No description available.

A knowledge of morphology can be useful for many natural language processing systems. Thus, much effort has been expended in developing accurate computational tools for morphology that lemmatize, segment and generate new forms. The most powerful and accurate of these have been manually encoded, such endeavors being without exception expensive and time-consuming. There have been consequently many attempts to reduce this cost in the development of morphological systems through the development of unsupervised or minimally supervised algorithms and learning methods for acquisition of morphology. These efforts have yet to produce a tool that approaches the performance of manually encoded systems. Here, I present a strategy for dealing with morphological clustering and segmentation in a minimally supervised manner but one that will be more linguistically informed than previous unsupervised approaches. That is, this study will attempt to induce clusters of words from an unannotated text that are inflectional variants of each other. Then a set of inflectional suffixes by part-of-speech will be induced from these clusters. This level of detail is made possible by a method known as alignment and transfer (AT), among other names, an approach that uses aligned bitexts to transfer linguistic resources developed for one language–the source language–to another language–the target. This approach has a further advantage in that it allows a reduction in the amount of training data without a significant degradation in performance making it useful in applications targeted at data collected from endangered languages. In the current study, however, I use English as the source and German as the target for ease of evaluation and for certain typlogical properties of German. The two main tasks, that of clustering and segmentation, are approached as sequential tasks with the clustering informing the segmentation to allow for greater accuracy in morphological analysis. While the performance of these methods does not exceed the current roster of unsupervised or minimally supervised approaches to morphology acquisition, it attempts to integrate more learning methods than previous studies. Furthermore, it attempts to learn inflectional morphology as opposed to derivational morphology, which is a crucial distinction in linguistics. / text

Despite a central role for morphological knowledge in supporting linguistic generalisation, the neural representations supporting its learning remain largely unexplored. This thesis addressed this gap by exploring the role of memory consolidation in morphological learning and generalisation. In three experiments, adult participants learned an artificial language in which stems (e.g. gleet, shiln) combined with plural affixes (e.g. –aff, -opp; gleetaff, shilnopp) to refer to the occupation of multiple male and female characters. Mimicking properties of morphological systems in natural languages, the plurals varied in their phonological consistency/ambiguity and type/token frequency. Two sets of plurals, distinguished by gender, were trained on two successive days. Experiment 1 revealed that generalisation to novel phonologically ambiguous forms measured on the second day showed a greater influence of token frequency for plurals trained on the previous day, suggesting overnight changes in their underlying representations. Experiment 2 examined this effect further by using fMRI to compare the neural representations underlying plurals learned on the day of scanning or on the previous day. Representational Similarity Analysis revealed increased similarity structure among high type frequency plurals and reduced similarity structure among high token frequency plurals following overnight consolidation in the left superior temporal gyrus (STG). These results are consistent with a Complementary Learning Systems (CLS) model in which overnight consolidation supports the development of overlapping representations among several items sharing the same feature (here, an affix; type frequency) and strengthens item-specific representations for frequently occurring items (token frequency). Additionally, connectivity analyses showed that the functional coupling between the left STG and the left dorsolateral prefrontal cortex was weaker for high type frequency plurals and stronger for high token frequency plurals following overnight consolidation. These results suggest that the engagement of prefrontal control processes in retrieving the newly-learned plurals is subject to overnight consolidation and sensitive to the similarity structure underlying the plurals to be retrieved. However, the overnight changes in similarity structure and functional networks observed in Experiment 2 were not mirrored by changes in generalisation to novel forms as were observed in Experiment 1. Experiment 3 aimed to address the discrepancy in consolidation-related changes in generalisation behaviour between the first two experiments. Type/token frequencies were manipulated to bias learning, consolidation, and generalisation towards high token frequency plurals. Despite this manipulation, no consolidation-related changes in generalisation were observed. Findings from all three experiments are interpreted in the context of the CLS model and a role for overnight consolidation in morphological learning and generalisation is discussed.

Impairment caused by non-fluent aphasia often results in the omission and substitution of inflectional markers. Cross-linguistic work has revealed differential patterns of aphasic impairment across languages. This study aims to determine how non-fluent aphasia is manifested in English-based Caribbean creole languages, namely Jamaican Creole English (JCE) and Guyanese Creole (GC). The use of inflectional morphology is variable in English-based Caribbean creole languages. Therefore, in aphasic creole speech, it is difficult to ascertain the status of a grammatical omission as a valid creole feature or as a sign of impairment. I argue that Seymour’s et al. (1998) contrastive-noncontrastive schema can be useful for differentiating between normal and disordered creole features. The data in this study was obtained from a creole speaker with aphasic impairments. The data was later transcribed and analyzed. The results of this study appear to suggest that grammatical markers may form a hierarchy of susceptibility to aphasic impairment. Tense, agreement and aspectual markers along with auxiliaries and copulas appear to be more susceptible to impairment in disordered creole speech than plural markers, personal and demonstrative pronouns and articles. / Thesis / Master of Science (MSc) / Non-fluent aphasia is a language disorder caused by stroke and other types of severe brain trauma. This disorder can affect an individual's ability to produce grammatical words and suffixes in different ways depending on the language that they speak. This study seeks to determine how non-fluent aphasia affects English-based Caribbean creole languages, namely Jamaican Creole English (JCE) and Guyanese Creole (GC). Grammatical suffixes are optional in English-based Caribbean creole languages; therefore, it is difficult to determine whether or not an omitted grammatical suffix is a valid creole feature or a sign of impairment. I argue that Seymour’s et al. (1998) contrastive-noncontrastive schema can be useful for differentiating between normal and disordered creole features. The data in this study was obtained from a creole speaker with aphasic impairments. The data was later transcribed and analyzed. The results of this study appear to suggest that certain grammatical markers are more prone error than others in aphasic creole speech. Markers that carry tense, agreement and aspectual information along with auxiliaries and copulas appear to be more susceptible to impairment than plural markers, personal and demonstrative pronouns and articles.

This literature-based thesis studies the phenomenon of the regular plural inside compounds according to Giegerich’s (1999) stratal model of English morphology. The strata of his model are defined by their bases: stratum 1 is root-based and stratum 2 is word-based. The model overcomes the failings associated with earlier stratal models defined by their affixes (e.g., Kiparsky, 1982). However, assigning compounding and the regular plural to the same word stratum following Giegerich’s (1999) model leaves an open question in terms of what restricts the interaction between both rules to prevent the generation of ill-formed compounds such as *toys box and *trucks driver. Another question emerges: Should the regular plural inflection be assigned to stratum 2? This question is important because the answer affects how we discuss the interaction between the regular plural and compounding. For example, how do we account for the interaction between a stratum-2 rule and a syntactic rule if we are not dealing with an interaction of two lexical rules at the same stratum? This thesis challenges the theory that inflectional morphology is separate from the lexicon (Anderson, 1988, 1982; Perlmutter, 1988) with supporting evidence from the properties of the possessive inflection. This research contributes significantly to the literature in its analysis of a number of compounds within texts extracted from books, which demonstrates that the internal regular plural morpheme has an evident semantic function that restricts it from appearing inside compounds (that is, on stratum 2 of the base-driven stratification model). The study thereby challenges Lieber and Štekauer’s (2009) view that the internal regular plural morpheme is purposeless and therefore should be regarded as a linking element. I also argue that the possessive inflection is assigned to stratum 2 and can interact with compounding to form possessive compounds, but is restricted by the semantic feature of the non-head element.

This study falls under the broad area of child language acquisition with specific focus on Shona morphosyntax. The understanding that knowledge of the nature of child language contributes to the sustainability of language acquisition matters forms the basis of the investigation. A qualitative approach is followed in the study, specifically focusing on the constraints on the development of inflectional morphemes (IMs) in the acquisition of nouns and verbs in child Shona. The study investigates the development of child Shona inflectional morphology and how morphology interacts with syntax. The constraints that operate in the acquisition of Shona are identified. The study refers to linguistic theories for an account of the development of child Shona morphosyntax. The study is based on the understanding that knowledge of the nature of child language contributes to the sustainability of language acquisition matters. The data used in this study is collected from four Shona speaking children. The ages of the children range from two years (2; 0) to three years and two months (3; 2). Two female and two male children participated in this study. The primary method of data collection used in this study is the naturalistic method, while elicitation is used to elicit plural formation. The results indicate that child Shona morphosyntax is characterized by omission of the various inflectional morphemes on nouns and verbs, while the lexical morphemes are retained. The child Shona IM is phonologically different from the adult Shona IM. This is because the children are constrained and hence use simplification strategies in order to try to be faithful to the input grammar. The noun and verb IMs are produced in the form of a reduced syllable, because the children dropped the consonant in the IM syllable and retained the vowel. The study reveals that the development of child Shona morphosyntax is based on pivotal constituencies of the sentence. These pivots are the nouns and verbs that are used by the children. The study identifies constraints that operate on the process of child Shona development as phonological, morphological, semantic, visibility and frequency constraints. The finding that is arrived at through this study is that syntax is in place before morphology. / African Languages / D. Litt. et Phil.

In Paradigm Function Morphology, it is usual to model affix position classes with an ordered sequence of inflectional rule blocks. Each rule block determines how (or whether) a particular affix position is filled. In this model, competition among inflectional rules is assumed to be limited to members of the same rule block; thus, the appearance of an affix in one position cannot be precluded by the appearance of an affix in another position. I present evidence that apparently disconfirms this restriction and suggests that a more general conception of rule competition is necessary. The data appear to imply that an affixation rule may in some cases override a rule introducing an affix occupying another, distinct position. I propose that each inflectional rule R carry two indices — the first, as usual, specifying the position of the affix introduced by R. The second, however, specifies the position(s) that R satisfies. By default, these two indices identify the same position. However, where one affix precludes another, the second index of the appearing affix specifies two affix positions: the one in which it appears and the one which it precludes. With both blocks satisfied, no other rules which fill either may be applied.

Number is a grammatical category found in nearly every language around the world (Corbett, 2000). The syntactic expression of number is referred to as grammatical number. In English and French, two number categories are in use: singular and plural. Nouns that are written more frequently in their singular form are called singular-dominant, while those that are written more frequently in their plural form are called plural-dominant. Several lexical decision and picture naming studies have found that grammatical number and noun dominance interact, resulting in a surface frequency effect for singular-dominant nouns only. Singular-dominant nouns are recognized/named significantly faster in their singular form than in their plural form, while plural-dominant nouns are recognized/named equally fast in both forms (e.g., Baayen, Burani, & Schreuder, 1997; Biedermann, Beyersmann, Mason, & Nickels, 2013; Domínguez, Cuetos, & Segui, 1999; New, Brysbaert, Segui, Ferrand, & Rastle, 2004; Reifegerste, Meyer, & Zwitserlood, 2017). The objective of this thesis is to extend our understanding of the singular-dominant noun surface frequency effect in English and French by adopting three procedures. First, advanced linear mixed modelling techniques were used to improve statistical power and accuracy. Second, the noun dominance ratio technique (Reifegerste et al., 2017) was applied to investigate whether the surface frequency effect remains significant when noun dominance was treated as a continuous variable. Third, a determiner-noun number agreement task was created to determine whether the surface frequency effect could be reproduced in a novel task. Three studies were conducted. In Study 1, two lexical decision tasks (LDTs) were conducted. Results revealed that in both English and French, singular nouns were recognized faster than plural nouns while the noun dominance effect was non-significant. The interaction between grammatical number and noun dominance was significant in French and marginally so in English. The interaction pattern was identical in both languages, singular-dominant nouns demonstrated a surface frequency effect while plural nouns did not. In Study 2, three determiner-noun number agreement tasks (NATs) were conducted. Results revealed that in both English and French, plural nouns were recognized faster than singular nouns. No other effects were significant. Incorporating irregular singular nouns (e.g., bonus) and plural nouns (e.g., mice) as foils produced the same results. In Study 3, two LDTs and one NAT were conducted. Lexical decision results revealed that in both English and French, singular nouns were recognized faster than plural nouns. However, the effects of noun collectivity and animacy were significant in English only; non-collective nouns were recognized faster than collective nouns while inanimate nouns were recognized faster than animate nouns. Number agreement results revealed that in English, plural nouns were recognized faster than singular nouns; no other effects reached significance. Taken together, my studies confirm that a strong surface frequency effect exists during visual word recognition for singular-dominant nouns. However, the surface frequency effect does not extend to the formation determiner-noun number agreement decisions, which were influenced nearly exclusively by grammatical number.

Submitted by Maike Costa (maiksebas@gmail.com) on 2016-01-07T13:13:31Z No. of bitstreams: 1 arquivototal.pdf: 1078165 bytes, checksum: c31c89090fcfb416a78c9db501f010fa (MD5) / Made available in DSpace on 2016-01-07T13:13:31Z (GMT). No. of bitstreams: 1 arquivototal.pdf: 1078165 bytes, checksum: c31c89090fcfb416a78c9db501f010fa (MD5) Previous issue date: 2014-11-18 / This study has as objective to investigate the children's sensitivity to the inflectional morphemes of Brazilian Portuguese (BP) in children between 3, 4 and 5 years old, it has taken as fundamental to the acquisition of verbal morphology of this language. Studies about this topic, already realized with children were based mainly on the production of data, which demonstrated that children produce, already very early, inflectional verbal morphemes in their language. These production data follow a pattern in which it reveals the presence of irregular verb forms with regular forms in the early stages of the process. In the second moment, around 3-4 years old, decays the presence of irregular shapes, appearing irregular verbs used as regular phenomenon known as over regularization. In the final stage, there is the return to the initial default. Our study is to investigate, from the data of understanding, if the child, in the stage that produces over regularization forms, would be sensible to the presence, in the input, to irregular verb in the over regularization form. It is assumed as a hypothesis the idea that children in the age between 3 and 5 years, even thy produce phases in the over regularization forms, they are able to identify the ungrammaticality in the these verbs in the sentence produced by someone else. For this, it was mounted a grammaticality judgment the experiment in which controlled , beyond the age of the participants , the type of verb (regular, irregular and over regularization) and invented verbs, it was divided into three lists of 30 stimuli and each one with 4 conditions . For the conditions 3 and 4, the sentences are repeated in the three lists. Thus, we have for the tree lists a total of 54 stimulus.The results support the hypothesis allow assumed here, showing a sensitivity to the grammaticality of the child's verbal forms present in the input. / Este trabalho investiga a sensibilidade infantil aos morfemas flexionais do português brasileiro (PB) em crianças com 3, 4 e 5 anos de idades, tomada como fundamental para a aquisição da morfologia verbal dessa língua. Os estudos sobre esse tópico, já realizados com crianças, basearam-se notadamente em dados de produção, os quais demonstraram que elas produzem, já desde muito cedo, morfemas flexionais verbais em sua língua. Tais dados de produção seguem um padrão no qual se revela a presença de formas verbais irregulares juntamente com formas regulares, nas etapas iniciais do processo. Num segundo momento, por volta dos 3-4 anos de idade, decai a presença de formas irregulares, surgindo verbos irregulares usados como regulares, fenômeno conhecido como superregularização. Na etapa final, observa-se o retorno ao padrão inicial. Nosso estudo vem investigar, a partir de dados de compreensão, se a criança, na fase em que produz formas superregularizadas, seria sensível à presença, no input, de formas verbais irregulares superregularizadas. Assume-se como hipótese a ideia de que crianças com idade entre 3 e 5 anos, mesmo na fase de produção das formas superregularizadas, são capazes de identificar a agramaticalidade desses verbos em sentenças produzidas por outras pessoas. Para isso, foi elaborado um experimento de julgamento de gramaticalidade no qual se controlou, além da faixa etária dos sujeitos, o tipo de verbo (regulares, irregulares e superregularizados) e verbos inventados, distribuídos em três listas com 30 estímulos e 4 condições cada uma, sendo que as sentenças da condição 3 ( verbos inventados) e da condição 4 ( verbo regular) se repetem nas três listas, somando assim, um total de 54 estímulos nas três listas. Os resultados permitem sustentar a hipótese aqui assumida, evidenciando uma sensibilidade da criança à gramaticalidade das formas verbais presentes no input.

“No aspect of Basque linguistics has received more attention over the years than the morphology of the verb.” (Trask 1981:1) The current study examines the complex morphological agreement patterns found in the Basque auxiliary verb system as a case in point for discussion of theoretical approaches to inflectional morphology. The traditional syntax-driven treatment of these auxiliaries is contrasted with an inferential, morphology-driven analysis within the Paradigm Function Morphology framework. Additionally, a computational implementation of the current analysis using the DATR lexical knowledge representation language is discussed.