A carregar...
Publicação
O papel da estrutura morfológica nos processos de leitura de palavras
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 7.94 MB | Adobe PDF |
Autores
Resumo(s)
O presente trabalho enquadra-se no domínio dos estudos de psicolinguística sobre leitura de palavras. Neste domínio, a atenção tem-se centrado, sobretudo, nos aspetos gráficos, fonológicos e semânticos das palavras. Também existem estudos que consideram aspetos morfológicos, aqueles a que o presente trabalho se dedica, uma vez que os modelos teóricos do reconhecimento visual das palavras não respondem, satisfatoriamente, a todas as questões que a observação empírica dos dados permite formular.
Um resultado comummente reportado em estudos que utilizam o paradigma de priming morfológico é a facilitação do reconhecimento da palavra quando esta é precedida por uma outra palavra com a qual está morfologicamente relacionada. O problema desta hipótese de facilitação da leitura, nestas circunstâncias, vem do tipo de relação morfológica considerada, que se manifesta também na sobreposição ortográfica, na proximidade fonológica e no nexo semântico. Pretendemos, assim, encontrar evidências de que o processamento morfológico não só é fulcral no reconhecimento visual da palavra, como decorre de forma independente do processamento ortográfico, fonológico e semântico. Procurámos também documentar essas evidências em dois momentos distintos da cronologia do uso da língua, em crianças a frequentar o quarto ano de escolaridade e em jovens adultos, estudantes universitários.
Uma outra questão que decidimos tornar central relaciona-se com o facto de muitos dos modelos de reconhecimento visual, assim como uma grande maioria dos estudos que lhes estão associados, terem sido desenvolvidos sobre dados do Inglês, língua que apresenta uma alta irregularidade ortográfica, com diversas inconsistências e complexidades, e uma morfologia habitualmente caracterizada como pobre. Neste estudo, propomos a análise de dados do Português, cujo sistema morfológico é geralmente considerado como mais rico e onde os dados poderão diferir dos já encontrados na literatura. O conhecimento do sistema morfológico do Português, numa abordagem global e sistemática parece-nos essencial como base da observação do processamento morfológico das palavras. Por essa razão, adotámos o quadro de análise morfológica de Villalva (1994, 2000, 2003, 2008) para estabelecer as condições morfológicas que pretendíamos analisar, nomeadamente, palavras derivadas por sufixação, e uma distinção entre palavras derivadas que têm uma estrutura composicional e palavras derivadas lexicalizadas.
Para o presente trabalho foram delineadas três experiências. Na primeira (Capítulo 6), foi utilizado o paradigma de priming associado a uma tarefa de decisão lexical onde testamos dois sufixos de nominalização: um afixado a um radical nominal (–eiro) e o outro afixado a um radical verbal (–dor). O desenho experimental permitiu testar três condições de pares de palavras (prime/target): (i) palavras morfologicamente relacionadas pela partilha do mesmo sufixo (e.g. mineiro/carteiro); (ii) palavras aparentemente relacionadas (e.g. mineiro/poleiro); e (iii) palavras não relacionadas (e.g. mineiro/ caneta). O prime foi exposto durante 50 ms e 150 ms, imediatamente seguido pelo alvo, que esteve disponível até a decisão lexical ser tomada. Para cada tempo de exposição ao prime foram utilizadas amostras distintas de crianças a frequentar o quarto ano de escolaridade, e adultos, estudantes universitários. Os resultados mostraram que tanto crianças como adultos apresentam maior tempo de reação nos alvos morfologicamente relacionados comparativamente aos alvos aparentemente relacionados e alvos não relacionados. Os resultados não vão ao encontro do que é referido na literatura, dado que os sujeitos mostraram tempos de reação mais elevados quando existe relação morfológica entre prime e alvo. Não obstante, os resultados mostram indícios de segmentação morfológica e de identificação dos sufixos independentemente da composicionalidade morfológica das palavras. Na segunda experiência (Capítulo 7), continuamos a utilizar o método de priming morfológico associado à tarefa de decisão lexical, constituindo três condições experimentais: (i) palavras derivadas com estrutura composicional (e.g. desejo/desejoso); (ii) palavras derivadas lexicalizadas com alomorfia no sufixo (e.g. luxo/luxuoso); (iii) palavras derivadas lexicalizadas com alomorfia na base (e.g. água/aquoso). Como prime foi apresentada a forma derivante, que é uma palavra simples, e como alvo, palavras derivadas. Utilizaram-se três tempos de exposição ao prime (50 ms, 100 ms e 150 ms), sendo que, uma vez mais, para cada tempo de exposição foram constituídas amostras distintas de crianças e adultos. Os resultados obtidos confirmam que o processamento das palavras derivadas é sensível à sua estrutura morfológica e mostram também que as estruturas composicionais acarretam menores custos de processamento do que as estruturas que contêm variação alomórfica. Os resultados foram consistentes em todos os tempos de exposição ao prime. Na terceira experiência (capítulo 8) utilizamos a tarefa de decisão lexical sobre cinco condições experimentais, resultantes de palavras utilizadas como alvo nas experiências anteriores: (i) palavras simples (e.g. janela); (ii) palavras derivadas composicionais (e.g. desejoso); (iii) palavras derivadas lexicalizadas sem relação com uma forma de base, (e.g. poleiro); (iv) palavras derivadas lexicalizadas com alomorfia do sufixo (e.g. luxuoso); e (v) palavras derivadas com alomorfia da base (e.g. aquoso). A amostra foi constituída por um grupo de crianças e um grupo de adultos. Os resultados mostraram que o acesso lexical é sensível a propriedades morfológicas das palavras, sendo que as palavras com estrutura composicional apresentam menores tempos de resposta do que as palavras simples ou do que as palavras lexicalizadas com alomorfia da base e do sufixo. A composicionalidade parece, portanto, ser um fator facilitador no acesso lexical.
A análise dos resultados obtidos nas três experiências permitem-nos admitir que o processamento visual de palavras complexas derivadas por sufixação, no Português, é sensível a propriedades da sua estrutura morfológica. Por um lado, verificou-se que as estruturas composicionais têm um custo de processamento menor do que as estruturas perturbadas por fatores como a alomorfia do sufixo ou a alomorfia da base, tanto quando são primadas como quando não o são. Por outro lado, verificou-se que quando primamos sufixos, o reconhecimento visual não é facilitado.
The present work is a psycholinguistic study with is focus on word reading. In this field, attention has been focused on graphic, phonological and semantic aspects. Regarding morphological aspects, the theoretical models of visual word recognition do not satisfactorily respond to all the questions that empirical observation of the data allows to formulate. A commonly reported result in studies that use morphological priming paradigm is the facilitation of word recognition when it is preceded by another word with which it is morphologically related. The problem of this hypothesis of reading facilitation, in these circumstances, comes from the type of morphological relation considered, which also manifests itself in orthographic overlap, in phonological proximity and in semantic nexus. We intend, therefore, to find evidence that morphological processing is not only central in the visual word recognition, but also occurs independently of orthographic, phonological and semantic processing. We also sought to document this evidence at two distinct times in the chronology of language use: in children, attending the fourth grade; and in young adults, college students. Another issue that we have decided to make central is related to the fact that many of visual recognition models, as well as a large majority of the studies associated with them, have been developed on data from English, a language with a high orthographic irregularity, with various inconsistencies and complexities, and a poor morphology. In this study, we propose the analysis of Portuguese data, whose morphological system is generally considered to be richer and where data may differ from the one already found in the literature. The knowledge of the morphological system of Portuguese in a global and systematic approach seems essential to us as the basis for the observation of morphological processing of words. For this reason, we adopted the morphological analysis of Villalva (1994, 2000, 2003, 2008) to establish morphological conditions that we wanted to analyse, namely derived words and a distinction between derived words that have a compositional structure and words that are lexicalized. For the present work were delineated three experiments. In the first one we used the priming paradigm associated to a lexical decision task. We tested two nominalization suffixes: one affixed to a nominal radical (–eiro) and the other affixed to a verbal theme (–dor). The experimental design allowed us to test three experimental conditions of prime/target word pairs: (i) morphologically related (e.g. mineiro 'miner' / carteiro 'mailman'); (ii) with apparently related words (e.g. mineiro 'miner' / poleiro 'perch'); and (iii) unrelated words (e.g. mineiro 'miner' / caneta ‘pen’). The prime was exposed for 50 and 150 ms, immediately followed by the target, which was available until the lexical decision was made. For each time of exposure, different samples of children and adults were used (the children attending the fourth grade and adults were undergraduate students). The results showed that both children and adults have longer reaction times on morphologically related targets compared to apparently related targets and unrelated targets. The data do not match with the one mentioned in the literature, since the subjects showed higher reaction times when there is a morphological relationship between the primes and the targets. Nevertheless, the results show that, for Portuguese language, there are indications of morphological segmentation, as well as identification of the suffixes independently of the morphological compositionality of the words. In the second experiment, we continued to use the morphological priming method associated with the lexical decision task, constituting three experimental conditions: (i) compound derived words (e.g. desejo 'desire' / desejoso 'desiring'); (ii) lexicalized derived words with suffix allomorphy (e.g. luxo 'luxury' / luxuoso 'luxurious'); and (iii) lexicalized derived words with base allomorphy (e.g. água 'water' / aquoso 'aqueous'). As the prime we always presented the simple word and as a target, the derived words. Three prime exposure times (50, 100 and 150 ms) were used, and once again, different samples of children and adults were tested for each exposure time. The results confirm that derived words processing is sensitive to morphological structure and also show that compositional structures involve lower processing costs than the structures that contain allomorphic variation. The results were consistent at all prime exposure times. In the third experiment we used a lexical decision task on five experimental conditions, which resulted from the words used as targets in previous experiments: (i) simple words (e.g. janela 'window'); (Ii) compositional derived words (e.g. desejoso 'wishful'); (iii) lexicalized derivative words with total allomorphy (e.g. poleiro 'perch'); (iv) lexicalized derived words an allomorphic suffix (e.g. luxuoso 'luxurious'); and (v) lexicalized derived words with radical allomorphy (e.g. aquoso 'aqueous'). The sample consisted in a group of children and a group of adults. The results showed, once again, that lexical access is sensitive to the words morphological properties, and the words with compositional structure have shorter response times than the simple structures or than the lexicalized structures with radical or suffix allomorphy. Compositionality, therefore, appears to be a facilitating factor in lexical access. The results obtained in the three experiments allow us to construct the hypothesis that the visual processing of complex derived words in Portuguese is sensitive to the morphological structure properties. On the one hand, it has been found that compositional structures have a lower processing cost than structures disturbed by factors such as suffix allomorphy or base allomorphy. On the other hand, it has been found that when we use suffixes, visual word recognition is not facilitated.
The present work is a psycholinguistic study with is focus on word reading. In this field, attention has been focused on graphic, phonological and semantic aspects. Regarding morphological aspects, the theoretical models of visual word recognition do not satisfactorily respond to all the questions that empirical observation of the data allows to formulate. A commonly reported result in studies that use morphological priming paradigm is the facilitation of word recognition when it is preceded by another word with which it is morphologically related. The problem of this hypothesis of reading facilitation, in these circumstances, comes from the type of morphological relation considered, which also manifests itself in orthographic overlap, in phonological proximity and in semantic nexus. We intend, therefore, to find evidence that morphological processing is not only central in the visual word recognition, but also occurs independently of orthographic, phonological and semantic processing. We also sought to document this evidence at two distinct times in the chronology of language use: in children, attending the fourth grade; and in young adults, college students. Another issue that we have decided to make central is related to the fact that many of visual recognition models, as well as a large majority of the studies associated with them, have been developed on data from English, a language with a high orthographic irregularity, with various inconsistencies and complexities, and a poor morphology. In this study, we propose the analysis of Portuguese data, whose morphological system is generally considered to be richer and where data may differ from the one already found in the literature. The knowledge of the morphological system of Portuguese in a global and systematic approach seems essential to us as the basis for the observation of morphological processing of words. For this reason, we adopted the morphological analysis of Villalva (1994, 2000, 2003, 2008) to establish morphological conditions that we wanted to analyse, namely derived words and a distinction between derived words that have a compositional structure and words that are lexicalized. For the present work were delineated three experiments. In the first one we used the priming paradigm associated to a lexical decision task. We tested two nominalization suffixes: one affixed to a nominal radical (–eiro) and the other affixed to a verbal theme (–dor). The experimental design allowed us to test three experimental conditions of prime/target word pairs: (i) morphologically related (e.g. mineiro 'miner' / carteiro 'mailman'); (ii) with apparently related words (e.g. mineiro 'miner' / poleiro 'perch'); and (iii) unrelated words (e.g. mineiro 'miner' / caneta ‘pen’). The prime was exposed for 50 and 150 ms, immediately followed by the target, which was available until the lexical decision was made. For each time of exposure, different samples of children and adults were used (the children attending the fourth grade and adults were undergraduate students). The results showed that both children and adults have longer reaction times on morphologically related targets compared to apparently related targets and unrelated targets. The data do not match with the one mentioned in the literature, since the subjects showed higher reaction times when there is a morphological relationship between the primes and the targets. Nevertheless, the results show that, for Portuguese language, there are indications of morphological segmentation, as well as identification of the suffixes independently of the morphological compositionality of the words. In the second experiment, we continued to use the morphological priming method associated with the lexical decision task, constituting three experimental conditions: (i) compound derived words (e.g. desejo 'desire' / desejoso 'desiring'); (ii) lexicalized derived words with suffix allomorphy (e.g. luxo 'luxury' / luxuoso 'luxurious'); and (iii) lexicalized derived words with base allomorphy (e.g. água 'water' / aquoso 'aqueous'). As the prime we always presented the simple word and as a target, the derived words. Three prime exposure times (50, 100 and 150 ms) were used, and once again, different samples of children and adults were tested for each exposure time. The results confirm that derived words processing is sensitive to morphological structure and also show that compositional structures involve lower processing costs than the structures that contain allomorphic variation. The results were consistent at all prime exposure times. In the third experiment we used a lexical decision task on five experimental conditions, which resulted from the words used as targets in previous experiments: (i) simple words (e.g. janela 'window'); (Ii) compositional derived words (e.g. desejoso 'wishful'); (iii) lexicalized derivative words with total allomorphy (e.g. poleiro 'perch'); (iv) lexicalized derived words an allomorphic suffix (e.g. luxuoso 'luxurious'); and (v) lexicalized derived words with radical allomorphy (e.g. aquoso 'aqueous'). The sample consisted in a group of children and a group of adults. The results showed, once again, that lexical access is sensitive to the words morphological properties, and the words with compositional structure have shorter response times than the simple structures or than the lexicalized structures with radical or suffix allomorphy. Compositionality, therefore, appears to be a facilitating factor in lexical access. The results obtained in the three experiments allow us to construct the hypothesis that the visual processing of complex derived words in Portuguese is sensitive to the morphological structure properties. On the one hand, it has been found that compositional structures have a lower processing cost than structures disturbed by factors such as suffix allomorphy or base allomorphy. On the other hand, it has been found that when we use suffixes, visual word recognition is not facilitated.
Descrição
Palavras-chave
Morfologia (Linguística) Reconhecimento das palavras Linguagem - Aquisição Psicolinguística Teses de doutoramento - 2017