Using a word-learning paradigm to investigate three forms of generalisation in the acquisition of lexical knowledge

Lead Research Organisation: Royal Holloway, University of London
Department Name: Psychology

Abstract

Abstracts are not currently available in GtR for all funded research. This is normally because the abstract was not required at the time of proposal submission, but may be because it included sensitive information such as personal details.

Publications

10 25 50

publication icon
Crepaldi D (2013) Seeing stems everywhere: position-independent identification of stem morphemes. in Journal of experimental psychology. Human perception and performance

publication icon
Lavric A (2012) Tracking hierarchical processing in morphological decomposition with brain potentials. in Journal of experimental psychology. Human perception and performance

publication icon
Merkx M (2011) The acquisition of morphological knowledge investigated through artificial language learning. in Quarterly journal of experimental psychology (2006)

publication icon
Rastle K (2011) Orthography influences the perception and production of speech. in Journal of experimental psychology. Learning, memory, and cognition

 
Description One of the hallmarks of language and literacy acquisition is our ability to extract general knowledge from exposure to a limited set of individual exemplars. These generalisation processes are particularly important in the domain of morphology, where exposure to words like 'repaint', 'retype', and 'redo' allow us to express and understand new words like 'retweet.



We studied the mental processes that underpin this generalisation process by developing a laboratory model of morphological learning. Adults are trained on a set of novel words that have an internal morphological structure (e.g. sleepnule, rugnule, teachnule), and are then tested on a variety of speeded and non-speeded tasks to probe what they know about the novel morphemes (in this case, -nule). Critically, the way that we probe this knowledge is by assessing the extent to which participants can generalise their knowledge to untrained words such as 'sailnule'. This laboratory method allows us the opportunity to intervene directly into the language systems of individuals, manipulating what is learned and how it is learned, in order to uncover the ingredients of successful generalisation.



Results of several experiments indicate that adults are able to discover information about the forms, meanings, and syntactic characteristics of the novel morphemes (e.g. -nule) through the training process that we employ. In fact, irrespective of a variety of manipulations to the parameters of our studies, if given sufficient time in the test task, participants are able to 'figure out' explicitly what the novel morphemes mean and the rules by which they combine with known stems, such that they can make decisions about the appropriateness of untrained exemplars (e.g. sailnule). However, when participants' knowledge is probed through speeded tasks that reflect access to stored linguistic representations, successful generalisation is constrained by three factors: consolidation, variability, and semantic consistency.



(a) Consolidation. Our studies demonstrate that participants have explicit knowledge of the novel morphemes and their characteristics immediately after training. However, participants seem unable to capitalize on this knowledge when they are assessed on speeded tasks that reflect access to stored linguistic representations until some days after training. This result suggests that the discovery of shared structure across the training set requires a period of offline consolidation, perhaps associated with overnight sleep. It is during this period of consolidation that knowledge of the novel morphemes is encoded into the long-term memory stores that support rapid generalisation.



(b) Variability. Our studies demonstrate that participants have explicit knowledge of the novel morphemes and their characteristics irrespective of the number of stems with which the novel morphemes combine during training (e.g. whether -nule occurs only with two stems or occurs with eight stems). However, generalisation in speeded tasks that reflect access to stored linguistic representations requires that the novel morphemes presented in training combine with multiple stems. This result suggests that the discovery of shared structure across the training set requires variability in initial learning.



(c) Semantic Consistency. Our studies demonstrate that participants have explicit knowledge of the novel morphemes and their characteristics irrespective of whether the novel morphemes always have the same meaning or whether there is semantic inconsistency in the training set. However, generalisation in speeded tasks that reflect access to stored linguistic representations requires that the novel morphemes have a consistent meaning during training. Our studies further show that inconsistency in initial learning can be overcome by spacing the learning of each meaning across successive days (i.e. allowing a period of overnight consolidation between the presentation of novel words comprising each meaning for a novel morpheme).



These findings together support and further constrain a theory of linguistic generalisation based on complementary learning systems in the brain, whereby rapid initial learning of episodes is followed by slower long-term encoding of the abstract knowledge that permits rapid generalisation. These findings also have important implications for the teaching of linguistic information which we are pursuing in a variety of ways.
Exploitation Route The discovery of generalised linguistic information is central to acquisition of language and literacy skills in the classroom. For example, children are exposed to a myriad of individual words in early literacy instruction, and from this exposure, derive general knowledge of the relationship between spellings and sounds (in recent years, this discovery process has been assisted by explicit phonics instruction). Our research has identified three of the key ingredients required for the acquisition of generalised knowledge: consolidation, variability, and consistency. While our studies have uncovered these principles in the domain of morphological learning, we believe that they may apply to the acquisition of generalised knowledge in any linguistic domain (e.g. reading, spelling, syntax). Thus, our findings may have important implications for the teaching of first and second languages and literacy.



Our findings suggest that if teachers want their pupils to acquire some general knowledge about the rules or statistics that govern some linguistic domain, then they need to think carefully about the nature of the items to which pupils are exposed. Pupils should be exposed to multiple exemplars of whatever rule or statistic is being taught, and there should be high consistency amongst these exemplars. For example, training of the phonic relationship [ai -> /A/ as in paid] would be best accomplished by providing multiple examples of this relationship (e.g. laid, paid, main, nail, sail, faint), and ensuring that no exceptions are included (e.g. said). Further, our research would suggest that the acquisition of generalised knowledge (in this case, that [ai sounds like /A/]) does not happen in the classroom, but rather arises after some period of offline consolidation.



We are excited about the potential uses of our research in educational settings, and have been pursuing impacts through a variety of routes, including grass-roots interaction with teachers, interaction at policy level, and through more generic outreach and media activities.
Sectors Education

URL http://www.rastlelab.com
 
Description The research conducted in this project addresses the mental and neural processes that underpin linguistic generalisation. This work constitutes basic science and we anticipate that the impacts will be largely academic. However, we have worked hard to build the pathways that would permit non-academic impacts to be realised in the mid- and longer-term. Further, we are delighted that the experimental method that we developed through this project, and our findings concerning the mechanisms that underpin linguistic generalisation, have already begun to yield impacts on academic and non-academic users. Our findings have been disseminated in three journal articles (Merkx et al., 2011; Tamminen et al., 2012; Tamminen et al., 2015) and two book chapters (Rastle & Merkx, 2011; Tamminen, 2013), one article in the publication of the Association for Language Learning (the subject association for language teachers in schools; Davis, 2013), and approximately 35 spoken presentations at scholarly conferences, departmental seminars, professional societies, public lectures and workshops, and government departments. The main academic outcome of the work- a very substantial article published in the highly-prestigious journal Cognitive Psychology - was praised by reviewers as 'immaculate', 'very interesting', and 'impressive' and continues to be one of the top downloaded articles in the journal. Finally, this project has supported the publication of an additional eight journal articles related to themes of morphological processing, reading, and lexical consolidation. These details of these outputs have been uploaded to the researchfish database and most are available on our lab website (www.rastlelab.com). From an academic perspective, citation data indicate that our work is already having impact on several areas of research including word learning, consolidation, language processing, and language acquisition. Further, several groups have requested our materials to use in their own studies involving related research questions and participant groups. In spite of the fact that our research seeks to uncover fundamental mechanisms, we are pleased that it is already beginning to have non-academic impacts. These have arisen largely through the grass-roots engagement and outreach that we have undertaken during the life of the grant and subsequent to it. These activities include: (a) a workshop on language learning that we organised for approximately 50 teachers, SENCOs, and educational professionals from local authorities in our region; (b) interaction that we have had with language teachers who are members of the Association for Language Learning; (c) interaction that we have had with members of the Association for University Language Teachers; (d) interaction that we have had with Dr Peter Bowers, international teacher trainer and founder of the WordWorks Literacy Centre in Canada; (e) interaction with the general public through participation in Science Festivals, through writing articles suitable for a general readership (including an article in The Conversation which had approximately 10,000 reads (https://theconversation.com/learning-a-language-sleep-on-it-and-youll-get-the-grammar-40605), and through public lectures; (f) interaction with publishers of 'learning to read' materials; and (g) interaction with relevant policy-makers through invitations to present to Rt Hon Nick Gibb (MP) and his Ministerial Steering Group in the Department for Education. Tangible non-academic impacts at this early stage are difficult to discern, and we would anticipate changes in educational policy or practice to arise only when there is a suitably persuasive mass of evidence relevant to a particular issue. However, we believe that our active engagement with relevant professionals and policy-makers is having an indirect impact For example, we believe that our engagement with teachers and teacher trainers is helping to inform the nature of the training that they deliver. Similarly, we believe that our engagement with language teachers and policy makers is helping to inform how languages are taught in schools. We believe that our contact with non-academic users of our research is vitally important, and will seek to continue to build on the impact pathways we have begun to develop in this project.
Sector Education
Impact Types Societal

 
Description Cognitive Neuroscience of Language Learning: Implications for Education 
Form Of Engagement Activity A formal working group, expert panel or dialogue
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Policymakers/politicians
Results and Impact Paper and discussion presented to Nick Gibb (MP) & Ministerial Steering Group, Department for Education, Whitehall, UK, 21st Feb 2012

Department for Education, Nick Gibb (MP and Minister for Education) and Ministerial Steering Group
Year(s) Of Engagement Activity 2012
 
Description Cognitive Neuroscience of Language Learning: Implications for Education 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach Regional
Primary Audience Professional Practitioners
Results and Impact Paper and discussion session provided to Literacy, Language & Communication (LiLaC) network, Royal Holloway, Egham, UK, 8th Nov 2012. This is a network of 50-or-so teachers, headteachers, SENCOs, and professionals from charity and local authorities interested in language and literacy education.

New professional connections and contacts
Year(s) Of Engagement Activity 2012
 
Description From specific examples to general knowledge in language learning 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach Local
Primary Audience Other academic audiences (collaborators, peers etc.)
Results and Impact Seminar presented at the Department of Psychology, Claremont Graduate School, California, USA

Increased academic interest in our work.
Year(s) Of Engagement Activity 2013
 
Description From specific examples to general knowledge in language learning 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Other academic audiences (collaborators, peers etc.)
Results and Impact Paper presented at the Meeting of the Psychonomic Society, Toronto.

Increased academic interest in our work.
Year(s) Of Engagement Activity 2013
 
Description From specific examples to general knowledge in language learning 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Other academic audiences (collaborators, peers etc.)
Results and Impact Paper presented at the Meeting of the British Psychological Society, Reading, UK.

Increased interest from academic colleagues
Year(s) Of Engagement Activity 2013
 
Description From specific examples to general knowledge in language learning 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach Regional
Primary Audience Professional Practitioners
Results and Impact Paper presented at the Meeting of the LiLaC Researcher-Practitioner Network, Egham, UK

New professional connections
Year(s) Of Engagement Activity 2012
 
Description From specific examples to general knowledge in language learning 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Other academic audiences (collaborators, peers etc.)
Results and Impact Seminar presented at the Forum for Research in Language and Literacy, Reading, UK.

Increased interest in our work; new connections from education
Year(s) Of Engagement Activity 2013
 
Description From specific examples to general knowledge in language learning 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach Local
Primary Audience Other academic audiences (collaborators, peers etc.)
Results and Impact Seminar presented at the Department of Psychology, University of Warwick.

Increased academic interest in our work; new academic contacts
Year(s) Of Engagement Activity 2013
 
Description From specific examples to general knowledge in language learning 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Other academic audiences (collaborators, peers etc.)
Results and Impact Paper presented at the meeting of the International Society for Morphological Processing, Cambridge UK

Increased interest in our work
Year(s) Of Engagement Activity 2013
 
Description From specific examples to general knowledge in language learning 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach Local
Primary Audience Other academic audiences (collaborators, peers etc.)
Results and Impact Seminar presented at the MRC Cognition and Brain Sciences Unit, Cambridge, UK.

Increased academic interest in our work; new cross-disciplinary connections and contacts
Year(s) Of Engagement Activity 2013
 
Description Language Learning and the Brain 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Other academic audiences (collaborators, peers etc.)
Results and Impact Cambridge University, Linguistics Society, Cambridge, UK, 28th November 2013

Increased interest in the work; new collaborations
Year(s) Of Engagement Activity 2013
 
Description Language Learning and the Brain 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Professional Practitioners
Results and Impact Association for Language Learning (ALL) annual meeting, Nottingham, UK, 22nd March 2013.

Increased interest from literacy professionals in our work; new connections
Year(s) Of Engagement Activity 2013
 
Description Learning about the 'nule' in 'sleepnule': Acquisition and abstraction of newly learnt morphemes 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach Regional
Primary Audience Other academic audiences (collaborators, peers etc.)
Results and Impact Invited talk given at: Workshop on neural mechanisms of word learning, Nijmegen, March 2013.

Increased interest in our work
Year(s) Of Engagement Activity 2013
 
Description Mechanisms of skilled reading 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach Regional
Primary Audience Public/other audiences
Results and Impact Public lecture given at Royal Holloway University of London.

N/A
Year(s) Of Engagement Activity 2012,2013
 
Description The Conversation 2015 
Form Of Engagement Activity A magazine, newsletter or online publication
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Public/other audiences
Results and Impact Wrote article for "The Conversation" titled "Learning a language? Sleep on it and you'll get the grammar"
https://theconversation.com/learning-a-language-sleep-on-it-and-youll-get-the-grammar-40605

Had around 10,000 views.
Year(s) Of Engagement Activity 2015
URL https://theconversation.com/learning-a-language-sleep-on-it-and-youll-get-the-grammar-40605
 
Description Word Learning Demonstration 
Form Of Engagement Activity Participation in an open day or visit at my research institution
Part Of Official Scheme? No
Geographic Reach Regional
Primary Audience Public/other audiences
Results and Impact Activity presented at Annual Science Open Day, Royal Holloway, University of London

N/A.
Year(s) Of Engagement Activity 2013