Learning Effective and Interpretable Semantic Models using Non-Negative Sparse Embedding

In this paper, we introduce an application of matrix factorization to produce corpus-derived, distributional
models of semantics that demonstrate cognitive plausibility. We find that word representations
learned by Non-Negative Sparse Embedding (NNSE), a variant of matrix factorization, are sparse,
effective, and highly interpretable. To the best of our knowledge, this is the first approach which
yields semantic representation of words satisfying these three desirable properties. Though extensive
experimental evaluations on multiple real-world tasks and datasets, we demonstrate the superiority
of semantic models learned by NNSE over other state-of-the-art baselines.

Documents and Dependencies: an Exploration of Vector Space Models for Semantic Composition

In most previous research on distributional semantics, Vector Space Models (VSMs) of words are built either from topical information (e.g., documents in which a word is present), or from syntactic/semantic types of words (e.g., dependency parse links of a word in sentences), but not both. In this paper, we explore the utility of combining these two representations to build VSM for the task of semantic composition of adjective-noun phrases. Through extensive experiments on benchmark datasets, we find that even though a type-based VSM is effective for semantic composition, it is often outperformed by a VSM built using a combination of topic- and type-based statistics. We also introduce a new evaluation task wherein we predict the composed vector representation of a phrase from the brain activity of a human subject reading that phrase. We exploit a large syntactically parsed corpus of 16 billion tokens to build our VSMs, with vectors for both phrases and words, and make them publicly available.

Discriminating Taxonomic Categories and Domains in Mental Simulations of Concepts of Varying Concreteness

Most studies of conceptual knowledge in the brain focus on a narrow range of concrete conceptual categories, rely on the researchers' intuitions about which object belongs to these categories, and assume a broadly taxonomic organization of knowledge. In this fMRI study, we focus on concepts with a variety of concreteness levels; we use a state of the art lexical resource (WordNet 3.1) as the source for a relatively large number of category distinctions and compare a taxonomic style of organization with a domain-based model (associating concepts with scenarios). Participants mentally simulated situations associated with concepts when cued by text stimuli. Using multivariate pattern analysis, we find evidence that all Taxonomic categories and Domains can be distinguished from fMRI data and also observe a clear concreteness effect: Tools and Locations can be reliably predicted for unseen participants, but less concrete categories (e.g., Attributes, Communications, Events, Social Roles) can only be reliably discriminated within participants. A second concreteness effect relates to the interaction of Domain and Taxonomic category membership: Domain (e.g., relation to Law vs. Music) can be better predicted for less concrete categories. We repeated the analysis within anatomical regions, observing discrimination between all/most categories in the left middle occipital and temporal gyri, and more specialized discrimination for concrete categories Tool and Location in the left precentral and fusiform gyri, respectively. Highly concrete/abstract Taxonomic categories and Domain were segregated in frontal regions. We conclude that both Taxonomic and Domain class distinctions are relevant for interpreting neural structuring of concrete and abstract concepts.

National identity in a foreign context: Irish women accounting for their children’s national identity in England

Social psychologists have attempted to capture the ideological quality of the nation through a consideration of its taken-for-granted quality, whereby it forms an unnoticed ‘banal’ background to everyday life and is passively absorbed by its members in contrast to its ‘hot’, politically created and contested nature. Accordingly, national identity is assumed to be both passively absorbed from the national backdrop and actively acquired through national inculcation. This raises the question of how national identity is expressed, transmitted and acquired in a foreign context, where the banal national backdrop is unavailable to scaffold identity and the national resources for identity transmission may be unavailable. The present article addresses this gap by examining the situation of Irish women raising children in England. Critical discursive analyses of the 16 interviews revealed that all women treated their children’s national identity and the issue of transmitting identity as dilemmatic: passive transmission risks children passively absorbing English, but active transmission contravenes the assumed naturalness of national identity and can furthermore conflict with children’s own personal choice. These results point to the complex interaction between the management of national identity and the broader personal and national context within which this occurs.

'You' and 'I' in university seminars and spoken learner discourse

You and I may be little words but they do a great deal. In spoken discourse they reference shared knowledge and mark stance. In pedagogical contexts, they maintain relations in teacher-student discourse. However, language classrooms may rarely explore this array of pragmatic meanings. A lack of awareness of the variety of these functions may be problematic for learners when seeking to construct interpersonal relations and operate successfully in particular spoken contexts. This paper presents a study of you and I in two spoken corpora: a corpus of English language learner task talk and a corpus of university seminar talk. Findings illustrate different patterns of I and you between the two corpora: I and you have a higher rate of occurrence in learner discourse, and pronoun repetition is more frequent in learner discourse, though it does not account for the higher rate of you and I. These findings suggest that language learner task talk displays more features tied to speech production and self-regulation and fewer features associated with attempting to point to the informational space of others, a key feature of university classroom talk. This paper concludes by outlining pedagogical applications to overcome features perceived as disfluent.

Scales and Hierarchies: A Cross-Disciplinary Perspective

The volume advances our understanding of the role of scales and hierarchies across the linguistic sciences. Although scales and hierarchies are widely assumed to play a role in the modelling of linguistic phenomena, their status remains controversial, and it is these controversies that the present volume tackles head-on.