More Like this

1. Mermaid Constructions in Lexical Functional Grammar

   Chihiro Taguchi (January 2022, Lexical-Functional Grammar Conference)

   This paper provides a cross-linguistic analysis of Mermaid Constructions in terms of Lexical Functional Grammar. Mermaid Constructions, coined by Tsunoda (2020), are grammaticalized monoclausal constructions in which a verb and a noun, sometimes with a copula, form a compound predicate. However, the work was chiefly descriptive, and the morphosyntactic nature of Mermaid Constructions in theoretical terms has not yet been explained. In this work, in opposition to Tsunoda’s (2020) hypothesis, the following points are shown: (1) Mermaid Constructions are not monoclausal but biclausal; (2) Mermaid Constructions do not comprise a compound predicate, but are control and raising with a nominal predicate; (3) these findings hold cross-linguistically. 

   more » « less
2. Understanding the Polarity of Events in the Biomedical Literature: Deep Learning vs. Linguistically-informed Methods

   https://doi.org/10.18653/v1/W19-2603  

   Noriega-Atala, Enrique; Liang, Zhengzhong; Bachman, John; Morrison, Clayton; Surdeanu, Mihai (June 2019, Proceedings of the Workshop on Extracting Structured Knowledge from Scientific Publications)

   An important task in the machine reading of biochemical events expressed in biomedical texts is correctly reading the polarity, i.e., attributing whether the biochemical event is a promotion or an inhibition. Here we present a novel dataset for studying polarity attribution accuracy. We use this dataset to train and evaluate several deep learning models for polarity identification, and compare these to a linguistically-informed model. The best performing deep learning architecture achieves 0.968 average F1 performance in a five-fold cross-validation study, a considerable improvement over the linguistically informed model average F1 of 0.862. 

   more » « less
3. Language classification in Western Amazonia: advances in favor of the Pano-Takana Hypothesis

   https://doi.org/10.20396/liames.v23i00.8670150  

   Valenzuela, Pilar; Zariquiey, Roberto (January 2023, LIAMES: Línguas Indígenas Americanas)

   The languages of the Pano and Takana families exhibit a considerable number of lexical and structural affinities that cannot be ascribed to mere chance and are not readily detectable instances of borrowing. After the comparative studies by Key (1968) and Girard (1971) the proposal of a genetic relationship between these two families was generally accepted (e.g. Loos 1973, 2005; Suárez 1973; Kaufman 1990; Campbell 1997). Without solid argumentation, however, this classification was later put into question (Fabre 1998; Loos 1999; Fleck 2013) and, even today, there is no full consensus as to whether the observed similarities are due to genetic inheritance or long-term language contact. The present paper offers lexical and grammatical evidence in support of the hypothesis that Pano and Takana are genetically connected. Comparing for the first time what can be considered Proto-Pano and Proto-Takana reconstructions, it is shown that 18 of the 40 items in the basic vocabulary list proposed by the Automated Similarity Judgment Program (asjp) (Holman et al. 2008) might be cognate; this includes 9 body-part terms. Also, a set of alleged grammatical cognates are assembled, and shared constructions involving motion verbal morphology, intransive and transitive auxiliaries, transitivity harmony restrictions, and switch-reference are discussed. 

   more » « less
4. Are cross-linguistically frequent systems easier to learn? The case of evidentiality.

   Saratsl, Dionysian; Bartell, Stefan; Papafragou, Anna (July 2019, Proceedings of the Annual Conference of the Cognitive Science Society)

   It is often assumed that cross-linguistically more prevalent distinctions are easier to learn (Typological Prevalence Hypothesis - TPH). Prior work supports this hypothesis in phonology, morphology and syntax but has not addressed semantics. Using an Artificial Language Learning paradigm, we explore the learnability of semantic distinctions within the domain of evidentiality (i.e. the linguistic encoding of information sources). Our results support the TPH, since the most prevalent evidential system was learned best while the most rare evidentiality system yielded the worst learnability results. Furthermore, our results indicate that, cross-linguistically, indirect information sources seem to be marked preferentially (and acquired more easily) compared to direct sources. We explain this pattern in terms of the pragmatic need to mark indirect, potentially more unreliable sources over direct sources of information. The Document 

   more » « less
5. Linguistically-driven Framework for Computationally Efficient and Scalable Sign Recognition

   Metaxas, Dimitris; Dilsizian, Mark; Neidle, C. (May 2018, Language Resources and Evaluation Conference 2018)

   We introduce a new general framework for sign recognition from monocular video using limited quantities of annotated data. The novelty of the hybrid framework we describe here is that we exploit state-of-the art learning methods while also incorporating features based on what we know about the linguistic composition of lexical signs. In particular, we analyze hand shape, orientation, location, and motion trajectories, and then use CRFs to combine this linguistically significant information for purposes of sign recognition. Our robust modeling and recognition of these sub-components of sign production allow an efficient parameterization of the sign recognition problem as compared with purely data-driven methods. This parameterization enables a scalable and extendable time-series learning approach that advances the state of the art in sign recognition, as shown by the results reported here for recognition of isolated, citation-form, lexical signs from American Sign Language (ASL). 

   more » « less