This article offers a descriptive and analytic view of the different stages leading to the constitution of a corpus that is representative of the issues of climate and energy justice. Overall, the corpus contains over five million words and gathers reports, newsletters and web-pages dealing with the most equitable ways of moving to a low-carbon future in the aim of limiting climate change. It can be divided into six sub-corpora, according to types of discourse communities, and methods of constitution. (...) We begin by presenting the small Do It Yourself (DIY) corpora which were used as a starting point. Three discourse communities were selected to observe possible variation in their treatment of the issue: Non-Governmental Organisations (NGOs), United-Nation institutions, and the Renewable Energy Sector (RES). The sources are selected according to author, date, keywords in title. Using the concordance Antconc and WMatrix software we test the reliability of the corpora for their thematic content, terminology and lexical unit classification. Our first results enable us to confirm variation between the discourse communities. The discrepancy in sizes and the time-consuming nature of the initial DIY corpus constitution lead us to use BootCat to extend them, using keywords from the corpora as seeds to retrieve and download webpages. We thus contrast a more traditional approach to corpus building to web-as-corpus data gathering methods. We compare the results found in the BootCat corpora to test if they are as specific as those in the DIY corpora. This enables us to draw conclusions on the possible uses and advantages of relatively small corpora for the study of specialised discourse. (shrink)

In 1990, Vermeerbergen started the first larger-scale corpus study with spontaneous language data from adult signers on the morpho-syntactic aspects of Flemish Sign Language. After this, a number of lexicographic projects, including the collection of a 90-h corpus, led to the launch of the first online bilingual Dutch/VGT—VGT/Dutch dictionary in 2004. Since then, researchers have developed several corpora of variable sizes, with the greatest realization being the VGT Corpus. The main focus of this chapter is twofold. On the one (...) hand the run-up to, the development and the use of the VGT Corpus will be discussed, while on the other hand smaller specific research corpora will be highlighted such as the corpus on early parent-child interaction and the multifocal eye-tracking corpus. The current chapter will discuss the research and community value of the corpora and future directions. Finally, it will elaborate on the need for corpus research, the associated advantages and disadvantages, and the obstacles faced in smaller deaf communities. (shrink)

Cet article interroge la question des corpus en analyse multimodale des corpus oraux d’interactions filmées dans le cadre d’une linguistique qui intègre l’étude des gestes à celle du discours pour rendre compte des pratiques langagières dans leur globalité. Ce type d’approche est traditionnellement associé à des petits corpus supports d’analyses qualitatives. À l’appui d’un tour d’horizon de recherches contemporaines en analyse multimodale des interactions orales, nous montrons que l’étude linguistique des gestes se prête à une grande variété d’approches innovantes au (...) plan méthodologique. En pratique, il reste à ce jour difficile de construire des grands corpus pour l’analyse gestuelle, parce qu’il s’agit d’une entreprise longue et coûteuse (enregistrement, transcription, annotation) : les petits corpus sont nombreux parce qu’ils sont plus faciles à construire pour un chercheur seul ou une petite équipe de recherche aux moyens limités. Cette situation tend cependant à évoluer, et ce de plusieurs manières, grâce aux collaborations nationales et internationales et aux nouveaux outils numériques. Les « petits » corpus n’en restent pas moins sources de résultats précieux en analyse gestuelle. Pour diverses raisons, les standards internationaux de l’évaluation de la recherche tendent à valoriser l’analyse quantitative aux dépens de l’analyse qualitative, faisant oublier que ces deux types d’approches ont des objectifs certes différents mais complémentaires : ils permettent de répondre à des questions de recherche différentes (étroites ou larges) et de proposer des savoirs d’origines différentes (généralisabilité des résultats ou généralisabilité des processus observés), dont la légitimité est elle aussi de nature différente (réfutabilité des résultats ou consilience des experts). Nombre de travaux récents en analyse gestuelle encouragent à remettre en question la dichotomie traditionnelle entre analyses qualitatives et quantitatives, en intégrant les deux approches de diverses manières. (shrink)

Cette contribution présente des types de petits corpus très hétérogènes, qui ne sont toujours pas clairement définis : les petits corpus oraux bilingues et plurilingues. Elle envisage d’abord de donner un aperçu de petits corpus et discute l’insécurité terminologique du domaine des corpus bi- et plurilingues. Ensuite, les bases de données de petits corpus bilingues Talkbank, HZSK, Sketchengine et CLAPOTY sont brièvement décrites pour discuter les possibilités de visualisation ainsi que de mise en disposition des données de corpus. Finalement, un (...) exemple concret est présenté : le corpus FM, un petit corpus bilingue français-anglais d’une variété du français canadien, parlée au Manitoba. (shrink)

Le présent article prend pour point de départ de sa réflexion sur les enjeux épistémologiques et méthodologiques des petits corpus un constat simple, à savoir que ceux-ci ont survécu à l’heure où l’essor de la linguistique de corpus et le développement du big data ont multiplié les ressources. Cette survie est, selon nous, la preuve d’une nécessité à la fois matérielle et logistique et d’une pertinence scientifique des petits corpus que nous explorons sous l’angle particulier de la langue orale et (...) du domaine spécifique de la phonologie. Nous montrons en quoi cette discipline entretient un rapport unique à l’objet langue et donc privilégié aux corpus et repensons la validité des corpus non pas en termes de taille, mais en termes de finalité. Nous défendons alors la méthodologie et les ambitions du programme PAC dont la base de données sur l’anglais oral contemporain repose exclusivement sur de petits corpus spécialisés constitués dans les différentes aires géographiques anglophones. Nous démontrons non seulement la pertinence individuelle de ces petits corpus pour l’étude quantitative et qualitative de phénomènes précis, comme le nivellement de la variété mancunienne dans le nord de l’Angleterre ; mais également l’apport concret de la comparabilité et de la jonction possible des analyses faites à partir de ces petits corpus pour la modélisation théorique des phénomènes de rhoticité et de r de sandhi dans les variétés d’anglais parlées à travers le monde. Ainsi, nous illustrons la manière dont les résultats obtenus grâce à l’étude en profondeur des petits corpus PAC permettent de penser la dynamique des systèmes de l’anglais ainsi que le poids sociolinguistique des identités, notamment nord-sud, dans la description et la compréhension de leurs caractéristiques phonético-phonologiques. Aussi, en défendant les petits corpus, nous défendons en creux la phonologie de corpus, et nous remettons en cause la supériorité automatique des grands corpus pour ce qui est de l’analyse en profondeur de la langue orale et de l’évolution des variétés d’anglais parlées à travers le monde. (shrink)

People are using innovative internet of things technologies to gain individualised management of their type 1 diabetes. The #WeAreNotWaiting movement supports them to build their own hybrid closed loop systems and access their real time blood sugar data via any web connected device. A small number of parents in Australia use such DIY looping systems to manage their child’s type 1 diabetes, but these systems have not been approved by the Therapeutic Goods Administration in Australia, creating ethical dilemmas for (...) clinicians about how to respond to the use of medical devices that are not registered on the Australian Register of Therapeutic Goods. This article considers whether the use of DIY looping is in the best interests of the child and, if not, whether intervention in parental decision making is justified to prevent harm to the child. It addresses the ongoing duty of healthcare professionals to provide care to children who are ‘looping.’ Reference is made to findings from a study, Personalised Closed Loop Systems for Childhood Diabetes, to illustrate stakeholders’ perceptions of benefits and harms of DIY looping systems. I conclude that the decision of parents to use DIY looping technology could be considered to be in a child’s best interests, broadly defined, and falls within the Zone of Parental Discretion, however healthcare practitioners who support parents may have professional concerns in doing so. (shrink)

Nous présentons un ensemble d’observations et de repères méthodologiques issus de la constitution et de l’annotation du corpus MC4 « Modélisation Contrastive et Computationnelle des Chaînes de Coréférences », corpus de (très) petite taille qui a permis des expérimentations, des études de faisabilité puis l’élaboration d’un corpus de beaucoup plus grande taille, celui du projet Democrat « Description et modélisation des chaînes de référence : outils pour l’annotation et le traitement automatique ». Nos remarques traitent notamment de la taille d’un (...) corpus annoté et du besoin de documentation complète, même pour un très petit corpus incompatible avec des exploitations ADT et TAL. (shrink)

Applying machine learning algorithms to automatically infer relationships between concepts from large-scale collections of documents presents a unique opportunity to investigate at scale how human semantic knowledge is organized, how people use it to make fundamental judgments (“How similar are cats and bears?”), and how these judgments depend on the features that describe concepts (e.g., size, furriness). However, efforts to date have exhibited a substantial discrepancy between algorithm predictions and human empirical judgments. Here, we introduce a novel approach to generating (...) embeddings for this purpose motivated by the idea that semantic context plays a critical role in human judgment. We leverage this idea by constraining the topic or domain from which documents used for generating embeddings are drawn (e.g., referring to the natural world vs. transportation apparatus). Specifically, we trained state-of-the-art machine learning algorithms using contextually-constrained text corpora (domain-specific subsets of Wikipedia articles, 50+ million words each) and showed that this procedure greatly improved predictions of empirical similarity judgments and feature ratings of contextually relevant concepts. Furthermore, we describe a novel, computationally tractable method for improving predictions of contextually-unconstrained embedding models based on dimensionality reduction of their internal representation to a small number of contextually relevant semantic features. By improving the correspondence between predictions derived automatically by machine learning methods using vast amounts of data and more limited, but direct empirical measurements of human judgments, our approach may help leverage the availability of online corpora to better understand the structure of human semantic representations and how people make judgments based on those. (shrink)

Small corpora remain an important resource for addressing questions requiring specialized data not found in large general purpose corpora. But do these small corpora have sufficient data to analyze low-frequency phenomena? This article presents a case study of the try-and-V pseudocoordination construction in English, which has an unusual morphosyntactic restriction: neither verb can be inflected. Diachronic corpora elucidate its origin as reanalysis from ambiguous usage of non-finite verbs in the late 1500s. Dialectal corpora (...) reveal variation in frequency but not grammatical properties. And child language corpora reveal productive usage even by young children. These results are considered successful and the outlook is optimistic for addressing low-frequency phenomena beyond contemporary, standard English, especially as more, and more diverse, corpora become available. (shrink)

Considerable work during the past two decades has focused on modeling the structure of semantic memory, although the performance of these models in complex and unconstrained semantic tasks remains relatively understudied. We introduce a two‐player cooperative word game, Connector (based on the boardgame Codenames), and investigate whether similarity metrics derived from two large databases of human free association norms, the University of South Florida norms and the Small World of Words norms, and two distributional semantic models based on large (...) language corpora (word2vec and GloVe) predict performance in this game. Participant dyads were presented with 20‐item word boards with word pairs of varying relatedness. The speaker received a word pair from the board (e.g., exam‐algebra) and generated a one‐word semantic clue (e.g., math), which was used by the guesser to identify the word pair on the board across three attempts. Response times to generate the clue, as well as accuracy and latencies for the guessed word pair, were strongly predicted by the cosine similarity between word pairs and clues in random walk‐based associative models, and to a lesser degree by the distributional models, suggesting that conceptual representations activated during free association were better able to capture search and retrieval processes in the game. Further, the speaker adjusted subsequent clues based on the first attempt by the guesser, who in turn benefited from the adjustment in clues, suggesting a cooperative influence in the game that was effectively captured by both associative and distributional models. These results indicate that both associative and distributional models can capture relatively unconstrained search processes in a cooperative game setting, and Connector is particularly suited to examine communication and semantic search processes. (shrink)

We show a number of improvements in the use of Hidden Conditional Random Fields for phone classification on the TIMIT and Switchboard corpora. We first show that the use of regularization effectively prevents overfitting, improving over other methods such as early stopping. We then show that HCRFs are able to make use of non-independent features in phone classification, at least with small numbers of mixture components, while HMMs degrade due to their strong independence assumptions. Finally, we successfully apply (...) Maximum a Posteriori adaptation to HCRFs, decreasing the phone classification error rate in the Switchboard corpus by around 1% – 5% given only small amounts of adaptation data. (shrink)

By the time they reach early adulthood, English speakers are familiar with the meaning of thousands of words. In the last decades, computational simulations known as distributional semantic models have demonstrated that it is possible to induce word meaning representations solely from word co-occurrence statistics extracted from a large amount of text. However, while these models learn in batch mode from large corpora, human word learning proceeds incrementally after minimal exposure to new words. In this study, we run a (...) set of experiments investigating whether minimal distributional evidence from very short passages suffices to trigger successful word learning in subjects, testing their linguistic and visual intuitions about the concepts associated with new words. After confirming that subjects are indeed very efficient distributional learners even from small amounts of evidence, we test a DSM on the same multimodal task, finding that it behaves in a remarkable human-like way. We conclude that DSMs provide a convincing computational account of word learning even at the early stages in which a word is first encountered, and the way they build meaning representations can offer new insights into human language acquisition. (shrink)

This paper reconsiders the diphone-based word segmentation model of Cairns, Shillcock, Chater, and Levy (1997) and Hockema (2006), previously thought to be unlearnable. A statistically principled learning model is developed using Bayes’ theorem and reasonable assumptions about infants’ implicit knowledge. The ability to recover phrase-medial word boundaries is tested using phonetic corpora derived from spontaneous interactions with children and adults. The (unsupervised and semi-supervised) learning models are shown to exhibit several crucial properties. First, only a small amount of (...) language exposure is required to achieve the model’s ceiling performance, equivalent to between 1 day and 1 month of caregiver input. Second, the models are robust to variation, both in the free parameter and the input representation. Finally, both the learning and baseline models exhibit undersegmentation, argued to have significant ramifications for speech processing as a whole. (shrink)

In this article we explore the relationship between semiotic analysis of images and quantitative analysis of vast image corpora, in particular the work produced by Lev Manovich and the Cultural Analytics Lab, called “Media Visualization.” Media Visualization has been chosen as corpus because of its metavisual operation (images are visualized and analyzed by images) and its innovating way of conceiving analysis: by visual instruments. In this paper semiotics is used as an approach to Media Visualization and taken as an (...) object of study as well, especially visual semiotics. In this sense, a comparison between visual semiotics (close reading of small corpora) and quantitative analyses of images (distant reading of vast collections) are conducted from a semiotic point of view. Post-Greimassian semiotics guides this study with respect to the issue of the image-within-an-image and metavisual visualization; Peircean semiotics is employed to explain and develop the notion of diagram. (shrink)

One of the tasks faced by young children is the segmentation of a continuous stream of speech into discrete linguistic units. Early in development, syllables emerge as perceptual primitives, and the wholesale storage of syllable chunks is one possible strategy for bootstrapping the segmentation process. Here, we investigate what types of chunks children store. Our method involves selecting syllabified utterances from corpora of child-directed speech, which we vary according to (a) their length in syllables, (b) the mutual predictability of (...) their syllables, and (c) their frequency. We then use the number of utterances within which words are contained to predict the time course of word learning, arguing that utterances which perform well at this task are also more likely to be stored, by young children, as undersegmented chunks. Our results show that short utterances are best-suited for predicting when children acquire the words contained within them, although the effect is rather small. Beyond this, we also find that short utterances are the most likely to correspond to words. Together, the two findings suggest that children may not store many complete utterances as undersegmented chunks, with most of the units that children store as hypothesized words corresponding to \textit{actual} words. However, dovetailing with an item-based account of language-acquisition, when children do store undersegmented chunks, these are likely to be short sequences -- not frequent or internally predictable multi-word chunks. We end by discussing implications for work on formulaic multi-word sequences. (shrink)

In a remarkable confirmation of OT in an empirical domain for which it was not originally intended, phonological and morphological variation has been successfully modeled by partially ranked categorical constraints (Anttila 1997, 2002). Poetic meter is a good place to extend and test this approach to variation, because there is abundant and diverse quantitative data available for it, and because it is typically governed by a relatively small number of well-understood constraints. I report the results of four such studies (...) here. They confirm that choices among metrical options are governed by the interaction of partially ranked constraints, in each case constraints that are grounded, and motivated independently of variation data by related systems in which they have a fixed rank. The partially ranked constraint systems turned out to predict not only the relative preferences among metrical options, but also their actual frequencies in the corpora, with surprising accuracy. These findings support the partial ranking model of variation, and provide an explanatory benchmark beyond the reach of intrinsically weaker stochastic approaches that posit a statistical component for metrical competence (Hayes & MacEachern 1998). (shrink)

MKM has been defined as the quest for technologies to manage mathematical knowledge. MKM “in the small” is well-studied, so the real problem is to scale up to large, highly interconnected corpora: “MKM in the large”. We contend that advances in two areas are needed to reach this goal. We need representation languages that support incremental processing of all primitive MKM operations, and we need software architectures and implementations that implement these operations scalably on large knowledge bases. We (...) present instances of both in this paper: the MMT framework for modular theory-graphs that integrates meta-logical foundations, which forms the base of the next OMDOC version; and TNTBase, a versioned storage system for XML- based document formats. TNTBase becomes an MMT database by instantiating it with special MKM operations for MMT. (shrink)

Social scientists and computer scientist are divided by small differences in perspective and not by any significant disciplinary divide. In the field of text analysis, several such differences are noted: social scientists often use unsupervised models to explore corpora, whereas many computer scientists employ supervised models to train data; social scientists hold to more conventional causal notions than do most computer scientists, and often favor intense exploitation of existing algorithms, whereas computer scientists focus more on developing new models; (...) and computer scientists tend to trust human judgment more than social scientists do. These differences have implications that potentially can improve the practice of social science. (shrink)

In The Value of the Humanities prize-winning critic Helen Small assesses the value of the Humanities, eloquently examining five historical arguments in defence of the Humanities.

Tammy A. Kinsey 'Let us Never Speak of It?: Edward S. Small's _Direct Theory_' _Film-Philosophy_, vol. 7 no. 51, December 2003.

Nietzsche's thinking on justice and punishment explores the motives and forces which lie behind moral concepts and social institutions. His dialogue with several writers of his time is discussed here. Eugen Dühring had argued that a natural feeling of ressentiment against those who have harmed us is the source of the concept of injustice, so that punishment, even in its most impersonal form, is always a form of revenge. In attacking this theory, Nietzsche developed his own powerful critique of moral (...) concepts such as responsibility and guilt. He borrowed his ‘historical’ approach to moral concepts from Paul Ree, who suggested that the utilitarian function of punishment had been obscured by its practice, which appears to be directly linked with moral guilt. Nietzsche responds that punishment has quite different purposes and meanings at different times, so that any single explanation or justification is inadequate. In this way, he rejects the pre-suppositions common to the retributivists and utilitarians of his time. (shrink)

GS 341 is one of the most familiar of Nietzsche’s writings. This article proposes a new reading that stands in contrast with most English-language Nietzsche scholarship. The text presents a communication and its reception. A ‘demon’ makes an announcement, and a hearer responds in one way or another. But there is also another narrative altogether, whose conceptual vocabulary comes from a dynamic world-view. In this an interaction of forces leads to a new situation. If the hearer is not crushed by (...) the ‘greatest weight’, there must be an inner force that counteracts its impact. I argue that what Nietzsche calls amor fati is a state that allows our drives to achieve full expression, and so makes possible a collective strength able to withstand the greatest impact. The final sentence refers to a “confirmation and sealing”. What happens is that the demon’s message makes an impression on the receptive hearer. For Nietzsche this is a working metaphor, not a turn of phrase. In a draft from this period he writes: “Let us stamp the image of eternity on our life!” That sudden and forceful act, embodying an evaluative judgement, is the event that GS 341 is all about. (shrink)

Qurʾans of the Umayyads: A First Overview. By François Déroche. Leiden Studies in Islam and Society, vol. 1. Leiden: Brill, 2014. Pp. xiii + 155, illus. $71, £55.

Sociologists have identified many factors that mitigate the progressive effects of the legal mobilization to end sexual violence. Within this body of research, however, there is little interrogation about the social construction of sexual harm. I use the case of child sexual abuse to investigate how prosecutors make sense of sexual harm. Data are qualitative interviews with 43 prosecutors. Findings reveal that prosecutors use a framework of sexual identity to construct sexual injury on the child’s body. The perceived harm centers (...) on the anticipated loss of the child’s heterosexual potential. Girl victims are thought to grow into sexual promiscuity, and boy victims are thought to grow into sex offenders. Prosecutorial constructions of child sexual abuse cases are future-oriented, which increases their urgency, and these constructions also imagine the child as a person in formation, rather than a fully actualized person with intrinsic rights. In revealing how the state of sexual victimization is not only deeply gendered but also heteronormative, this research has theoretical implications for childhood studies, queer studies, and anti-violence advocacy. (shrink)

Deductive reasoning is criticized by hegel for its failure to show the purpose and necessity of its thinking. It may be acceptable in other sciences and in everyday life, But not in philosophy. Dialectical reasoning, In contrast, Is not an instrument for attaining truth but is inseparable from the development of truth itself. This argument is not a "critique of dialectical reason"; the validation of dialectic is the task of dialectic alone.

Food policy increasingly attempts to accommodate a wider and more diverse range of stakeholder interests. However, the emerging influence of different communities and networks of actors with localized concerns and interests around how food should be produced and traded, can challenge attempts to achieving more open, sustainable and globally-integrated food chains. This article analyses how cultural factors internal to a developed country can disrupt the export of food to a developing country. A framing analysis is applied to examine how activists (...) using social media to interact with the traditional news media in Australia were able to inflame public opinion and provoke outrage to disrupt the policy agenda. The paper contains a case study analysis of the media controversy in 2011 around the slaughter of beef cattle in Indonesian abattoirs and the subsequent banning of live cattle exports to Indonesia by Australia. The analysis draws on the theory of binary cultural oppositions to examine how practices in relation to the slaughter of beef cattle in Indonesia were reframed, through extensive media coverage of moral outrage into a critique of the values and cultural practices of Indonesian society. (shrink)

Smooth muscle cells have developed a contractile machinery that allows them to exert tension on the surrounding extracellular matrix over their entire length. This has been achieved by coupling obliquely organized contractile filaments to a more‐or‐less longitudinal framework of cytoskeletal elements. Earlier structural data suggested that the cytoskeleton was composed primarily of intermediate filaments and played only a passive role. More recent findings highlight the segregation of actin isotypes and of actin‐associated proteins between the contractile and cytoskeletal domains and raise (...) the possibility that the cytoskeleton performs a more active function. Current efforts focus on defining the relative contributions of myosin cross‐bridge cycling and actin‐associated protein interactions to the maintenance of tension in smooth muscle tissue. (shrink)

In 2000, the publication of ten texts marked the completion of a two-year project entitled Le Devoir de mémoire, which had brought together ten writers from across Africa to write in response to the Rwandan genocide. This article looks at how the project was posited from the outset as a specifically African response, setting this in the context of older problems of voice, selfrepresentation and the renegotiation of miswritten histories in the postcolonial context. This aspect of the project is made (...) all the more urgent by the actuality of the genocide, and the period of residence the writers spent in Rwanda in 1998. This article argues that the project succeeded in creating a space in which Africa as a whole is made part of the genocide, and vice versa, by raising complex questions of responsibility and by drawing on familiar themes in modern African writing, namely history, memory and identity; exile and dislocation. (shrink)

Some ethical dilemmas in health care, such as over the use of age as a criterion of patient selection, appeal to the notion of life expectancy. However, some features of this concept have not been discussed. Here I look in turn at two aspects: one positive — our expectation of further life — and the other negative — the loss of potential life brought about by death. The most common method of determining this loss, by counting only the period of (...) time between death and some particular age, implies that those who die at ages not far from that one are regarded as losing very little potential life, while those who die at greater ages are regarded as losing none at all. This approach has methodological advantages but ethical disadvantages, in that it fails to correspond to our strong belief that anyone who dies is losing some period of life that he or she would otherwise have had. The normative role of life expectancy expressed in the ‘fair innings’ attitude arises from a particular historical situation: not the increase of life expectancy in modern societies, but a related narrowing in the distribution of projected life spans. Since life expectancy is really a representation of existing patterns of mortality, which in turn are determined by many influences, including the present allocation of health resources, it should not be taken as a prediction, and still less as a statement of entitlement. (shrink)