Abstracs and Bios for JWLLP-23 (2017/12): The 23rd Joint Workshop on Linguistics and Language Processing
and
the Fifth International Workshop on the Linguistics of Ba
abstracts
- Ba based thinking and unconscious dimension in communicative interaction, Sachiko Ide
Little attention has been paid on the topic of the unconscious dimension of communicative interaction, while our predecessors Sapir (1949) and Jacobson (1980) wrote about it. It is because of the lack of method to investigate the unconscious dimension in communication. This presentation attempts to explore the communicative interaction by introducing ba based thinking, an innovating frame of thinking which complement the scientific way of reductionism, and discusses how sensibilities and sensitivities in communicative interaction are presented in our language practice. The data to be discussed are an utterance by a baseball player Ichiro, quantitative data gathered by the questionnaire on expressions in borrowing a pen, and the teacher-student task discourse in Japanese. It is argued that the assumption of the basho domain of self functions to express the sensibilities and sensitivities in communicative interactions which is performed unconsciously by speakers.
- Speech levels in matrix and subordinate clauses: Where Japanese and Korean meet and diverge, Kaoru Horie
Speech levels, the grammaticalized encoding of a speaker's interactive stance toward the addressee, are prominent linguistic phenomena in Japanese and Korean. Speech levels in both languages are commonly divided into non-polite and polite levels, though the numbers of distinguished levels differ between the two languages.
While speech levels in Japanese and Korean have been extensively researched in respective linguistic traditions, few comparative studies have been carried out to this day (excepting, e.g. Song 1980). This study aims to fill the gap from a functional and pragmatic perspective.
In matrix clauses, speech levels in Japanese and Korean serve important pragmatic functions utterance-finally. They do so either by being maintained throughout, thereby setting the basic interactive tone of a communicative situation (informal or formal), or by being shifted (e.g. from polite to non-polite). The latter phenomena, known as "speech level shift" or "style shifting" (Ono and Jones 2008), refer to a speaker's opting out of a hitherto maintained speech level and switching to another level. Speech level shift is not randomly exercised. Our study has shown that it serves some specific pragmatic functions, e.g. (i) soliloquy, (ii) attention to upcoming new information, (iii) sudden realization, and the (iv) declaration of one's intent.
In subordinate clauses, speech levels in Japanese and Korean present a striking contrast. Speech level suffixes in Korean are undoubtedly main clause elements. As such, they are in principle prevented from occurring in subordinate clauses of any type (relative, complement, adverbial clauses). In sharp contrast, polite speech level suffixes in Japanese (i.e. verbal -masu and nominal -desu) can readily occur in adverbial clauses of various types. Though less regularly and frequently, these suffixes, particularly verbal -masu, can also occur in relative and complement clauses when the speaker is expected to behave very politely and formally. This suggests that polite speech level suffixes in Japanese are employed in a wider range of non-matrix syntactic environments when there is interactional/sociolinguistic motivation.
The foregoing discussion indicates that speech levels in Japanese and Korean, in spite of extensive parallelism in matrix clause usage, may turn out to be rather different morpho-syntactic structures governed by different interactional/sociolinguistic conventions.
- Lexical Effects on Processing Doubly Quantified Sentences in Chinese, Yu-Yin Hsu
When one says "Every student likes a teacher," does it mean every student likes a different teacher (i.e., the surface, universal-wide, scope reading), or the same teacher (i.e, the inverse, existential-wide, scope reading)? English doubly quantified sentences are often considered to allow either the surface or the inverse scope thus denoting both universal and existential readings (e.g., May, 1985). One may wonder what factors contribute to scopal preferences in reading similar sentences in Chinese, a language that is typologically different from English.
In English, Kurtzman and MacDonal (1993) reported that active sentences prefer the surface scope. Dwivedi et al. (2010) reported that sentences are ambiguous initially, and the structural computation occurs later. Both studies showed that stimuli did not behave uniformly. Concerning Chinese sentence processing, some report that no inverse scope is available (Scontras et al., 2014); some others reported that the inverse scope is available (Zhou & Gao, 2009; Hsu & Lin, 2014), or at least initially (Su et al., 2013), but there is no consensus on what kind of condition licenses what type of scope.
Assuming that verbs within each verb type are the same, this study investigates the effects of different markers of aspectuality and of verb types, separately, and reports how theses lexical features influence the interpretation of quantifier scope during language comprehension.
- Querying Universal Dependencies: Asian languages special, Natalia Klyueva
Universal Dependencies (UD) present a huge project aiming to unify morphosyntactic annotation and to collect treebanks in many languages according the standard. The project started mainly with European languages in mind and first treebanks were for European languages. Now more and more Asian languages are included into the UD, e.g. Vietnamese, Chinese, Cantonese, Hindi, Urdu etc. The text corpora are either manually annotated or converted from the existing treebanks according to the UD standard. So far the treebanks served mainly as data for training parsers and not much linguistic research was done directly on the UD. This may be attributed to the fact that the corpus query systems for the data are not largely introduced to linguistic community and it is not straightforward to search in a mostly machine-friendly UD format called CoNLLU. Here I demonstrate several tools (NoSketch Engine, PML-TQ, TurkuNLP) to make a search over the treebanks. Examples will come from the Asian treebanks and involve morphosyntactic queries rather than lexica-based.
- Title TBA, Yasuhiro Katagiri
abstract
- What is Exceptional?, Stanley Peters
Exceptions are items in a generalization's domain of applicability that deviate from its general pattern. Exceptive expressions like English "except," "but," and "unless" claim that some number of items in a specified set, or particular ones, are exceptions. The exceptive also eliminates these exceptions from the generalization's domain of applicability. Exceptives occur with the many generalizations that are not universal, as well as with universal generalizations. Many reasons exist for highlighting certain exceptions by means of an exceptive. Exceptives do not serve always to save a generalization that has exceptions, and do not characterize which items are exceptions.
- From General back to Japanese Linguistics (Around Japanese Noun Particles), Andre Wlodarczyk
Generalisations of phenomena (discovered when studying the Japanese language and scrutinising the theories of Japanese linguistics) laid the basis for the Meta-Informative Centering (MIC) theory belonging to the General Linguistics domain. Recently, MIC has been integrated within an even more general Distributed Grammar framework (DG). The new information structure theory highlights the most prominent components of meaning as conveyed by major phrases of utterances in any language. These are the layers of meta-information, ortho-information and para-information which exhibit elements of pragmatics, semantics and ontology,respectively.
One important characteristic of the theory is that it sheds new light on predication and word order as universal problems, provided that we distinguish properly multi-modal conceptual representations from language utterances. Somehow a rebours, DG makes it possible to explain in a coherent theory some problems of the Japanese language, itself.
http://celta.paris-sorbonne.fr/DG-Biblio.html
- Major Phrases Word Order in Utterances (in the Distributed Grammar Program), Helene Wlodarczyk
The problem of the order of major phrases (subject, object and verb) in simple, declarative utterances in different types of languages is crucial for understanding the linear ordering of phrases which are centred by attention. In the Distributed Grammar framework, the two major noun phrases (subject and object) are seen not only as grammatical functions as expressed by morphological and syntactic markers, but also as expressing respectively the main (Global) and secondary (Local) centres of attention of the speaker. In a base utterance, both the subject and the predicate have the same meta-informative status: either Old or New. Two other attention-driven phrases can be added to base utterances giving rise to extended utterances in which the topic (with the Old information status) and the comment (with the New information status), on the one hand, and in which the focus (with the New information status) and the background (with the Old information status) on the other hand, form a contrast.
Introducing (a) the distinction between base and extended utterances and (b) the uniform treatment of subject, object, topic and focus as attention-driven phrases makes it possible to understand, in a more coherent and general way, the word order problem of different types of languages: ones with “rigid” word order (e.g.: English, Japanese) and others with “free” word order (e.g.: Polish, Russian). In the DG framework, word order (together with intonation, pragmatic and case markers) is a meta-informative device for building attention-driven phrases, not exclusively in extended utterances, but in base utterances too.
http://celta.paris-sorbonne.fr/DG-Biblio.html
- The quantity adjective 'mahn-' (many) in Korean, Jae-Woong Choe
In this presentation, we discuss a quantity adjective 'manh-' in Korean which translates into 'many' in English. Unlike English 'many' which is used either as a determiner or as an adjective, 'manh-' shows general distributional characteristics of an adjective. As for the semantic aspect of 'manh-' we find that it does not show some key logical properties that a quantification determiner is supposed to have, like conservativity and extension. We then address the issue of ambiguity and vagueness, checking if it is possible to specify the meaning of 'mahn-' more clearly. Considering all of these characteristics, especially its heavy dependency on the context, we propose a very simple hypothesis on the meaning of 'mahn-': 'more than the (contextually) given number'.
- When and how to use the response particles 'yes' and 'no': From a cross-linguistic perspective, Jong-Bok Kim
abstract
- How (normal) people respond to (potentially deviant) sentences is not as simple as linguists thought to be: An initial report on Acceptability Rating Database of Japanese (ARDJ) project, Kow Kuroda
We are developing a new kind of linguistic data called "acceptability rating data of Japanese (ARDJ)". It is intended to compensate our damning ignorance of what the acceptability space looks like, in the case of Japanese stimuli.
ARDJ collects the acceptability ratings by diverse real humans on a large number of sentences with varying degrees of acceptabilities, from fully acceptable to fully unacceptable ones. The product is expected to allow us to explore the possibility space of acceptability. This is the first report of an ongoing project.
Method:
For stimuli, we generated Japanese sentences with varying degrees of acceptabilities using a method called "mutation-inspired generation." In this method, (normal or possibly subnormal) sentences are modified (i) by randomly replacing either a noun (n), a verb (v) or a case-marker (p for positional) with a similar item, or (ii) by swapping a randomly selected pair of phrases (bunsetsu) of the same POS (n or v). This gave us four types, n-, v-, p-, and s-type, for mutation. We had 33 originals and 167 mutations (200 stimuli in total) for our pilot experiment.
We collected responses to these stimuli from 213 effective raters and analyzed them by applying cluster analysis (k-means method) to random samples with or without preservation of mutation type.
Results:
First, the analysis suggested that normal people do not simply classify stimuli into dichotomy: deviants or non-deviants. Rather, their responses had three dimension: normal stimuli (= without deviance) and two subtypes of deviant stimuli, though we are not yet able to characterze their diffrence properly.
Second, simple classification by mutation type , revealed no differentiation in response. We admit that this was surprising: we expected otherwise.
Third, we confirmed that deviances were brought about in two distinct dimensions, though we are not certain what they really are. Further investigation awaits this finding.
Fourth, and perhaps most interestingly, cluster analysis suggested differentiated effects of mutation types into clusters. While this runs counter to the first point, this would simply tell us that superficial classification is not informative enough.
- A corpus study of the construction of evaluative stance in Introduction in Psychology and Radiology journals, Winnie Cheng
In academic writing, which is essentially interpersonal and persuasive in nature, research writers need to strategically choose potentially evaluative or value-laden lexis to express (explicitly or implicitly) their attitudinal stance, to convey their level of commitment towards propositions, and to engage appropriately with the readers. Among the linguistic devices under this category, the reporting verb (RV) is a prominent one used by the writer to report, interpret and evaluate the literature throughout the article. The present study examines a collection of 128 “Introduction-Method-Results- Discussion”-structured empirical research articles published in leading journals with high impact factors in the disciplines of Radiology, Nuclear Medicine and Medical Imaging and Psychology. It compares the frequency, distribution, denotation and evaluative functions of reporting verbs (RV) associated with the moves in Introduction. The purpose is to compare how expert writers from the two disciplines use RVs to express their attitudinal stance, to convey their level of commitment towards propositions, and to engage appropriately with the readers.
- The Qualitative Difference among Tones: Evidence from an Acceptability Judgement of Taiwanese Tone Sandhi, Chuyu Huang, Tzu-Yin Chen, Yuki Hirose, and Takane Ito
This study aims to investigate the qualitative difference among various lexical tones in Taiwanese tone sandhi. Taiwanese tone sandhi has been hitherto viewed as a circular change in which all seven lexical tones are involved and every non-final syllable in a specific domain obligatorily undergoes a tonal change. This change occurs regardless of the phonological characteristics of the following tone. The theoretical challenge is whether tone sandhi should be analyzed as a rule (Chuang et al., 2011), an allomorph selection (Tsay and Myers, 1996), or a lexical retrieval with tonal information. The aim of the present study is two-fold: First, the issue of the nature of Taiwanese tone sandhi (rule-driven or lexical) is addressed through an acceptability judgement of different violation types in Taiwanese tone sandhi. Secondly, differences in acceptability of different types of sandhi violation are carefully examined. The results showed that: (i) The lexical retrieval analysis of Taiwanese tone sandhi is untenable, (ii) the violation involving the high-low contour was perceived significantly more unnatural than the violation involving the middle level tone and the mid-low contour, which could possibly be accounted for by the markedness hierarchy: If a word contains a tone violation, the wider markedness difference between the ill-formed tone and the base tone there is, the less acceptable the word would be.
- Phonological Integrity or Orthographic Integrity: Challenging traditional views on how sounds in a language is defined, Chu-Ren Huang
The concept of phonological encoding in current theories of language processing underlines the hypothesis of phonological integrity. Note that phonological encoding requires recognition of an inventory of sound categories (i.e. phonemes) defined by a linguistic system (i.e. phonology). This predicts the necessity of phonological adaptation as sounds outside of that phonological inventory defined for each language cannot be encoded. It also accounts for the effectiveness of language processing as encoding focuses on a small number of categories. However, in each language, there are some non-lexical sounds are clearly not in the phonological system and require recognition of additional categories. Examples of non-lexical (conversation) sounds (Ward 2006) include English clicks (Ogden 2013) or back-channel sounds (Ward 2000), and Mandarin atonal fillers (Yuan et al., 2016). In fact, there are also exceptional lexical sounds that are contextually dependent, such as onomatopoeia or other mimicry sounds. We also recently observed that the sub-system of Mandarin Alphabetic Words (MAWs) as exceptional lexical sounds (Huang and Liu 2016, Ding et al. 2017). The general agreement is that non-lexical sounds exist external to the core linguistic system. Exceptional lexical sounds, however, poses a serious dilemma. Being lexical, they should be part of the linguistic systems. However, the rules accounting for them phonologically may be incompatible with rules governing the core system. How do these two systems interact and how are potential inconsistencies resolved?
- Innovation and Abductive Reasoning, Eiichi Yamaguchi
To understand the role of "abduction" for innovation, I present the method of innovation diagram, which express innovation processes in a two-dimensional space spanned by "knowledge embodiment (deduction)" and "knowledge creation (abduction)". It is found that all innovationsoccur from a chain of "knowledge creation" and "knowledge embodiment" with the "field of resonance" as nodes of the chain. I discover a new structure of innovation, "paradigm disruptive innovation" on which abduction gives an essential role. I then discuss how abduction can be educated.
- Language Understanding and Abductive Reasoning: Abductive Reasoning in L2 English Dictation Tasks and in Interpreting L1 Japanese "NP1-no NP2" Constructions, Yasunari Harada, Sachiko Shudo and Kazuo Sakai
The adnominal particle “-no” in Japanese can attach to almost any nominal expressions to form adnominal expressions, thus “NP1-no NP2” is syntactically admissible for any combination of NP1 and NP2. Real-time interpretation of adnominal constructions, however, depends on several different and potentially conflicting factors such as lexical knowledge, understanding of how the real-world is and hearers' perception of the current context of discourse. Such differences in the source of information that are brought to the real-time resolution for plausible interpretations may affect the reaction time and "acceptability judgements" on the part of the hearers for those expressions, which may reveal how different thought processes are involved and utilized in understanding natural language interactions.
- Impact of Recording Devices on Students Performance in English Classes in Japanese Universities, Lisa Nabei, Miwa Morishita and Yasunari Harada
This study aims to investigate the effect and the impact of recording devices (camcorder) on students performance in English classes in Japanese University. It is well known that Japanese college students show general reluctance to communicate in English among themselves. When organized into groups of three or four and instructed to engage in some speaking tasks in English they easily change the language to their native language Japanese, which is to be expected in a foreign language class where all or almost all of the members share the same mother tongue. However, when each group is given an audio recorder or a camcorder to archive their oral interactions, they tend to keep the interaction going in English, even among students with low proficiency. (Harada, Morishita and Shudo, 2015) In addition, unlike their usual behaviour that students tend to remain quiet and sit still, trying to be invisible so that they won’t be called on during the class, some students are found to become active, even using body language extensively. Pedagogical implications and limitations of using recording devices in language classes will be discussed.
- Voices of a priest: Participation framework and language socialization in a household Shinto ceremony, Tomoko Endo
Variations in language use within an individual speaker have lately been attracting researchers’ attention (Iwasaki 2015). Adopting the notion of voice (Hill 1995), this study analyzes a Shinto priest’s linguistic behaviors in a household ceremony. It will be shown that the priest uses different prosodic patterns and grammar/vocabulary, depending on the roles that he plays in each stage of the ceremony. A further analytic emphasis is placed on the differences of linguistic behaviors triggered by the recipient, i.e., adults or children, and it is argued that language socialization and shift in participation framework play a crucial role in the generation of variations in the priest’s behaviors.
short biographical notes of the invited participants
- Winnie Cheng is Associate Dean of Faculty of Humanities, Professor of English and Director of Research Centre for Professional Communication in English (RCPCE) of Department of English, The Hong Kong Polytechnic University. Her research interests include pragmatics, intercultural and professional communication, corpus linguistics, conversation analysis, critical discourse analysis, critical genre analysis, discourse intonation, and English for Specific Purposes. (winnie.cheng@polyu.edu.hk)
- Jae Woong Choe is a professor of Linguistics at Korea University in Seoul. He was President of the Korean Society for Language and Information (1999-2001), served as a Steering Committee member of PACLIC (2005-2012), and was a JSPS fellow (2015). His areas of interest include Theoretical Linguistics, mainly on Semantics and Pragmatics, and Computational and Statistical Linguistics.
http://faculty.korea.ac.kr/jchoe
- Yasunari Harada is Professor and Associate Dean at Faculty of Law, Director at the Institute for Digital Enhancement of Cognitive Development and researcher at the Institute for the Study of Language and Information, Waseda University and visiting researcher at the Institute for Service Innovation Studies, Meiji University.
- Yu-Yin Hsu is Assistant Professor of Chinese linguistics in the Department of Chinese and Bilingual Studies at Hong Kong Polytechnic University. Her research interests include syntax theory, information structure, focus prosody, psycholinguistic sentence processing, pedagogical grammar, and teaching Chinese as a second/foreign language.
- Chu-Ren Huang PhD (Cornell), Chair Professor of Applied Chinese Language Studies, The Hong Kong Polytechnic University, Editor, Lingua Sinica, Studies in Natural Language Processing, The Humanities in Asia, Frontiers in Chinese Linguistics, Studies in East Asian Linguistics
- Sachiko Ide is former President of International Pragmatics Association, Professor Emeritus at Japan Women's University and Research Fellow at the Institute for the Study of Language and Information of Waseda University.
- Yasuhiro Katagiri, President of Future University of Hakodate
- Mayumi Kawamura is a language annotator, working with Yasunari Harada for the past 11 years on segmentation, transcription and behavioral annotation of Japanese EFL learners' interactional learning activities.
- Jong-Bok Kim, Professor at School of English, Director of Institute for the Stdudy of Language and Information and Dean at College of Humanities, Kyung Hee University
http://web.khu.ac.kr/~jongbok/
- Miwa Morishita, Associate Professor at Faculty of Global Communication, Kobe Gakuin University, visiting researcher at the Institute for Digital Enhancement of Cognitive Development, Waseda University and visiting researcher at the Institute for Service Innovation Studies, Meiji University
http://researchmap.jp/read0142634/?lang=english
- Stanley Peters is professor emeritus of linguistics and symbolic systems at Stanford University, where he chaired the Department of Linguistics and directed the Center for the Study of Language and Information. His research touches on mathematical properties of syntax and semantics, theoretical and descriptive grammar and meaning, spoken dialogue and computational dialogue systems, and artificial intelligent tutoring systems. Peters is a Fellow of the Linguistic Society of America. His publications include over a hundred articles and several books.
- Sachiko Shudo is Professor of English and Linguistics in Waseda University's Faculty of Law in Tokyo. She obtained her PhD in linguistics from Georgetown University in 1998. Her dissertation, The Presupposition and the Discourse Function of the Japanese Particle mo, was published in Routledge's Outstanding Dissertation in Linguistics Series in 2002. Her main research interests include pragmatic issues involving presupposition. She is currently Principal Investigator for Waseda University Grant for Special Research Projects (Continuing Research) entitled "Politeness-triggered manipulation of pragmatic constraints and its influence on semantic change."
- Makiko Takekuro is Associate Professor at the Faculty of Law at Waseda University and Principal Investigator for the Grant-in-Aid for Young Scientists (B) No.24720193, entitled, "A linguistic anthropological analysis of social interaction: The intersection of language, gesture, and environment in Ishigaki." She is editor-in-chief of the proceeding volume (to be published online soon) for the first International Workshop on Linguistics of BA held in December of 2011 at Waseda University.
- Andre Wlodarczyk, Universite Ch. de Gaulle (Lille) and Centre de Linguistique Theorique et Appliquee (CELTA) at Universite Paris-Sorbonne. His fields are Japanese linguistics, computational and theoretical linguistics, and semiotics. He is the initiator of the Meta-Informative Centering Theory (MIC) which is now part of the Distributed Grammar and Generalised Idiomatics (DGGI) Program.
https://www.researchgate.net/profile/Andre_Wlodarczyk
- Helene Wlodarczyk, Universite Paris-Sorbonne, Slavic Studies Department and Centre de Linguistique Theorique et Appliquee (CELTA). Her fields of research are Slavic and theoretical linguistics. She takes an active part in the development of both MIC and DG theories.
https://www.researchgate.net/profile/Helene_Wlodarczyk
- Dr. Eiichi Yamaguchi is Professor of Kyoto University (Graduate School of Advanced Integrated Studies in Human Survivability) and Visiting Professor of Doshisha University (Graduate School of Policy and Management).
notices
- Copyright © 2015-2017 Institute for DECODE, Waseda Universtiy, except for the titles and abstracts of the talks announced and short bios of the speakers, and organizers. All rights reserved.
- First drafted September 8th, 2015. Last revised December 17th, 2017.
- The meeting and talks announced in this web page are subject to change without prior notice. The organizers should not be held responsible for any purported or actual damages by prospective participants due to those changes.