Search Results

A fully lexicalized grammar can represent the knowledge of a multilingual speaker simply by putting lexical items from the various languages together. This conception suggests that multilingualism should be a quite natural state, an idea that fits well with a conception according to which every adjustment in “register” or dialect for context is regarded as a kind of codeswitching (CS). This conception of CS does not invoke any special mechanisms to control the interaction among the languages a learner knows. Nonetheless, some restrictions appear to be needed, since CS does not occur at arbitrary positions in an utterance. Exploring these restrictions, MacSwan (1999, 2014) proposes that the syntactic elements of various languages can mix freely, but the morphophonologies have different properties and cannot be mixed. Hence, CS is possible only at those points in an utterance that would not disrupt morphophonological dependencies. A computational model of language analysis proposed by Stabler (2001), which separates morphophonological and syntactic processing, is explicit and general enough to allow these CS proposals to be realized. This chapter shows such a computational model at work, concluding that (1) when the syntax is lexicalized, so all crosslinguistic variation is due to the lexicon, CS is immediately predicted for mixed utterances; (2) the extent of codeswitching in the syntax is based on the extent to which the categories of the different languages are assimilated, which allows for the influence of learned CS patterns; and (3) morphophonology, based on linear position and ranked constraints, does not allow word-internal CS.

This chapter describes the detailed relationship software and language; it states that software processes the natural language and constructs in programming languages. The chapter explores the in-depth analysis of software implementation languages and software written languages, and provides a discussion on the role of computer control languages, machine languages, common human languages, and different language variants. It concludes with the differentiation between the computer programming languages and coding concepts in the computing sector.

Computational approaches in spatial language understanding nowadays distinguish and use different aspects of spatial and contextual information. These aspects comprise linguistic grammatical features, qualitative formal representations, and situational context-aware data. In this chapter, we apply formal models and machine learning techniques to map spatial semantics in natural language to qualitative spatial representations. In particular, we investigate whether and how well linguistic features can be classified and automatically extracted and mapped to region-based qualitative relations using corpus-based learning. We separate the challenge of spatial language understanding into two tasks: (i) we identify and automatically extract those parts from linguistic utterances that provide specifically spatial information, and (ii) we map the extracted parts that result from the first task to qualitative spatial representations. In this chapter, we present both tasks and we particularly discuss experimental results of the second part of mapping linguistic features to qualitative spatial relations. Our results show that region-based spatial relations can indeed be learned to a high degree and that they are distinguishable on the basis of different linguistic features.

Combinatory Categorial Grammar (CCG) allows semantically equivalent alternate surface derivations typified by the sentence “Harry admires Louise,” as well as English noun phrases, to have all of the type-raised categories allowed by a full-blown morphological case system. Many critics have argued that this so-called spurious ambiguity makes CCG quite impracticable to apply to useful tasks such as parsing and question answering in open domains, regardless of its linguistic attractions. This chapter examines how CCG can be used for efficient natural language processing. It first considers algorithms that have formed the basis of a number of practical CCG parsers before turning to logical forms and how they are built with CCG. The chapter also discusses processing scope and pronominal reference in CCG, generation of strings from logical forms using CCG, the use of scope for rapid inference in support of question answering or textual entailment, and human sentence processing.

Computational linguists and psycholinguists show little interest in the formal semantics of quantification, even though quantifiers and other scoping elements play an important role in human and computational natural language processing. Quantifiers, along with negation and other categories such as modality, tense, and aspect, matter because they support inference and entailment. To understand the relevance of quantifier-scope, it is necessary to look at how rare events reveal the nature of the system and where quantifers might actually have practical applications. Questions provide few opportunities to deploy quantifiers, other than simple definites and indefinites. Some quantifiers convey subtle implicatures about the completeness or otherwise of the speaker’s knowledge.

A review of statistical facts of language especially morphology and children’s acquisition of morphology with focus on productivity. Contrary to popular beliefs, productivity should be understood as a categorical notion in language, judging from the now extensive cross-linguistic studies of language acquisition. Why language must make use of a small set of wide-ranging rules, rather than memorized expressions, and why this is a difficult task for the child learner who acquires language in a few short years.

Natural language processing (NLP) is the field of decoding human written language. This chapter responds to the growing interest in using machine learning–based NLP approaches for analyzing open-ended employee survey responses. These techniques address scalability and the ability to provide real-time insights to make qualitative data collection equally or more desirable in organizations. The chapter walks through the evolution of text analytics in industrial–organizational psychology and discusses relevant supervised and unsupervised machine learning NLP methods for survey text data, such as latent Dirichlet allocation, latent semantic analysis, sentiment analysis, word relatedness methods, and so on. The chapter also lays out preprocessing techniques and the trade-offs of growing NLP capabilities internally versus externally, points the readers to available resources, and ends with discussing implications and future directions of these approaches.

In this chapter, two signed language technologies are introduced: signed language recognition and signed language synthesis/animation. The modality-specific challenges of these technologies are discussed, originating in the multichannel nature of signs and the lack of a standardized writing system. The state of research of each technology is outlined, demonstrating that the existing body of research is considerably smaller than that of the field of automatic spoken language processing. The chapter describes the combination of the two signed language technologies in real-world applications. Most importantly, it explores the potential application of each technology to second language (L2) signed language assessment. Finally, an example of an existing use case is given: the application of signed language recognition to a vocabulary test for adult L2 learners of Swiss German Sign Language.

This chapter discusses state-of-the-art applications of covariate shift adaptation techniques to various real-world problems. It covers non-stationarity adaptation in brain-computer interfaces; speaker identification through change in voice quality; domain adaptation in natural language processing; age prediction from face images under changing illumination conditions; user adaptation in human activity recognition; and efficient sample reuse in autonomous robot control.

In this dialogue, Bernardo Cuenca Grau, a computer scientist at the Department of Computer Science, University of Oxford, begins by explaining his research in technology based on ontologies and knowledge representation, somewhere between mathematics, philosophy, and computer science. He goes on to argue why we need to represent knowledge in a way that it can be processed by a computer and therefore enable automated reasoning of this knowledge using artificial intelligence. Later he explains how his investigation probes the limits of mathematics to find the most appropriate languages for developing practical applications. For example, the large-scale processing of structured information linked to comprehensive health systems. Bernardo is supportive of collective tools such as Wikipedia. He also discusses why in his opinion the success of a scientific or technological idea depends very much on luck, and why the semantic web has not been defined. Furthermore, he argues why bureaucracy confuses process with progress.

This chapter shows how probabilistic topic modeling drives interest in the study of topics, and how topic modeling has proven a successful tool for identifying coherent linguistic categories within collections of texts. Yet despite interest in topic models, no one has yet asked the question “What is a topic?” (either in classical rhetoric or computational study). If we derive large-scale semantic significance from texts, how does this relate to the longer philosophical/philological tradition? Beginning with an overview of the long (pre-computational) history of topics (from Aristotle to Renaissance commonplace books to nineteenth-century encyclopedism), then moving to a quantitative approach to topic modeling’s link to the past, the chapter uses a single topic (a single model run on a collection of German novels written over the course of the long nineteenth century) to explore larger metaphorical constellations associated with this topic (through the close reading of individual passages), and to apply a more quantitative approach (chapter 2’s method of multidimensional scaling, where word distributions are transformed into spatial representations). The topic’s semantic coherence (when it’s more or less present) at different states of likelihood within a text is compared to the spatial relationships and interconnectedness of topics (the way they coalesce/disperse).

The author and his family went back to Europe for good, and had to acculturate to a new culture. His boys needed a bit of help during their first years but then things worked out well. The author set up his laboratory and started collaborating with firms involved in natural language processing (NLP). One of the main projects he worked on with his team was an English writing tool and grammar checker for French speakers. He also developed a long-term partnership with the Lausanne University Hospital (CHUV). The author explains how he helped students do experimental research in his laboratory. The chapter ends with some statistics showing how successful the laboratory was over a span of twenty years.

Quantity has a role to play in understanding the nature of characters and the process of characterization (the writerly act of generating animate entities through language). With Alex Woloch’s question of “the many” in mind, the chapter begins with a survey of an estimated 85 characters per novel in the nineteenth century, a conservative estimate of 20,000 novels published during this period in English, producing ca. 1.7 million unique characters appearing in one century in one language. Simultaneously, the process of characterization poses challenges of scale: the great number of characters, plus the vast amount of information surrounding even a single, main character. Characters (like other textual features) are abundant across the pages of novels. Through an examination of over 650,000 characters using new techniques in natural language processing and entity recognition, this chapter explores the "character-text" of novels (how characters are activated, described, objectified). The (surprising) evidence here suggests that the process of characterization is best described as one of stylistic constraint, aligning the practice of characterization more closely with a character’s etymological origins (as representative, general, or “characteristic”—not individualistic). The chapter then explores the rise of “interiorly oriented” characters and Nancy Armstrong’s notion of strongly gendered “deep character.”

This book studies the meaning of repetitions of reading instead of its more singular moments--the quantities of texts and the quantities of things within texts--that give meaning to a reader's experience. The introduction establishes a new theoretical framework for thinking about literary quantity through three key terms: implication, distribution, and diagram. First, implication puts the idea of modeling and its contingency into the foreground of research. Focusing on the implicatedness of computational modeling allows a rethinking of our investments in either purely empirical or subjective reading experiences. Implicatedness acknowledges both the constructedness of our knowl­edge about the past and an underlying affirmatory belief contained in any construction of the world. Distributed reading, by contrast, suggests a fundamentally relational/reflexive way of thinking about literary meaning, the way the sense of a text is always mutually produced through the construction of context. Distributional models allow for a more spatial, contingent modeling of texts and contexts. Finally, diagrammatic reading indicates how the practice of computation produces meaning by "drawing together" different sign systems (letter, number, image). Replacing the haptic totality of the book--its graspability--computational reading relies on the diagram's perspectival totality, the continual mediation between letters and numbers.

This chapter is about words, not in the individual sense, but in the distributional sense (a larger set of patterns/behaviors, as a form of usage). Contra a single luminous word, distributional semantics shows relationships existing between words; meaning shaped through probabi­listic distributions. Understanding­ texts as word distributions, a way of thinking about plot (the way actions/beliefs are encoded in narrative form), tracks the shift/drift of language in a text as it signals to readers a change in the text’s concerns. Using a trilingual collection (450 mostly canonical novels from the long nineteenth century), the chapter shows these novels as distinctive in their lexical contraction. Though for much of their history novels have been imagined as an abun­dant, often exceedingly long form, multiplying dramatically over time, vector space model techniques show these novels pushing against their perceived history of imagined excess. These novels are unique in how the linguistic “space” within them contracts as they explore social constraint experienced through language. Here the art of lack offers insights into what it means to contract in­ward and, in so doing, potentially saying more. The art of lack is the dream of insight where there is increasingly less and less to say.

The “corpus” (or body of work, since Cicero) of an author is meant to be organic, integral--well connected--but also distinct and whole; it marks limits; it is the material complement to the author’s life. What does it mean to imagine writing as a body, something with a distinct shape/form, but also subject to vulnerability? How do we understand those moments when a writer opens herself or her corpus up to change, how radical or gradual are these movements, how permanent, fleeting, or even recurrent? Is there something called “late style,” a distinctive signature that characterizes the end of a career as contours of an aging body mapped onto the weave of writing? When and how do we intellectually/creatively exfoliate? While we have very successful ways of detecting “authors” or “style,” we have considerably fewer techniques for talking about change, the nature of the variability within an author’s corporal outline, the variety of measures to study the shape of a writer’s career. Working with a trilingual collection of roughly 30,000 poems in French, German and English, this chapter explores questions of local/global vulnerability and late style, concluding with a computationally informed reading of the work of Wanda Coleman.

Instruction modeling is still in its early stages. This chapter discusses promising directions in which instruction modeling could develop in coming years. This includes increasing the richness of interfaces used in instruction modeling programs (e.g., by allowing students to enter responses in free form and have them graded via natural language processing); applying instruction modeling to subjects beyond mathematics, including English, foreign language, and science; using educational data mining to create automated “coaches” to help teachers better implement instruction modeling programs in their classrooms; creating approaches to instruction modeling that allow for rapid authorship of content; redesigning schools (in schedules as well as architecture) to optimize the use of instruction modeling; and putting in place government policies to encourage the use of comprehensive blended learning programs (such as those developed through instruction modeling).