the interface between second language acquisition and language teaching so how do we become better language teachers which has to start with how do we acquire a second language but the what we want to build on further Maria and I and the rest of the team is to within the quail framework is really building from the fantastic research our students are doing undergrad and postgraduate because we have as part of the school of humanities and social sciences we are offering languages French Spanish Italian and Mandarin and we have a very high level of languages and students are doing a language research project and we really would like to have them embedded in in the quail talks as well in terms of not only research sharing research but also employability as well to showcase what they're doing because they're doing a fantastic job and and in terms of inclusivity equality diversity we really need to to showcase as well what what we're doing there is for example one of our family are highest language level students who's looking at gendered language or languages and the impact on society and what it can do and that can lead to some very interesting discussion around our students and colleagues and research papers on this in terms of of inclusivity and equality and a very important agenda so it's about bringing second language teaching and our second language learners into the center and really building from that research undergrad and our postgraduate students as well of course and our own research because our team of colleagues working ranging from translation to language testing and also second language acquisition or applying linguistics in general so very very excited and we will have some interesting project coming on hopefully you are waiting maybe a little bit before time too to have something on campus and make a make it a bit different than online but thank you very much hey well thank you thank you Cecile thank you very much for this yes all right so this was the overview of what is coming up in thank you very much to all the colleagues who are going to be in charge of leading all these initiatives because this is teamwork completely teamwork and other things that we have in on this slide as is the international summer school and multilingualism which we hope to run again after last year which involves our visiting professor Mike Putnam from Penn State University who I think is on the call today as well and our literature conferences like the Victorian popular fictions annual conference so we are very much looking forward to all these thank you very much to everybody for listening to what is coming up now you have an idea when you receive invite emails what you can find and finally I want to introduce to everybody the speakers of today's of today's talk we you know we mark has already said we have a talk about language in autism that is going to be delivered by Ileana grandma Harriet Reynolds and Janet Shaffer they all belong to the language in autism laboratory of the University of Amsterdam so no surprise and there that they are here but well in any case they are leading scholars in the topic uh Janet Shaffer Professor Janet Shaffer is a worldwide recognized expert on the links between language development and extra linguistic cognitive knowledge across populations and languages she has studied at the University of Utrecht UCLA and MIT she worked at Ben Gurion University in Israel and has been in Amsterdam since 2011 she has to be recognized for many achievements in the linguistics in the linguistics discipline and one of them is that what I put in the blurb that she's been the founding lead of international networks such as the language abilities in children with autism the um where the analysis of language development and intelligence executive function theory of mind and coherence provided a more refined picture of human cognition uh together with her we have uh two colleagues from her lab Dr Ileana grandma who lectures at the University of Amsterdam and you may have seen that she got her BA from the University of Bucharest and MA and PhD from Utrecht University in linguistics she's an expert in statistical learning which I know is a topic that some of the colleagues here at Creola very much interested in and its potential role in language development in infants adults and clinical populations finally uh Harriet Reynolds um very close to the heart of many of us because she did her BA at the University of Sheffield and the MA in linguistics in Amsterdam she's currently developing a doctoral study on information and structure in autistic and neurotypical subjects connecting linguistic areas such as pragmatics or prosody with cognition and psychology so you can imagine that we are very lucky and I'm thrilled that they are giving us some of their time um and without further ado I'm going to stop sharing and and and see if you guys can share if you cannot share just let me know and I will share your slides from my end okay I will try one more time yes okay so then I have your presentation and I'll I'll retrieve it don't worry just tell me when you want me to pass the slide although I follow you very well for Harriet I think I can also it seems like it might be working for me we need to do it okay okay okay better if you can do it go ahead I would like to share and Harriet that would be great I think it's my mark that hasn't been set up well for the teams thinking thinking oh yes we can see your your screen Harriet thank you yes it's very good so you I'm going to mute my mic now that we have you all in place thank you very much again and the floor is yours thank you so much and what an impressive introduction of this crel center I kept saying C rel to Ilana and Harriet but now I know how to pronounce this crel what an incredibly diverse and multidisciplinary and impressive center you have there in Greenwich um you set the expectations very high so I hope we can meet them just a little bit um I'm very happy that you uh invited us and and yeah I think I can talk for Harriet and Ilana here as well that we feel very very honored to be the speakers of this first crel event this academic year so Maria thank you very much for inviting us um I move let's move to the second slide yeah so here are today's speakers but Maria did a fantastic job introducing us already and you can see each other you can see us on the screen as well so we can move on to the to the next slide where I will give an an outline here I I myself will start with a little introduction to language in autism and um then Ilana will tell us something about language and statistical learning in autism and then we'll move on to the topic of pragmatics in autism and all of us will tell you something about that Ilana will say something about an ongoing study on implicatures I will say something about an ongoing study on topic encoding and Harriet will say something about noun phrases and pronouns which is at the heart of her current PhD project um so what we are intending to do here is just give you a little glimpse into the ongoing studies at the moment so nothing is really finished we have no big results that we are going to present to you um but we would like to give you insight in in in our group language in autism and and we hope that you will have good questions and comments about this so that it will raise discussion in this very diverse group of people okay so um what is autism let's start with that with the autism spectrum condition as it is called nowadays and I think one of the colleagues here already mentioned you know the the meta language that is used um to refer to autism we will not say a lot about that but we are very aware of all the changes and the rapid changes in in the language that we use to refer to autism so a while ago it was the best way to refer to it or the most accepted way to refer to also by the population themselves was to say people with autism or people with autism spectrum disorder nowadays we move to the so-called disability first um terminology where we refer to them as autistic people or autistic individuals autistic children um or people with autism spectrum condition rather than a disorder I myself am uh not used yet very much to say condition so I may actually make errors I'm going to try very hard to to use the terminology that the autistic population prefers although I also realize that within the autistic population there are different preferences so forgive me for making errors in this but um we're trying to do our best to keep up also with the most accepted terminology for autism spectrum condition what is this well according to the dsm 5 which is a sort of a manual that is used in diagnosis of autism there's two main characteristics that are being mentioned the first one is persistent deficits in social interaction and communication and the second one is restricted repetitively patterns of behavior interests or activities okay now none of these is per se about language but we can imagine that's the first one deficits in social interaction and communication communication comes a little closer to language that that has repercussions for longer so we're going to zoom in to language now um and I'm going to tell you a few things about general things about language in autism later on we're going to talk more in detail about the studies that we're doing right now on language in autism but um many of you may know that 25 to 30 percent of the autistic population is actually minimally verbal or non-verbal and what does that mean well it means as far as we can describe this that these people this part of the autistic population doesn't have productive language we don't know so much about their comprehension of language so that in and of itself is a would be a very interesting field of research so 25 to 30 percent minimally verbal or non-verbal and of the verbal autistic individuals we most of us agree that they have that they're different from neurotypicals in the pragmatic area of language and but in addition to that and this was traditionally not assumed that of the verbal autistic individuals many have of them and in some studies we see this up to 50 percent that they're different from neurotypicals in structural language and structural language we mean grammar more for syntax but also phonology as the structural sounds so if we take all this knowledge together about this large percentage that is minimally or non-verbal that verbal autistic individuals have problems with at least the pragmatic part of language but also often with structural language it's quite surprising that language plays a very minimal role in autism diagnosis and in intervention and it's even more surprising and this is something that that you people in the in the Krell Center realize very well I think that language is the key to social success to academic success to life quality to partnerships I mean language is such a crucial instrument so one of the missions of our language in autism lab in Amsterdam and also of LAKA is really to put it very simply to put language on the map of autism let's go to the next slide oh and what I wanted to add that the previous you don't have to go back to the previous slide is that what is pragmatics of course pragmatics is a huge area in in language and it's basically the the use of language in a context and a context can be a text textual written context or a spoken context it can be a social situation but it's language in context and that varies from turn taking when you know how do you know when to use a question or a declarative or an imperative to the choice between a definite and an indefinite article something close to my heart is when do you choose to say ah when you choose to say the scalar implicatures information structure but also all the non-literal meaning that we get from a piece of written or oral texts including inferences figurative language irony idiomatic expressions metaphors so there's tons of research that would be very interesting to do in this autistic population okay so what about existing theories on autism if you browse websites on autism or of autistic association associations for people for autistic people you often come across three main theories and I will go through them very briefly here the first one revolves around theory of mind and the idea is that theory of mind is delayed or incomplete with all the repercussions maybe for language as well and in the second one executive function central role it's hypothesized that executive functions such as inhibitions switching attention working memory that they're incomplete or delayed or part of them is incomplete or delayed and the third theory revolves around the idea that there is a delayed or incomplete acquisition of central coherence or information integration you may or may not be familiar with with these theories but these are three psychological theories I would say that have been trying to say something about the underlying nature of autism and I also wanted to make a couple of remarks about another non-linguistic part of cognition in autism which is IQ well when we're working with a potentially language impaired population then it's very important to realize that a verbal IQ test assesses IQ through language and that may skew the scores right and second we know from previous research that low IQ does not always go hand in hand with low language ability so recent research have shown that autistic people with low IQ can go with good language abilities and vice versa that high IQ can go with weak language abilities and that is something that a lot of people are not aware of and this morning I was teaching a class and one of the students asked so so those those children those autistic children who are minimally verbal do they always have low IQ and that is not the case it doesn't always go hand in hand so so that would be important to point out um yeah so our more specific lia goals are to describe the language profiles in autistic individuals and how they may or may not relate to extra linguistic cognitive abilities and so that's a descriptive goal and a more explanatory goal is to figure out what the underlying nature is of these language profiles and the underlying nature we could look for them in within linguistics itself but also within other cognition and and also in the brain next slide please thank you so um in the spring of this year our group carried out a project on language cognition and statistical learning with autistic adults we had hoped to work with children but because of covid we couldn't visit school schools were closed etc you know all about it and we managed to set up a online test battery through the platform prolific and these are the tests that most of the tests depicted here is what we managed to administer online to autistic adults so we had tested statistical learning and social cognition in language and some baseline cognitive tests such as nonverbal intelligence and and memory tests and um if we go to the next slide then ilana will take over the floor and tell you something about the part on statistical learning and after that on scalar implicatures which is part of pragmatics ilana can you live with um yeah go through the slides or do you want to request control i've uh requested control i don't know from who but if you have some animations also right i think this might might be for you to um to have control yourself i have to request it again if it's not working then i can just say next but it will be a lot of nexts i should have foreseen this alternatively i can try to share my own i'm gonna request control again because i don't seem to be able to so what we're seeing now is the slides that maria is sharing something's happening harry um okay can you see the slide uh my slide okay and can you see it now at presenter mode yes okay perfect okay thank you very much uh yeah so thank you uh i'm ilana grandma and uh yeah thank you for that nice introduction also i am i wouldn't say an expert in statistical learning but that's definitely mostly what i've been working on in the past few years um so for those of you who don't know but i i would assume there are very few if there's already an interest in statistical learning um i would define statistical learning as the ability to detect patterns and regularities in the input whether that input is spoken input or auditory non-linguistic input or perhaps visual input the ability to detect structure and regularities effortlessly and often without being explicitly aware of it um so many of you would recognize here a short demonstration of the serial reaction time task so this is one simple way to test statistical learning and indeed one way that has been used quite often with individuals with autism to test their ability to learn the regularity in the movement of the um sorry of the target on the screen if you look at it repeatedly we'll start to see that there is a particular sequence and we know that statistical learning is related in some way to language because in developmental language disorders both children and adults diagnosed with a dld show systematically show a deficit in statistical learning we also know that sometimes statistical learning in some tasks correlate with the speed and efficiency with which we process complex syntactic constructions and that in brain imaging studies it's shown that it engages some of the same brain areas as language processing so it's immediately interesting to look at what the statistical learning abilities are in a population such as autism which has a diversity of language profiles from as shunet said people who are minimally verbal to people who only exhibit subtle pragmatic deficits and otherwise typical language development so this is an interesting relationship to study and in fact it has been studied people have been interested in the statistical learning abilities of autistic individuals for a long time for decades and the studies that have been produced actually have quite contradictory findings some studies find that autistic individuals do have are challenged by these statistical learning tasks although the validity of these studies is questioned based on their small samples and more recent meta analyses have not really found this that this is the case fmri studies show that autistic individuals might show less brain activation while they're learning but there are also many studies that actually find that autistic individuals are not challenged by statistical learning that the for instance the learning outcome so whether they learn is not different in autistic populations versus neurotypical populations although other studies find that while autistic individuals don't have problems with statistical learning they do in fact show different learning strategies from neurotypical individuals and there are also studies that find in specific test paradigms that actually individuals with autism perform better than the neurotypical individuals or that within the population of autistic individuals those who show poor social cognition more pronounced autistic traits actually show better learning so this is a very puzzling reality and our study is partly dedicated to understanding why these findings are so mixed and to get a better understanding of what the statistical learning abilities of autistic individuals actually are and also to fit them within the broader frame of language and social cognition does statistical learning underline language development in autistic individuals and to one extent perhaps to a greater or a lesser extent than neurotypical individuals and how is this relationship perhaps mediated by social cognition so our aim in the past year has been to develop better statistical learning paradigms which are more robust identifying the true statistical learning abilities of individuals with autism so for one we would like to have a paradigm that shows us different kinds of learning of multiple regularities within the same statistical input statistically structured input that is in a way that is most similar to natural languages we also wanted to develop a task that is simple and accessible to individuals of different ages IQ and autism severity so a task that is really also accessible to children as well as adults and also to to minimally verbal individuals or individuals with lower IQ a task that captures not only the learning outcomes but also looks at whether the trajectory of learning is different in autistic individuals and to understand the relationship between these statistical learning and of course language and social cognition measures so in order to develop this task we've looked at some of the previous research uh can we hear I think Ilana's frozen we I don't hear her either oh yeah she's frozen she may be undergoing some bad of bad connection I'm really sorry I seem to have dropped out of the meeting I can still hear you we can't hear you I'm sorry I will try to reshare mm-hmm this is where I left you or did I drop out earlier here here okay yeah okay and at the top of the slide okay um so I was talking about a task that we adapted that was developed by other researchers like Maria Vendor and Doug Sadie which is a task based on Fibonacci grammars and these grammars are interesting because they display some of the structural phrase structure properties of natural languages as you can see in a Fibonacci grammar is derived from the recursive application of some very simple rules in this case that zero always rewrites as a one and one always rewrites as a zero one and this generates longer and longer strings that actually as you can see in B this is the 11th elaboration so the 11th level of recursive application of the rules which generates a very long binary string that is interesting because it displays some really interesting structural properties so for instance you can see that in this grammar zero is always followed by a one so all the red ones are predicted based on the fact that they are immediately after zero and two ones are always followed by a zero so all the yellow zeros are predictable by being preceded by two ones but also higher order regularities so a sequence zero one zero one is always followed by a one and this can be easily derived from the rules and structure of the grammar because at the higher hierarchical level sequence zero one zero one is the rewrite of a sequence one one which predictably from rule two must be followed by a zero so these kinds of grammars have a series of increasingly complex rules they are very simple in that they are binary there are only two sort of vocabulary elements and they're also interesting because each element in these strings is actually either predictable by one of the three rules or less predictable or what we what I will improperly call random so this rule is actually this this grammar is actually very amenable to uh setting up in a visual statistical learning task which is what we did so as you can see every one in the string corresponds to the let's say left hand location of a stimulus on the screen and every zero corresponds to the right hand location so you can present the Fibonacci string generated in B by presenting sequences of these stimuli presented to the left and the right and we can see how fast individual learners react to the presentation of the stimuli so to show you this is what our statistical learning task actually looks like so it's like a whack-a-mole game you get these birds either on the left or on the right hand side of the screen and you need to react by pressing the left or the right corresponding button as soon as you see the stimulus and we analyze the data in terms of the reaction time for each individual stimulus and we predict that when the rules have been learned the reaction times will be faster for the elements that are predictable according to those rules than for the elements that are unpredictable so if you see again here for for each rule you can measure the reaction time separately and then compare it to the random elements to see if the reaction times are significantly slower and so as Jeanette mentioned we already conducted an online pilot study with adults this spring on the platform prolific we found a variety of individuals that were willing to participate in our research some of them were diagnosed with autism officially some of them considered themselves to some of them were self diagnosed or they believe they had autism but had no no official diagnosis and some of them were neurotypical adults and we also looked at their social challenges by testing them on the social responsiveness scale so what we actually found in our statistical learning task is that the simplest regularity in our language was learned by both individuals with autism and neurotypical individuals both with and without a social challenge suggesting that in this respect there are no differences in statistical learning but surprisingly the more complex structural regularity show the significant group difference in that the individuals with autism and social challenges as reflected by the social responsiveness scale were actually better learners and were the only ones that showed robust learning of this rule whereas the neurotypical individuals did not neither did the self diagnosed individuals actually and the more most complex rule was not really learned by either group because it's a very complex rule and it's actually rarely learned by more than a handful of individuals in a sample but yeah you can see here in the graphs how we compare the reaction times to the for instance rule two in this case to the random to the non predictable elements and you see on the left that the neurotypical individuals show an overlap and no discrimination between the two therefore show no learning of rule two whereas the individuals with autism on the on the right actually show a decreasing reaction time for the complex adjacent so rule two regularity in the final two blocks of learning and a significant difference from the non predictable elements so we also can show the trajectory of learning and the fact that that learning happens towards the end of the experiment since our individuals here were high functioning as one would say or at least had very good IQ scores and ceiling level language scores we believe that these these individuals are maybe have been protected in their language development by their excellent statistical learning skills so we would like to extend this research with individuals with lower IQ or poorer language abilities to see if the autism advantage still holds across the general across the spectrum or it's something that is specifically a protective factor for individuals that have strong language skills and finally we were talking about the fact that we want to look not only at statistical learning but of course at how language and statistical learning and social cognition interact in autism so part of my research also has to do with scalar implicatures because they are a reflection of the use of language in a social context so we know that language development in general is dependent on social interaction and that is something that affects individuals artistic individuals in general but we also know that there are parts of language that are specifically centered around understanding the communicative intentions of interlocutors and a phenomenon within this is the use of implicatures where the listener has to infer what the correct communicative intentions are of the speaker so we know of scalar implicatures where when you say on my cake some candles are burning that is pragmatically supposed to be interpreted as some but not all so that you would in on the left hand picture choose the target picture with some candles burning and not the competitor picture with all candles burning and we know that implicatures can also be ad hoc when you see for instance several boys with glasses and you hear the utterance my friend has glasses you infer that the speaker is being maximally informative and that therefore the boy that is referred to is the one that has only glasses and not glasses and the hat because otherwise the speaker would have also mentioned the hat so research into scalar and ad hoc implicatures has been done extensively including with children and adults with autism and we know that while adults seem to have no particular problem interpreting these implicatures that younger children may perform more poorly than neurotypical children in these tasks but not always the problem with these kinds of tasks is that when you present four pictures that really give away the possible alternative interpretations of these sentences it could be that you are in fact giving away to the child the purpose of the experiment so in our new research direction we are looking at the fact that individuals with autism might be able to outsmart this kind of task and we are trying to develop a new task that is more naturalistic and obscures the purpose of the experiment for the child or the adult taking this experiment sorry so we are collaborating with colleagues from Uject, Emmanuela Pinto and Shalom Zuckerman who have developed a coloring book an innovative method of testing language acquisition in children where we test for instance scalar implicatures by letting children color a picture with for instance several monkeys and giving them the sentence some monkeys are red and instead of having the explicit alternatives for the interpretation of the scalar implicature the children actually have to decide for themselves what the correct coloring strategy is in order to make the sentence true. So I think I've taken up enough time and I give the floor back to Jeanette. Thank you very much Ilana I don't know how much time do we have left Maria? Yes you have five to ten minutes but so then I would like to pass on the floor to Harriet because I think we would have too much otherwise so Harriet are you able to questions including questions you have until half past so you have plenty of time yeah we would love to have some questions so we'll see if we get to my bits but I think Harriet's project is more important okay all right I'm gonna do it again when I share my screen last time I got a bit stuck at the end and I couldn't get back to Teams to unshare so maybe at the end someone's gonna have to do that for me um all right let me find where we want to start from from here yes we can see it great all right so I'm doing a ph project on on the use and acquisition of topic which is essentially what we are talking about and and what we know already from the literature is that autistic individuals may on the one hand produce fewer pronouns overall yet at the same time when they do produce pronouns they may be more likely to be ambiguous this is done there they're neurotypical peers and we don't know so much about comprehension although there are some findings to suggest that there may also be some difficulties with comprehension and we also know that interpreting prosody and especially when driven by pragmatics may be challenging for autistic individuals so how this how the information is actually conveyed in terms of intonation and we'll come back to that for the second study and so firstly I have well I have two studies that are sort of in progress at the moment fortunately I don't have any results for either and the first study is about the effect of the visual stimuli on noun phrase choice in narratives and the second study is about is a comprehension study about processing prosody cues in order to resolve ambiguous pronouns so firstly for my for my study one about noun phrase choice and visual stimuli so some of you may be familiar with these pictures they come from the frog story which is often used to elicit narratives and generally discourse from participants in order to then later analyze it so essentially we give them the pictures and then they talk and then we analyze this as a block of discourse but often what happens in this process is that this then gets dissociated from the pictures which is the original context and in these pictures the the discourse itself is already kind of split into separate events and then the speaker of course has a choice of what to do with that they can describe each event individually they can join them together they can split each event into smaller sub events and so on and the way that we conceptualize and then talk about events like this has an effect on the noun phrases that we use so if you start talking about something with a full noun phrase like the boy this kind of shows that you are creating a new section of discourse whereas if you refer to a referent as he throughout the discourse then you're creating some sort of coherence between between your utterances and therefore between parts of the discourse and so you are kind of connecting events um by using um by using more pronouns um so this has been shown in um typical in neuro typical speakers um that we have this tendency to do this um and so what this made me think about is that we uh what we collected this data with our narrative task but um ultimately the way that we present the the task does this have an effect on the noun phrases they use so do that noun phrase choices reflect um the visual discourse segmentation um and does this actually differ between autistic and neuro typical speakers um and we may think that it does because um previous findings have um have found that autistic speakers um may um connect events less in their discourses and of course we also know that there is this uh general tendency for them to use more full noun phrases in general so less pronouns and so could these two things be linked could it be that actually they are just segmenting the discourse into smaller sections so that's something that I would like to look at and then um I would also like to look at the fact that in terms of ambiguous pronouns when when again when uh shown all these pictures and asked to present to produce a discourse um the speaker has to um when they are shown a new picture they have to take in the events in the new picture they have to remember what the events in the previous picture were they have to remember what they said about it and plan their current utterance and so as we change pictures in the narrative in the narrative we uh we may actually be putting our participants um under increased cognitive load and so is it at these points that um that people start uh producing more ambiguous pronouns and might this be um affected by working memory abilities and so um the way that I'm looking at the data is I'm taking two contexts from narratives um contexts where a pronoun would be felicitous you don't have to use a pronoun but we kind of expect that people um in many of these cases would um and then on the other hand cases where a pronoun would be potentially ambiguous and then where um you could use a pronoun we are then looking at when people don't use the pronoun instead use a full noun phrase um and how this might be related to discourse segmentation and on the other hand when people use um a pronoun where it could be potentially ambiguous we will look and see whether that is related to the cognitive load both both of these relating to um how the pictures then interact um so we have again from the same data collection as Liana described um we recorded narratives from 12 autistic participants with a diagnosis there are also five um self identifying who gave us narratives and we have 25 neurotypical participants um so first of all I would like to do a group analysis where we take um all of the autistic participants and we match them as best as we can on gender, age and IQ to um neurotypical participants and then I would also like to do an analysis of social responsiveness scale um scores so that we can um use that as an alternative alternative way of measuring this um and using more of the population um and so we collected frog story narratives uh via zoom and they've been transcribed um and now need to be coded um and then for working memory we used an n back task then moving on to uh my second study which is about um the comprehension of ambiguous pronouns um if we have a series of um of sentences like this um Elena was happy after the tennis tournament the silver medal was a great achievement the coach applauded proudly at the prize ceremony for the next tournament though she hopes for gold in the last sentence we have the pronoun she and this could refer to either Lena or the coach um so if I ask you who hopes for go for gold um some people might say um Lena and some people might say the coach and ultimately what this depends on is whether you take this middle section of the discourse as being a subordinate structure which is therefore telling you more about Lena and why she was happy um or she was happy because the silver medal was great and the coach was really happy with her um and so then for the next tournament though she is Lena we're still talking about Lena or on the other hand if we see these as a series of coordinate um sentences then um we start talking about Lena then we start talking about the silver medal then we talk about the coach and then it makes more sense that she is the coach as this is who we were most recently talking about um and we can also manipulate this in the prosody to suggest one or other of the structures um so we can do this by taking the pitch range which is what is represented in these um diagrams um and also the pause length and so in a subordinate um representation of this discourse um we well we always start with expanded when we're talking about something new and then it goes to normal compressed to show that something is ending and then a longer pause and we go back to expanded which kind of creates a um a bridge back to um the first sentence and to Lena and so then um it triggers that we're still talking we're going back to talking about Lena whereas when it is coordinate um we can go from expanded normal normal and compressed to show that this is just one chunk of discourse and it's been uh shown previously for German that although it's um perhaps not as big a difference as we might like that um neurotypicals are sensitive to this and they do perform differently um depending on these conditions um and so what we would like to do is take the same paradigm and use it with autistic and neurotypical teenagers um and so we would like to know whether these global prosodic cues affect how adolescents resolve ambiguous pronouns um are there any differences between autistic and neurotypical adolescents um and so we would test such high school students who are um that's 12 plus uh we will have an autistic and neurotypical group with roughly 30 members per group um and we will also uh have a short working memory task as a control so we backward digits span um just as by listening to the um to the stimuli they do have to to some extent remember what happened in the discourse so just to check that this is not having an effect um and we're currently working on an online pilot with neurotypical adults um just to check the task design um and it will look roughly like this so they will listen to um one of the two prosodic manipulations um and then they will be asked who hopes for gold and so they make a connection hopes for gold is it Leina or the coach and then they press uh keys on the on the keyboard to select uh which person they associate with this uh with this fragment so that is uh everything I kind of raised through that because I know that we want to have uh time for questions um so maybe now I will attempt to stop sharing my screen fantastic uh thank you thank you very much both very interesting