Linguistics Commons^™

Combining Learning Approaches For Incremental On-Line Parsing, Deryle W. Lonsdale, Michael B. Manookin

Faculty Publications

This paper discusses the integration of two different machine learning approaches to modeling language, NL-Soar and analogical modeling (AM). The resulting hybrid system is capable of functionality that is not possible when using only one of the systems in isolation. After a brief introduction of each system, an explanation is given of how AM is used to provide information useful to NL-Soar for two tasks. Examples are given, and related issues are outlined.

Connections Between Teacher And Student Attitudes Regarding Script Choice In First-Year Japanese Language Classrooms, Dan Dewey

Faculty Publications

This study investigates connections between teacher and student attitudes related to the delayed (or immediate) introduction of Japanese script in the Japanese language classroom. Two groups of students completed questionnaires concerning their attitudes toward the immediate or delayed introduction of Japanese script and the use of romaji (romanized Japanese) in the Japanese as a second language classroom. The first group was introduced to Japanese script immediately, and romaji was not used in their textbook. The second group used a romaji textbook and was introduced toJapanese script at a later time and a slower rate. Teacher attitudes were also assessed using …

The Language Contact Profile, Barbara F. Freed, Dan Dewey, Norman Segalowitz, Randall Halter

Faculty Publications

Efforts to gather data of various sorts—demographics, language-learning history, contact with native speakers, use of the language in the field—as they relate to participants in SLA research studies are inherent to understanding more about language acquisition and use. Scholars frequently develop questionnaires of their own, which are rarely shared widely in the profession. Consequently, much time and effort is invested in reinventing the process of gathering the types of data that are commonly needed.

The document we present here—a questionnaire to assess second language contact for students entering and completing language study programs in various contexts of learning (academic classrooms, …

A Computational Approach To Resolving Certain Issues In Spanish Stress Placement, David Eddington

Faculty Publications

Previous research into Spanish stress assignment suggests that accentuation involves learning and storing individual word tokens along with their inherent stress patterns, and analogizing on stored words when the need arises to determine the stress pattern of an unknown or unavailable word. Exactly what variables are most relevant to stress assignment has been hotly debated. In the present study, a number of analogical simulations were carried out to determine whether elements of the CV tier, syllable weight, phonemic representation, or a combination of these, best accounts for accentuation in Spanish.

Expanding Tree Adjoining Grammar To Create Junction Grammar Trees, Deryle W. Lonsdale, Ronald Millett

Faculty Publications

Junction Grammar (JG) combines junction operators, multiple linked syntax/semantics trees, and flexible traversal algorithms. The multiple tree and flexible ordering characteristics of MC-TAG and other TAG extensions are somewhat analogous. This paper proposes that these similarities can be integrated to form a new approach, JG-TAG. Relevant aspects of both theories and the proposed new model are discussed in turn, and representative examples are sketched.

Integrating Perception, Language And Problem Solving In A Cognitive Agent For A Mobile Robot., Deryle W. Lonsdale, D. Paul Benjamin, Damian M. Lyons

Faculty Publications

We are implementing a unified cognitive architecture for a mobile robot. Our goal is to endow a robot agent with the full range of cognitive abilities, including perception, use of natural language, learning and the ability to solve complex problems. The perspective of this work is that an architecture based on a unified theory of robot cognition has the best chance of attaining human-level performance.

This agent architecture is an integration of three theories: a theory of cognition embodied in the Soar system, the RS formal model of sensorimotor activity and an algebraic theory of decomposition and reformulation.

These three …

Adapt: A Cognitive Architecture For Robots, Deryle W. Lonsdale, D. Paul Benjamin, Damian M. Lyons

Faculty Publications

The goal of the ADAPT project is to create a robot that can model its environment accurately in real time, and use that model to perform tasks and interact with people using natural language.

Nl-Soar Update, Deryle W. Lonsdale

Faculty Publications

No abstract provided.

To Be Or Not To Be Ideophonically Impoverished, Janis B. Nuckolls

Faculty Publications

This paper addresses a current debate over the universality of ideophones, a class of expressions that are used to simulate, through performative foregrounding, the salient processes and perceptions of everyday life experience. Using data from Quechua-speaking Runa in Amazonian Ecuador, I argue for a view of ideophones as a type of cultural discourse through which speakers align themselves with nonhuman life forms and forces of nature. This alignment is suggested by the special performative properties of ideophones, which collapse the distinction between a speech event and a narrated event, thus compelling a speaker to become an action, event, or process, …

Memory-Based Tone Recognition Of Cantonese Syllables, Michael William Emonts

Theses and Dissertations

Speech recognition has only recently been applied to Cantonese. Considerable effort, however, has been spent in recognizing Mandarin, the standard dialect of Chinese. Prior to this thesis, the only published work on monosyllabic Cantonese tone recognition is from Tan Lee et al. (1993,1995). This thesis is the first of its kind in that it explores memory-based learning as a viable approach for Cantonese tone recognition. The memory-based learning algorithm employed in this thesis outperforms the highly respected and widely used neural network approach. Various numbers of tones and features are modeled to find the best method for feature selection and …

Doing Loshootseed Morphology By Analogy, Deryle W. Lonsdale

Faculty Publications

Analogical modeling (AM) is an exemplar-based general modeling theory that is being applied to an increasing range ofnatural language processing problems. This paper introduces AM as a viable approach to morphological analysis for Lushootseed, and shows results from applying the system to analyze the contents of the transcription of a well-known Lushootseed story. Subsequent discussion mentions the strengths and current weaknesses of the approach. Possible improvements and future applications are also sketched.

Automated Rating Of Esl Essays, Deryle W. Lonsdale, Diane Strong-Krause

Faculty Publications

To date, traditional NLP parsers have not been widely successful in TESOLoriented applications, particularly in scoring written compositions. Re-engineering such applications to provide the necessary robustness for handling ungrammatical English has proven a formidable obstacle. We discuss the use of a nontraditional parser for rating compositions that attenuates some of these difficulties. Its dependency-based shallow parsing approach provides significant robustness in the face of language learners’ ungrammatical compositions. This paper discusses how a corpus of L2 essays for English was rated using the parser, and how the automatic evaulations compared to those obtained by manual methods. The types of modifications …

A Memory-Based Approach To Cantonese Tone Recognition, Deryle W. Lonsdale, Michael Emonts

Faculty Publications

This paper introduces memory-based learning as a viable approach for Cantonese tone recognition. The memorybased learning algorithm employed here outperforms other documented current approaches for this problem, which is based on neural networks. Various numbers of tones and features are modeled to find the best method for feature selection and extraction. To further optimize this approach, experiments are performed to isolate the best feature weighting method, the best class voting weights method, and the best number of k-values to implement. Results and possible future work are discussed.

Nl-Soar And Lg-Soar: Ongoing Work, Deryle W. Lonsdale

Faculty Publications

Goals:

Expand Soar knowledge and explore possible uses on-campus

Provide and support an NL capability to the Soar research community

Toolkits, resources, knowledge repositories

Carry out research into the cognitive modeling of linguistic performance

Book Of Mormon Stories Diglot Reader On Computer, Neal S. Harmon

Theses and Dissertations

This report describes the design, development, and evaluation of a computer-based diglot reader of the Book of Mormon Stories of the Church of Jesus Christ of Latter-day Saints. Di means two and glot means language, thus a diglot reader combines two languages into one reader in order to teach a person to read in a new language. The program, which runs on both Macintosh and Windows computer platforms, contains fifteen chapters of the Book of Mormon Stories and introduces about four hundred Spanish words. This report includes a literature review on the diglot method and related materials, a description of …

The Effects Of Computer-Based Metacognitive Strategy Training For Adult Second Language Learners, Heidi D. Hyte

Theses and Dissertations

The purpose of this study was to investigate the effects of a metacognitive language learning strategy training program that was implemented into computer-assisted language learning (CALL) software on second language learners' independent use of metacognition and language learning strategies (LLS). Questions under investigation included what effect this metacognitive training had on learners' independent use of metacognition, the differences in use of metacognitive strategies between fast and slow language learners, and the effect of learners' perceptions of metacognition on their use of LLS and specific metacognitive strategies.

The subjects included 239 missionaries learning Spanish as a second language at the Missionary …

A Niche At The Nexus: Situating An Nlp Curriculum Interdisciplinarily, Deryle W. Lonsdale

Faculty Publications

This paper discusses the establishment and implementation of a curriculum for teaching NLP. At the core are two classes which involve some theoretical background, extensive hands-on experience with state-ofthe-art technologies, and practical application in the form of an intensive programming project. Issues involving interdisciplinary coordination, curriculum design, and challenges in teaching this discipline are discussed.

Are Plurals Derived Or Stored?, David Eddington, Patricia Lestrade

Faculty Publications

The ability to produce and comprehend human language involves at least two processes: the storage of units and the computations performed on those units. The role that storage and computation play in the processing of inflectional morphology is highly debated. At one end of the spectrum is the full storage model, which holds that all known words, whether morphologically complex or not, have separate entries in the mental lexicon (Jackendoff 1975, Manelis and Tharp 1977, Butterworth 1983, Bybee 1985, 1988, Rumelhart and McClelland 1986, McClelland 1988, Stemberger 1994, Rueckl, Mikolinski, Raveh, Miner, and Mars 1997). This position puts an enormous …

A Categorial Grammar Fragment For Lushootseed, Deryle W. Lonsdale

Faculty Publications

This paper establishes a basic framework for a Categorial Grammar (CG) description ofLushootseed morpho syntax and semantics. After a brief review of English CG principles, a preliminary attempt is made to apply these fundamentals to Lushootseed free and bound morphemes. Several sample derivations are presented. including one rather complex syntactic ambigiuty previously noted in the literature with its alternative descriptions. Once the syntactic elements have been put in place a short overview is presented of how semantic interpretation can be established in tandem with syntactic composition principles. Finally, the implementation of this framework within the Attribute Logic Engine environment is …

Why Quantitative?, David Eddington

Faculty Publications

It is generally the case that special volumes include studies on a narrowly defined topic of linguistic investigation from a single theoretical standpoint. A glance at the table of contents of the present issuem on the other hand, reveals an eclectic group of papers with topics ranging from second-language acquisition to feature theory. In addition to the variety of topics covered, it should be noted that a wide variety of theoretical frameworks is represented. The present volume contains the work of researchers who assume structuralist, generative, functionalist, and psycholinguistic perspectives. Nevertheless, the thread that ties these studies together is that …

Peppering Knowledge Sources With Salt: Boosting Conceptual Content For Ontology Generation, Deryle W. Lonsdale, Yihong Ding, David W. Embley, Alan Melby

Faculty Publications

This paper describes work done to explore the common ground between two different ongoing research projects: the standardization of lexical and terminological resources, and the use of conceptual ontologies for information extraction and data integration. Specifically, this paper explores improving the generation of extraction ontologies through use of a comprehensive terminology database that has been represented in a standardized format for easy tool-based implementation. We show how, via the successful integration of these two distinct efforts, it is possible to leverage large-scale terminological and conceptual information having relationship-rich semantic resources in order to reformulate, match, and merge retrieved information of …

A Speech Interface To Genealogical Data, Deryle W. Lonsdale

Faculty Publications

This talk presents an interface that has been developed to enable users to access genealogical information via speech. Whereas the more traditional approaches for accessing data are prevalent in today’s commercial genealogical products (e.g. windows, icons, and point-and-click methods), hands-free access to information is becoming increasingly popular. There would seem to be a potential demand for speech-based access to genealogical information, particularly among enthusiasts who might require mobile access, or those who are uncomfortable with or unable to use manual methods. Yet software to perform this function for genealogical access does not appear to be widely available yet.

An Integrated System For Processing Information From Genealogical Text, Deryle W. Lonsdale, Merrill Hutchison, Tim Richards, William Taysom

Faculty Publications

This presentation introduces an integrated software system that has been designed and implemented to allow processing of difficult text. The genealogical text subdomain often exhibits characteristics that do not occur in more general written language. The system we describe has been developed specifically to address the extraction and analysis of information from such specialized text.

A Usage-Based Simulation Of Spanish S-Weakening, David Eddington

Faculty Publications

The pronunciation of Spanish /s/ as an aspirate [h] (or its deletion) is studied in a usagebased simulation that assumes that prior linguistic experiences are stored as memory traces, and that these stored experiences, instead of generalizations abstracted from them, are the source of the knowledge speakers have of their language. This exploratory simulation was carried out using the algorithm of Analogical Modeling of Language (Skousen 1989, 1992).

Spanish Epenthesis: Formal And Preformance Perspectives, David Eddington

Faculty Publications

Epenthesis of /e/ before sC- clusters in Spanish is documented word-initially (e.g., eJfera) as well as in word-internal contexts (e.g., subestimar). Cases of alternating epenthesis also exist (e.g., arteriosclerosis I arterioeclerosis). Epe nthesis is examined in the formal frameworks of lexical phonology and optimality theory. Both formal analyses are able to account for the majority of the data presented, however, each requires some arguably ad hoc manipulation in order to do so. From a performance standpoint, it is argued that epenthesis plays no role in the processing of native Spanish words. The appearance of epenthesis in loan words (e.g., stress …

Lg-Soar: Parsing For Information, Deryle W. Lonsdale, Merrill Hutchison, Tim Richards, William Taysom

Faculty Publications

Mine content from problematic text

Address complicated linguistic issues

Output information into a usable format

Integrate components within an agent architecture

Surface Analogy And Spelling Rules In English Vowel Alterations, David Eddington

Faculty Publications

Previous psycholinguistic research into the English vocalic system suggest that the vowel alternations [ei-re, ii-£, ai-1, ow- a, u"-A], termed SPELLING RULE ALTERNATIONS, are psychologically significant for English speakers, in that they can influence the pronunciation of unfamiliar words. In this paper, the influence of these salient alternations on nonce words and neologisms is assessed. The effect of su1face analogy on the pronunciation of novel words is also examined. Surface analogy may be defined as the influence exerted by the fu lly specified surface forms of the morphemic relatives of a word. It assumes that all allomorphs of a given …

A Two-Level Implementation For Lushootseed Morphology, Deryle W. Lonsdale

Faculty Publications

This paper describes the implementatio of a computer system that processes Lushootseed word forms. Built on the two-level model and leveraging finite-state technology, the system is able both to parse surface forms to arives at the underlying morphemic decomposition, and to generate surface forms given such a representation. Components of the system are discussed including the lexicon architecture, rule formulation and specification, ans the word-structurer grammar. Issues relevant to processing Lushootseed are discussed including reduplication, allomorphic variation, inflectional and derivational affixation, and morphophonemic alternations.

An Operator-Based Account Of Semantic Processing, Deryle W. Lonsdale, C. Anton Rytting

Faculty Publications

This paper explores issues of psychological plausibility in modeling natural language understanding within Soar, a symbolic cognitive model. It focuses on constructing syntactic and semantic representations in simulated real time, with particular emphasis on word sense disambiguation (WSD). We discuss (i) what level of WSD should be modeled and (ii) how to use resources such as WordNet to inform these models. A preliminary model of coarse-grained WSD is included to show how syntactic, semantic, and other knowledge sources interact in Soar. Finally, we explore issues of interleaving, learning, and integrating other WSD approaches with Soar's native model of learning.

The Effect Of The Auap Study Abroad Experience On The Students' Cultural Perceptions, K. James Hartshorn

Faculty Publications

As the Asia University America Program (AUAP) commences its second decade, those who have closely observed this unique study abroad experience are aware of a variety of ways the students seem to benefit from their participation. Some of these benefits are academic and can be seen in the development and improvement of specific language skills, gains in standardized test scores and earned university credit applied toward graduation. However, there seem to be a variety of other valuable benefits that may be a little more difficult to measure. such benefits could include the enlargement of the students' international perspective as world …