Resumo A dissertação analisa as ParaGrafias cometidas sob ditado auditivo-visual de 560 palavras incomuns por parte de 154 estudantes (61 do Ensino Superior e 93 do Ensino Fundamental, do 5º ao 9º ano). Precisão de escrita foi analisada como função do grau médio de cifrabilidade, que é a média aritmética dos índices ponderados de cifrabilidade das relações entre Unidades de Fala (UF) e Unidades de Escrita (UE) contidas na palavra falada a ser escrita sob ditado auditivo-visual. O estudo ofereceu 521.640 oportunidades de cifragem de UF em UE, 285.943 (54,82%) de UF vocálicos, e 235.697 (45,18%) de UF consonantais. Destas 521.640 oportunidades, 421.126 (80,73%) foram apropriadas (231.546 vocálicas [80,98%] e 189.580 [80,43%] consonantais), e 100.514 [19,27%] consistiram em ParaGrafias (54.397 [19,02%] vocálicas e 46.117 [19,57%] consonantais). Destas 100.514 ParaGrafias, houve 96.649 [96%] substituições (51.528 [53%] vocálicas e 45.121 [47%] consonantais), e 3.865 [4%] omissões (2.577 vocálicas [67%] e 1.288 [33%] consonantais). Das 96.649 substituições, 79.518 [82%] foram canônicas (cf. modelo de Capovilla & Casado, 2014), e 17.131 [18%] foram não canônicas. Das 51.528 substituições vocálicas, 40.843 [79%] foram canônicas e 10.685 [21%] foram não canônicas. Das 45.121 substituições consonantais, 38.675 [86%] foram canônicas e 6.446 [14%] foram não canônicas. Este estudo descreve as ParaGrafias vocálicas. Ele cobriu 58 relações FonoGrafÊmicas vocálicas além de 14 conjuntos de OtoLalEmas vocálicos HomoScópicos (e.g., {\e\, \ej\, \ej\};{\a\, \a\, \\}; {\\, \j\}; {\i\, \\, \ij\}; {\o\, \oj\, \oj\}; {\wa\, \w\}; {\aj\, \j\}; {\iw\, \w\}; {\wi\, \w\}; {\ja\, \j\}; {\wej\, \we\}; {\w\, \w\}; {\wo\, \w\}; {\u\, \\, \uw\, \w\}, cf. Capovilla, 2015, 2018). Corroborando o modelo de Capovilla e Casado (2014), a precisão da cifragem foi função positiva da prevalência com que cada GrafEma cifra cada FonEma no Português. Ou seja, a cifragem correta foi função do IPC da relação UFA-UE. A grande incidência de substituições cometidas sob controle visual de OtoLalEmas HomoScópicos sugere a importância do processamento visual em reduzir a variabilidade das ParaGrafias cometidas sob ditado auditivo-visual ao vivo. Tais dados com ParaGrafias vocálicas corroboram os de Graton-Santos (2017) com ParaGrafias consonantais. Os dois estudos revelaram que as ParaGrafias (vocálicas e consonantais) são restritas ao local de articulação, de modo que FonEmas visualmente semelhantes (OptoLalEmas HomoScópicos) envolvendo o mesmo local de articulação foram confundidos entre si

This dissertation analyzes spelling errors committed by students while spelling under both auditory and visual live dictation. A sample of 154 volunteer students took part of the experiment. There were 61 college students, and 93 elementary-school students from 5th to 9th grade levels. The task consisted of spelling under dictation 560 spoken words of very low orthographic familiarity, so as to maximize phonological encoding processes. The list of 560 words to be spelled under dictation had very low orthographic familiarity, as ascertained by a mean annual search frequency in the Google AdWord database of under 16 searches. The task of spelling under dictation consisted in encoding Phonemes into Graphemes. Encoding precision was analyzed as function of the word mean encodability degree, which was the average of the encodability degree of all the Phoneme- Grapheme relationships contained in the spoken word to be spelled. This dissertation analyzes spelling errors committed under auditory-visual dictation of 560 uncommon spoken words by 154 students (61 college students, and 93 5th-9th grade elementary-school students). Encoding precision was analyzed as function of the word mean encodability degree, i.e., the average of the encodability degrees of all the Phoneme-Grapheme relationships contained in the spoken word to be spelled under auditory-visual dictation. There were 521,640 discrete Phoneme-Grapheme encoding opportunities, 285,943 (54.82%) vowel phonemes, and 235,697 (45.18%) consonant phonemes. Out of the 521,640 cyphering opportunities, 421,126 (80.73%) were appropriate spellings (231,546 vowel phonemes [80.98%] and 189,580 [80.43%] consonant phonemes), and 100,514 [19.27%] were misspellings (paragraphias) (54,397 [19.02%] vowel phonemes and 46,117 [19.57%] consonant phonemes). Out of the 100,514 vowel and consonant misspellings, there were 96,649 [96%] substitutions (51,528 [53%] vowel phonemes and 45,121 [47%] consonant phonemes), and 3,865 [4%] were omissions (2,577 pertaining to vowels [67%] and 1,288 [33%] pertaining to consonants). Out of the 96,649 substitutions, 79,518 [82%] were cannonical ones (cf. Capovilla & Casados 2014 model), and 17,131 [18%] were uncannonical ones, which reflect problems with literacy acquisition. Out of the 51.528 vowel substitutions, 40,843 [79%] were cannonical ones and 10,685 [21%] were cannonical ones. Out of the 45,121 consonant substitutions, 38,675 [86%] were cannonical ones and 6.446 [14%] were uncannonical ones. The present study describes vowel misspellings. It assessed the cyphering of 58 vowel phonemes presented auditorily. Out of the 58 phonemes, 38 phonemes were grouped in 14 sets of visually identical phonemes: {\e\, \ej\, \ej\};{\a\, \a\, \\}; {\\, \j\}; {\i\, \\, \ij\}; {\o\, \oj\, \oj\}; {\wa\, \w\}; {\aj\, \j\}; {\iw\, \w\}; {\wi\, \w\}; {\ja\, \j\}; {\wej\, \we\}; {\w\, \w\}; {\wo\, \w\}; {\u\, \\, \uw\, \w\}. Results showed that the phoneme cyphering precision was a positive function of the phoneme cypherability index in Portuguese (cf. Capovilla & Casado, 2014). The prevalence of substitutions that occurred under visual presentation suggests the importance of visual information in reducing the variability of misspellings committed in the task of spelling under auditory-visual dictation. The present results with vowel misspellings corroborate GratonSantos (2017) findings with consonant misspellings. Taken together, both studies revealed that vowel and consonant misspellings tend to become restricted to the articulation point, so that visually similar phonemes involving a common articulation point tend to be mistaken one for the other