The main aim of this thesis is to develop a new account of concepts and word meaning which provides a fully adequate basis for inferential accounts of linguistic communication, while both respecting philosophical insights into the nature of concepts and cohering with empirical findings in psychology on memory processes. In accord with the ‘action’ tradition in linguistic theorising, I maintain that utterance/speaker meaning is more basic than sentence meaning and that the approach to word meaning that naturally follows from this is ‘contextualism’. Contextualism challenges two assumptions of the traditional ‘minimalist’ approach to semantics: (i) that semantics (rather than pragmatics) is the appropriate locus of propositional content (hence truth-conditions); and, (ii) that words contribute stable, context-independent meanings to the sentences in which they appear. I set out two stages in the development of an adequate contextualist account of utterance content. The first provides an essential reformulation of the early insights of Paul Grice by demonstrating the unavoidability of pragmatic contributions to truth-conditional content. The second argues that the ubiquity of context-dependence justifies a radically different view of word meaning from that employed in all current pragmatic theorising, including relevance theory: rather than words expressing concepts or encoding stable meanings of any sort, both concepts and word meanings are constructed ad hoc in the process of on-line communication/interpretation, that is, in their situations of use. Finally, I show how my account of word meaning is supported by recent research in psychology: context-dependence is also rampant in category and concept formation, and multiple-trace memory models show how information distributed in memory across a multitude of previous occasions of language use can come together to build an occasion-specific word meaning, thereby bypassing the need for fixed word meanings.

In an age where technological advancements are providing people with new forms of communication, or increasing the communicative potential of forms previously available, translation is an activity which is growing more and more complex and cannot be accounted for in linguistic terms only. Translation Studies has traditionally dealt with meaning as a linguistic product; however, source texts nowadays very often include resources like images and/or sounds, which interact with the linguistically communicated message, considerably affecting meaning. More accurately, it can be said that linguistic, visual and aural meaning influence each other and create a multimodal message whose interpretation requires different types of literacy and the ability to combine them. Appropriate models analysing multimodal texts, however, are still missing. Furthermore, as no area of translation has been left untouched by the multimodal phenomenon, future translators need to be competent ‘readers’ of multimodal texts. However, the theoretical resources available to train translators are mostly concerned with texts in which the message is communicated verbally; this creates a gap between translation theory and practice as well as a gap between the training translators receive and the reality of the translation industry they need to face, in which translators find themselves working on texts where the message is communicated by more than ‘just’ words. Addressing these gaps, the main aim of this work is to develop a new model for source text analysis for translation purposes. The model brings together aspects of meaning production as it is viewed in Pragmatics, Multimodality, Translation and Semiotics and merges them in a single theoretical framework that can be applied to the analysis of any multimodal source text in order to gain a better understanding of how it conveys meaning. The model aims to contribute to a better general understanding of meaning not just as a linguistic, but as a multimodal product and it is also proposed as a theoretical resource for trainee translators.

This thesis provides insights into the effect of sociophonetic variation on speech production and speech perception among language users in a recently multilingual classroom environment in Northern Ireland. It does so in order to investigate how adolescent children acquire and adapt to the variable phonological input to which they are exposed. While research has been conducted on the role that sociolinguistic and sociophonetic variation pray in SLA for other varieties of English, little research has been conducted on the effect that NIE sociophonetic and phonological variation might have on newcomers' L2 acquisition in the NI classroom. The current study has attempted to bridge that gap. The linguistic and social factors relevant to the study are investigated using a novel research framework that combines statistical, impressionistic, instrumental and quasi-experimental techniques adapted from methodologies in the fields of sociophonetics, variationist SLA, socioperception, and accent recognition. The findings reveal that the study's participants were sensitive to NIE phonological variation and that they evaluated it in relation to their own native phonological systems. It was also found that the newcomer pupils acquired phonological variants typical among their native English-speaking peers, and that their acquisition was most consistent for NIE vowel variants. Further, comments made by the schoolchildren implied that they shared an appreciation for linguistic diversity and language learning, but that the newcomer pupils had experienced difficulties communicating with peers and teachers in the school environment. However, the findings provide evidence that sociophonetic variation has a limited effect on mutual intelligibility, and suggest that perceptions of language proficiency are more fluid than the dichotomous notions of "native" and "non-native" speaker might suggest.

The representation of word meaning has received substantial attention in the psycholinguistic literature over the past decades, yet the vast majority of studies have been limited to words referring to concrete objects. The aim of the present work is to provide a theoretically and neurally plausible model of lexical-semantic representations, not only for words referring to concrete objects but also for words referring to actions and events using a common set of assumptions across domains. In order to do so, features of meaning are generated by naïve speakers, and used as a window into important aspects of representation. A first series of analyses test how the meanings of words of different types are reflected in features associated with different modalities of sensory-motor experience, and how featural properties may be related to patterns of impairment in language-disordered populations. The features of meaning are then used to generate a model of lexical-semantic similarity, in which these different types of words are represented within a single system, under the assumption that lexical-semantic representations serve to provide an interface between conceptual knowledge derived in part from sensory-motor experience, and other linguistic information such as syntax, phonology and orthography. Predictions generated from this model are tested in a series of behavioural experiments designed to test two main questions: whether similarity measures based on speaker- generated features can predict fine-grained semantic similarity effects, and whether the predictive quality of the model is comparable for words referring to objects and words referring to actions. The results of five behavioural experiments consistently reveal graded semantic effects as predicted by the feature-based model, of similar magnitude for objects and actions. The model's fine-grained predictive performance is also found to be superior to other word-based models of representation (Latent Semantic Analysis, and similarity measures derived from Wordnet).

This thesis explores the role of previously acquired languages in third language (L3) acquisition, providing evidence that L3 transfer is determined by the underlying structural similarities and differences between previously acquired languages and the target L3. The analysis proposes a ‘feature-based’ approach to L3 transfer studies, highlighting the importance of linguistic features and the way in which they are assembled in different languages. I examine the acquisition of null and overt arguments by L1 English-L2 Spanish-L3 Chinese earners [+SP], conducting a comparative analysis with a group of L1 English-L2 non-null subject language-L3 Chinese learners [-SP]. The three languages of the [+SP] participants are not typologically related, but the L1 and L2 have similarities related to subject pronouns in the L3. Null subjects are allowed in Spanish and Chinese (although there are syntactic differences) and overt subjects behave in a similar way in English and Chinese. Therefore, transfer can occur from the L1 or L2. A Written Production Task (WPT) and a Pronoun Interpretation Task (PIT) test the use and interpretation of arguments, and a Language Relations Questionnaire (LRQ) explores the learners’ perceptions of the relatedness between languages. The results show transfer from L1 English (overt pronouns) and L2 Spanish (null subjects), indicating that L3 transfer can occur from either the L1/L2, depending on the property being acquired (i.e. it is selective). Furthermore, the data shows that the [+SP] group outperform the [-SP] group with null subjects, indicating that reassembling features associated with null subjects is straightforward despite syntactic differences between the languages. In addition, learners’ perceptions of language relatedness do not play an important role for typologically unrelated languages. Therefore, the study concludes that future L3 transfer studies can make more specific predictions regarding the source of transfer if linguistic features are taken into account.

Speaker identification (SID) aims to identify the underlying speaker(s) given a speech utterance. In a speaker identification system, the first component is the front-end or feature extractor. Feature extraction transforms the raw speech signal into a compact but effective representation that is more stable and discriminative than the original signal. Since the front-end is the first component in the chain, the quality of the later components is strongly determined by its quality. Existing approaches have used several feature extraction methods that have been adopted directly from the speech recognition task. However, the nature of these two tasks is contradictory given that speaker variability is one of the major error sources in speech recognition whereas in speaker recognition, it is the information that we wish to extract. In this thesis, the possible benefits of adapting a biologically-inspired model of human auditory processing as part of the front-end of a SID system are examined. This auditory model named Auditory Image Model (AIM) generates the stabilized auditory image (SAI). Features are extracted by the SAI through breaking it into boxes of different scales. Vector quantization (VQ) is used to create the speaker database with the speakers’ reference templates that will be used for pattern matching with the features of the target speakers that need to be identified. Also, these features are compared to the Mel-frequency cepstral coefficients (MFCCs), which is the most evident example of a feature set that is extensively used in speaker recognition but originally developed for speech recognition purposes. Additionally, another important parameter in SID systems is the dimensionality of the features. This study addresses this issue by specifying the most speaker-specific features and trying to further improve the system configuration for obtaining a representation of the auditory features with lower dimensionality. Furthermore, after evaluating the system performance in quiet conditions, another primary topic of speaker recognition is investigated. SID systems can perform well under matched training and test conditions but their performance degrades significantly because of the mismatch caused by background noise in real-world environments. Achieving robustness to SID systems becomes an important research problem. In the second experimental part of this thesis, the developed version of the system is assessed for speaker data sets of different size. Clean speech is used for the training phase while speech in the presence of babble noise is used for speaker testing. The results suggest that the extracted auditory feature vectors lead to much better performance, i.e. higher SID accuracy, compared to the MFCC-based recognition system especially for low SNRs. Lastly, the system performance is inspected with regard to parameters related to the training and test speech data such as the duration of the spoken material. From these experiments, the system is found to produce satisfying identification scores for relatively short training and test speech segments.

This thesis examines children’s argument structure overgeneralisation errors (e.g. *Don’t giggle me!). Errors of this kind arise from children observing that certain verbs can appear in more than one argument structure (e.g. The ball rolled/Homer rolled the ball). This pattern can be usefully generalised to allow children who have heard a verb produced in only one of these structures (e.g. The window opened) to produce it in the other (e.g. Marge opened the window). The ability to generalise patterns to new items is key to children becoming productive language users. However, if they overgeneralise this pattern, errors will result: Bart giggled is grammatical, but *Lisa giggled Bart (meaning Lisa made Bart giggle) is not. This thesis tested three hypotheses designed to explain how children retreat from such overgeneralisation errors, or, indeed, avoid making them altogether: the semantic verb class hypothesis (Pinker, 1989); the entrenchment hypothesis (Braine & Brooks, 1995); and the preemption hypothesis (Goldberg, 1995). Chapter 3 uses a novel-verb grammaticality judgment paradigm to investigate overgeneralisation errors in the locative construction (e.g. *Marge filled tea into the cup). Chapter 4 investigates overgeneralisation errors in the transitive and intransitive constructions, using a grammaticality judgment paradigm with known verbs, as well as a production priming paradigm designed to elicit errors from young children (e.g. *Homer swam the fish). Finally, in order to investigate the role of semantics in language development more generally, Chapter 5 moves beyond overgeneralisation errors to investigate children’s acquisition of the passive construction (e.g. Bart was helped by Lisa). This thesis adds to a growing body of work demonstrating that none of the individual theories (semantics, entrenchment, preemption) alone is able to explain children’s retreat from overgeneralisation, and that an integrated approach, such as that proposed by Ambridge and colleagues’ FIT account, is required to account for the data. The thesis moves our understanding forward by demonstrating both that this account can explain error patterns in production, and that the role of verb-in-construction semantic compatibility (a key aspect of the FIT account) can explain children’s acquisition of argument structure more widely.

Technology discloses man’s mode of dealing with Nature, the process of production by which he sustains his life, and thereby also lays bare the mode of formation of his social relations, and of the mental conceptions that flow from them (Marx, 1990: 372) My thesis is a Sociological analysis of UK policy discourse for educational technology during the last 15 years. My framework is a dialogue between the Marxist-based critical social theory of Lieras and a corpus-based Critical Discourse Analysis (CDA) of UK policy for Technology Enhanced Learning (TEL) in higher education. Embedded in TEL is a presupposition: a deterministic assumption that technology has enhanced learning. This conceals a necessary debate that reminds us it is humans that design learning, not technology. By omitting people, TEL provides a vehicle for strong hierarchical or neoliberal, agendas to make simplified claims politically, in the name of technology. My research has two main aims: firstly, I share a replicable, mixed methodological approach for linguistic analysis of the political discourse of TEL. Quantitatively, I examine patterns in my corpus to question forms of ‘use’ around technology that structure a rigid basic argument which ‘enframes’ educational technology (Heidegger, 1977: 38). In a qualitative analysis of findings, I ask to what extent policy discourse evaluates technology in one way, to support a Knowledge Based Economy (KBE) in a political economy of neoliberalism (Jessop 2004, Fairclough 2006). If technology is commodified as an external enhancement, it is expected to provide an ‘exchange value’ for learners (Marx, 1867). I therefore examine more closely what is prioritised and devalued in these texts. Secondly, I disclose a form of austerity in the discourse where technology, as an abstract force, undertakes tasks usually ascribed to humans (Lieras, 1996, Brey, 2003:2). This risks desubjectivisation, loss of power and limits people’s relationships with technology and with each other. A view of technology in political discourse as complete without people closes possibilities for broader dialectical (Fairclough, 2001, 2007) and ‘convivial’ (Illich, 1973) understandings of the intimate, material practice of engaging with technology in education. In opening the ‘black box’ of TEL via CDA I reveal talking points that are otherwise concealed. This allows me as to be reflexive and self-critical through praxis, to confront my own assumptions about what the discourse conceals and what forms of resistance might be required. In so doing, I contribute to ongoing debates about networked learning, providing a context to explore educational technology as a technology, language and learning nexus.

This thesis is based on fieldwork carried out in Guyana, South America, between 1999 and 2002, looking at the use of English in negotiating situations between the indigenous Makushi communities of the North Rupununi Savannah, the Government of Guyana, and non-governmental and international development organisations. It considers notions of liminality and third-space encounters as they relate to the evolution within the Discourse of Development of fora that temporarily accommodate both indigenous and external modes of discourse, so opening up for debate the ideologies behind them while fostering the interactive development of a third mode capable of expressing relevant aspects of both ideological systems simultaneously and so capable of expanding each individually. The need for such a hybrid space is placed in relation to the material situation of the Makushi people and wider socioeconomic and cultural issues of power, ideology and the limits of agency. The ongoing development of this space is illustrated through detailed analysis of linguistic interaction within the specific forum of the North Rupununi District Development Board. A theory is developed for relating micro and macro issues of discourse and power and a two-way relationship established between top-down pressures towards conformity and bottom-up processes of agentive change. This framework is then related to the current subordinate position of Makushi communities within Guyana and their existing achievements in developing hybrid discourses to challenge this situation. The thesis concludes with a consideration of possible applications within the existing development context in the North Rupununi, within international development in general, and within the education system.

No description available.