Synchronic morphological parsing and suffixal productivity in the Hittite adjective: Considerations on some illustrative examples

This work presents the results of a masked lexical decision experiment in which we explore the morphological parsing of Spanish suffixed or pseudosuffixed words through the suffix priming effect. Priming the bases or pseudobases with their suffixed or pseudosuffixed forms is the standard process in experiments aimed at understanding the processes underlying morphological parsing in visual word recognition with masked priming lexical decision (e.g., darkness-DARK; corner-CORN). We, however, compare the effect of suffix priming on the lexical decision of suffixed (ero-JORNALERO) and pseudosuffixed words (ero-CORDERO), as well as the effect of orthographic priming on nonsuffixed words (eba-PRUEBA). The results show that in the case of suffixed and pseudosuffixed words, related primes (ero-JORNALERO; ero-CORDERO) significantly accelerated response latencies in comparison to unrelated primes (ista-JORNALERO; ura-CORDERO), while for simple words there was no facilitation from the orthographically related prime in comparison to the unrelated prime (eba-PRUEBA; afo-PRUEBA). These results are consistent with the so-called morpho-orthographic segmentation process in the course of visual word recognition, which might also be independent of orthographic and purely semantic factors. Our results also support the view that morphological parsing takes place regardless of whether a stem is present in a word. These results complement findings from studies dealing with CORNER- and BROTHEL-like stimuli.

Morphological analysis is a very critical issue especially for natural language processing related tasks on agglutinative languages. This study gives the implementation details of a rule-based morphological analyzer of Kazakh language which is an agglutinative language. A detailed computational analysis of Kazakh language morphology such as formalization of alternation and morphotactic rules for Kazakh language is worked out in order to create the morphological analyzer. In the implementation of the morphological analyzer, alternation and morphotactic rules of Kazakh language are represented by two-level morphology rules and Foma finite state compiler is employed. This is the first detailed computational analysis of Kazakh language from morphological view. A word can have more than one morphological parse but only one of its morphological parses is valid in a given sentence. A morphological disambiguator disambiguates words by selecting one of possible parses of words. In this paper, we also present a transformation-based morphological disambiguator for Kazakh language and it is a variation of Brill tagger.

One of the major problems one is faced with when decomposing words into their constituent parts is ambiguity: the generation of multiple analyses for one input word, many of which are implausible. In order to deal with ambiguity, the MOR-phological Parser MORPA is provided with a probabilistic context-free grammar (PCFG), i.e. it combines a conventional context-free morphological grammar to filter out ungrammatical segmentations with a probability-based scoring function which determines the likelihood of each successful parse. Consequently, remaining analyses can be ordered along a scale of plausibility. Test performance data will show that a PCFG yields good results in morphological parsing. MORPA is a fully implemented parser developed for use in a text-to-speech conversion system.

Morphologically rich languages seem to benefit from joint processing of morphology and syntax, as compared to pipeline architectures. We propose a graph-based model for joint morphological parsing and dependency parsing in Sanskrit. Here, we extend the Energy based model framework (Krishna et al., 2020), proposed for several structured prediction tasks in Sanskrit, in 2 simple yet significant ways. First, the framework's default input graph generation method is modified to generate a multigraph, which enables the use of an exact search inference. Second, we prune the input search space using a linguistically motivated approach, rooted in the traditional grammatical analysis of Sanskrit. Our experiments show that the morphological parsing from our joint model outperforms standalone morphological parsers. We report state of the art results in morphological parsing, and in dependency parsing, both in standalone (with gold morphological tags) and joint morphosyntactic parsing setting.

Morphological disambiguation is the task of determining the contextually correct morphological parses of tokens in a sentence. A morphological disambiguator takes in a set of morphological parses for each token, generated by a morphological analyzer, and then selects a morphological parse for each, considering statistical and/or linguistic contextual information. This task can be seen as a generalization of the part-of-speech (POS) tagging problem, for morphologically rich languages. The disambiguated morphological analysis is usually crucial for further processing steps such as dependency parsing. In this chapter, we review the morphological disambiguation problem for Turkish and discuss approaches for solving this problem as they have evolved from manually crafted constraint-based rule systems to systems employing machine learning.

Morpheme-based reading aloud: Evidence from dyslexic and skilled Italian readers

The morphology of the fairy tale marks the dawn of structural analysis of folktales. Although it has been vastly applied in recent years to written genres, we shall go back to the origins and discuss universal and specific traits of the Japanese fairy tale via morphological parsing. The objective of this essay is to examine if, and to what extent, Propp’s morphology can be applied to Japanese fairy tales and thus find out whether there is a solution for some dilemmas in the Japanese folktale research environment. For this purpose, a group of tales on the marriage between a man and a supernatural/animal bride, considered by many Japanese and non-Japanese folklorists as representative of the Japanese fairy tale heritage, are examined. The tale of the Crane-wife serves as a model case study, which is parsed to morphologically support two of the most popular theories about the Japanese tale, namely the theory about the circular plot development by Toshio Ozawa and the result of nothingness by Hayao Kawai. The study provides a possible explanation for the difficulties in perception of Japanese fairy tales abroad. In the broader sense, the results of the study suggest that, on a morphological level, the Japanese fairy tale possesses universal features.

Large annotated datasets in NLP are overwhelmingly in English. This is an obstacle to progress in other languages. Unfortunately, obtaining new annotated resources for each task in each language would be prohibitively expensive. At the same time, commercial machine translation systems are now robust. Can we leverage these systems to translate English-language datasets automatically? In this paper, we offer a positive response for natural language inference (NLI) in Turkish. We translated two large English NLI datasets into Turkish and had a team of experts validate their translation quality and fidelity to the original labels. Using these datasets, we address core issues of representation for Turkish NLI. We find that in-language embeddings are essential and that morphological parsing can be avoided where the training set is large. Finally, we show that models trained on our machine-translated datasets are successful on human-translated evaluation sets. We share all code, models, and data publicly.

No abstract available.

This study presents results from a cross-modal priming experiment investigating inflected verb forms of German. A group of late learners of German with Russian as their native language (L1) was compared to a control group of German L1 speakers. The experiment showed different priming patterns for the two participant groups. The L1 German data yielded a stem-priming effect for inflected forms involving regular affixation and a partial priming effect for irregular forms irrespective of stem allomorphy. By contrast, the data from the late bilinguals showed reduced priming effects for both regular and irregular forms. We argue that late learners rely more on lexically stored inflected word forms during word recognition and less on morphological parsing than native speakers.

We discuss here the potential use of a morphologically aware spelling corrector to aid second language learners in finding words in an electronic dictionary, using Korean as the language of instantiation but with a method extensible to many languages. The tool discussed is a composition of two parts: a spelling corrector and morphological analyzer. When these two parts are composed, it allows for the spelling correction of morphologically complex words. The results include correct spellings as well as one or more possible morphological parses, which can aid the student both in their spelling and in their understanding of the components of the words. It can be particularly useful when students hear a word (rather than constructing it themselves) and may, therefore, not know the correct spelling or all the morphemes in it. The performance of this method is tested with data obtained from a Korean student learner error corpus and then compared with that of another publicly available morphologically aware spellchecker—the online Pusan National University spelling and grammar checker. Testing and findings reveal that traditional spellcheckers will not find many student errors, because they are not designed to; the word they typed is a correct word, but the incorrect one for the context. Providing students with multiple search results at all times allows them to choose the correct one for their context, and doing so allowed our tool to offer spelling correction, where a more traditional spellchecker did not.

In recent years, the need for Machine Translation (MT) has grown, especially for translating legal contracts between languages like Arabic and English. This study primarily investigates whether Google Translator can adequately replace human translation for legal documents. Utilizing a widely popular free web-based tool, Google Translate, the research method involved translating six segments from various legal contracts into Arabic and assessing the translations for lexical and syntactic accuracy. The findings show that although Google Translate can quickly produce English-Arabic translations, it falls short compared to professional translators, especially with complex legal terms and syntax. Errors can be categorized into: polysemy, homonymy, legal doublets, and adverbs at the linguistic level, and morphological parsing, concord, and modality at the syntactic level. The study concludes with recommendations for enhancing machine translation systems and suggests caution in using Google Translate for legal purposes, advocating for continued reliance on human expertise in legal settings.

This article presents a selective overview of studies that have investigated how advanced adult second language (L2) learners process morphologically complex words. The studies reported here have used different kinds of experimental tasks (including speeded grammaticality judgments, lexical decision, and priming) to examine three domains of morphological processing (regular and irregular inflection, derived word forms, and morphosyntactic phenomena) in L2 learners from typologically different first language (L1) backgrounds. The results from these studies demonstrate clear differences between native and nonnative processing in all three domains, indicating that adult L2 learners are less sensitive to morphological structure than native speakers and rely more on lexical storage than on morphological parsing during processing.

How the brain encodes morphological constraints during Chinese word reading: An EEG-fNIRS study

Linguistic structures exhibit a rich array of global phenomena, however commonly used Markov models are unable to adequately describe these phenomena due to their strong locality assumptions. We propose a novel hierarchical model for structured prediction over sequences and trees which exploits global context by conditioning each generation decision on an unbounded context of prior decisions. This builds on the success of Markov models but without imposing a fixed bound in order to better represent global phenomena. To facilitate learning of this large and unbounded model, we use a hierarchical Pitman-Yor process prior which provides a recursive form of smoothing. We propose prediction algorithms based on A* and Markov Chain Monte Carlo sampling. Empirical results demonstrate the potential of our model compared to baseline finite-context Markov models on three tasks: morphological parsing, syntactic parsing and part-of-speech tagging.KeywordsStructured predictionInfinite markov modelChinese restaurant process