Portuguese Corpus Research Articles

A Diachronic Overview of the Prepositional Accusative in Portuguese

One of the several differences between Modern European Portuguese (EP) and Modern Brazilian Portuguese (BP) is the prepositional expression of complements licensed by the preposition a. While in EP the preposition a occurs in several contexts, this element has been substituted by other strategies in BP, as is extensively discussed in the literature. The aim of this paper is to investigate the historical behavior of a-marked prepositional accusatives (PP-ACC) in Portuguese. In order to do so, a search was conducted for PP-ACCs in the Historical Portuguese Corpus Tycho Brahe. The results showed an increase of PP-ACCs in the 17th century, followed by a decrease in the 18th century. Thereafter, unmarked accusatives (NP-ACC) were analyzed in the corpus, which resulted in 7756 sentences, contrasting with 624 PP-ACCs in the same contexts. This result shows that the a-marked accusative is far less common than bare accusatives in Historical Portuguese. Psych verbs, however, behaved differently, showing a constant increase in PP-ACCs. In EP, the preposition a still introduces Experiencer arguments in structures with some psych verbs (O vinho agradou ao João—lit. ‘The wine pleased ‘to’ John’). In BP, the preposition a has disappeared in psych predicates (O vinho agradou Ø o João—‘The wine pleased John’). In both Modern EP and BP, most PP-ACCs have become typical unmarked direct objects. In the context of psych verbs, however, structural accusative assignment has shifted to structural dative Case in Modern EP, so as to ascertain the interpretation of the Experiencer in the internal argument via the preposition a. While in Modern BP, the argument is not overtly marked since it receives inherent accusative case in the derivation.

Multipluralism with education in Google books: Mechanisms descriptive about transsexuality and social representations and rights of adolescents

The work presented was prepared according to the need to understand the representations of human beings in their recognized historical rights identified throughout mainly adolescence as an important stage of life. This research is also motivated to understand public health policies identified in alignment with the representations of the human being in the phases of life, especially those aimed at transsexuals. I used the following lexical items added in English using the search engine through Google Books and, concomitantly, scanning in Brazilian Portuguese in a correspondence search engine called Brazilian Corpus made available by the Linguateca1 initiative. Therefore, these are recognized masculine genders, women, gays, lesbians, homosexuals, transgender people; person, transgender, transgender person, transsexual person, spiritual person, transgender person, transsexual person, gender queer, gender dysphoria, gender identity, gender diverse, gender non-conforming minors, gender non-conforming, gender affirming, non-binary; non-binary; transgender, gender diversity, non-conforming adolescents, LGBTQ, non- conforming people, minority genders, gender transition, family, adopted family (as well as their own plural forms). The phases of life have been operationalized by the following measures: man, woman, adolescent, adolescence, adult, boy, girl, child, elderly, child, young person and adolescent (as well as their plural forms). The data for the analysis consists of publications made available by Google Books in the period 1800 to 2008, totaling approximately 500 million words, in addition to the use of the SKETCH ENGINE Corpus which brings together publications in English and Brazilian Portuguese in this analysis. The Portuguese Corpus contains around 1 billion words of data from newspapers and magazines on the Internet in four Portuguese-speaking countries between 2012 and 2019 and is called the Portuguese Web (ptTenTen20). To this end, several analyzes were carried out, focusing in particular on the results related to the rights of transsexual people and when these rights began to be recognized in publications and the possible health policies addressed to adolescents and education.

A study on methods for revising dependency treebanks: in search of gold

Reliably annotated corpora with reliable annotation are a valuable resource for Natural Language Processing, which justifies the search for methods capable of assisting linguistic revision. In this context, we present a study on methods for revising dependency treebanks, investigating the contribution of three different strategies to the corpus review: (i) linguistic rules; (ii) an adaptation of the n-grams method proposed by Boyd et al. (2008) applied to Portuguese; and (iii) Inter-Annotator Disagreement, a linguistically motivated approach that draws inspiration from the human annotation process. The results are promising, and taken together the three methods can lead to the revision of up to 58% of the errors in a specific corpus at the cost of revising only 20% of the corpus. We also present a tool that integrates treebank editing, evaluation and search capabilities with the review methods, as well as a gold-standard Portuguese corpus from the oil and gas domain.

A comparison of syntactic patterns of Brazilian Portuguese and English: using comparable corpora of museum texts

The aim of this paper is to analyze two corpora (one in Brazilian Portuguese and another English), from PUCRS’ Science and Technology Museum, to unveil which of these languages could be rendered to be more or less syntactically and rhetorically elaborate. The corpora were processed on Natural Language Processing tools, such as NILC-Metrix, Coh-Metrix 3.0 and AntConc. Results point to the feature regarding the higher number of words before the main verb as the main factor rendering syntactic and rhetoric elaboration and, consequently, more complexity in the Portuguese corpus.

A Use Case of Patent Classification Using Deep Learning with Transfer Learning

Abstract Purpose Patent classification is one of the areas in Intellectual Property Analytics (IPA), and a growing use case since the number of patent applications has been increasing worldwide. We propose using machine learning algorithms to classify Portuguese patents and evaluate the performance of transfer learning methodologies to solve this task. Design/methodology/approach We applied three different approaches in this paper. First, we used a dataset available by INPI to explore traditional machine learning algorithms and ensemble methods. After preprocessing data by applying TF-IDF, FastText and Doc2Vec, the models were evaluated by cross-validation in 5 folds. In a second approach, we used two different Neural Networks architectures, a Convolutional Neural Network (CNN) and a bi-directional Long Short-Term Memory (BiLSTM). Finally, we used pre-trained BERT, DistilBERT, and ULMFiT models in the third approach. Findings BERTTimbau, a BERT architecture model pre-trained on a large Portuguese corpus, presented the best results for the task, even though with a performance of only 4% superior to a LinearSVC model using TF-IDF feature engineering. Research limitations The dataset was highly imbalanced, as usual in patent applications, so the classes with the lowest samples were expected to present the worst performance. That result happened in some cases, especially in classes with less than 60 training samples. Practical implications Patent classification is challenging because of the hierarchical classification system, the context overlap, and the underrepresentation of the classes. However, the final model presented an acceptable performance given the size of the dataset and the task complexity. This model can support the decision and improve the time by proposing a category in the second level of ICP, which is one of the critical phases of the grant patent process. Originality/value To our knowledge, the proposed models were never implemented for Portuguese patent classification.

SemClinBr - a multi-institutional and multi-specialty semantically annotated corpus for Portuguese clinical NLP tasks

BackgroundThe high volume of research focusing on extracting patient information from electronic health records (EHRs) has led to an increase in the demand for annotated corpora, which are a precious resource for both the development and evaluation of natural language processing (NLP) algorithms. The absence of a multipurpose clinical corpus outside the scope of the English language, especially in Brazilian Portuguese, is glaring and severely impacts scientific progress in the biomedical NLP field.MethodsIn this study, a semantically annotated corpus was developed using clinical text from multiple medical specialties, document types, and institutions. In addition, we present, (1) a survey listing common aspects, differences, and lessons learned from previous research, (2) a fine-grained annotation schema that can be replicated to guide other annotation initiatives, (3) a web-based annotation tool focusing on an annotation suggestion feature, and (4) both intrinsic and extrinsic evaluation of the annotations.ResultsThis study resulted in SemClinBr, a corpus that has 1000 clinical notes, labeled with 65,117 entities and 11,263 relations. In addition, both negation cues and medical abbreviation dictionaries were generated from the annotations. The average annotator agreement score varied from 0.71 (applying strict match) to 0.92 (considering a relaxed match) while accepting partial overlaps and hierarchically related semantic types. The extrinsic evaluation, when applying the corpus to two downstream NLP tasks, demonstrated the reliability and usefulness of annotations, with the systems achieving results that were consistent with the agreement scores.ConclusionThe SemClinBr corpus and other resources produced in this work can support clinical NLP studies, providing a common development and evaluation resource for the research community, boosting the utilization of EHRs in both clinical practice and biomedical research. To the best of our knowledge, SemClinBr is the first available Portuguese clinical corpus.

Autorreferência generalizadora com a pessoa, uma pessoa e outros nomes humanos referencialmente vagos

This article analyzes data extracted from the Syntax-oriented Corpus of Portuguese Dialects (CORDIAL-SIN) and Corpus Africa with the aim of demonstrating that the arbitrary/indefinite expressions uma pessoa and a pessoa convey first-person-based genericity, that is, what Moltmann (2010) calls ‘generalizing detached self-reference’. Differences in contextual adequacy between uma pessoa and a pessoa are then discussed using intuitive data. It is suggested that such distributional differences are a consequence of the greater interpretative flexibility of a pessoa (which allows an inclusive or exclusive reading) relative to uma pessoa(which only allows an inclusive reading). Finally, the geolinguistic distribution of uma pessoaand a pessoa in the Portuguese territory is compared to the geolinguistic patterns found for other expressions with human general nouns that also seem to convey generalizing detached self-reference (i.e. homem, fulano, fulana, gajo, tipo, indivíduo, mulher).

Listas de sentenças em português: avaliação da percepção da fala em crianças

ABSTRACT Purpose: to develop lists of sentences in Brazilian Portuguese based on daily vocabulary to evaluate auditory speech perception in children. Methods: a Brazilian Portuguese corpus with adaptations to attend the criteria established for this study and analyzed by judges was used to obtain the sentences. Thirty-seven lists, each composed of thirteen sentences with fifty phonological keywords, were constructed. The material was presented to ten children, aged between 6 and 10 years, with no hearing or communication complaints. The Friedman test was applied to verify possible differences among the lists. The level of significance adopted was 5% (0.05). Results: the mean speech recognition index presented by the participants was 96.8% for all lists. The thirty-seven lists were considered statistically similar (p=0.140). Conclusion: the developed material proved to contain sentences that represent situations of children’s day-to-day communication and its application is easy and quick.

Uma análise para a distinção entre evidencialidade e modalidade epistémica o caso dos adjetivos alegado e suposto e respetivas formas adverbiais

The modal and evidential domains are not always easily distinguishable. Even though some languages do have specific modal and evidential markers, European Portuguese does not behave the same way, having, therefore, several ways and mechanisms to mark both evidentiality and modality. The present study, based on examples retrieved from the European Portuguese corpus CETEMpúblico, thus, intends to evaluate if the traditionally considered modal adjectives, suposto (supposed) and alegado (alledged), and their corresponding adverbial forms, can convey evidential values and contribute to signal a distinction between the two domains. The examples show that these adjectives and adverbs behave distinctively from each other: suposto (and supostamente) conveys suppositional evidentiality, while alegado (and alegadamente) conveys reportative evidentiality. Even though the preferred interpretation is an evidential one, they both allow for an epistemic modal reading. This means, put differently, that, although these adjectives and adverbs have the capacity to represent both domains, modality seems to always come second: the transmission of evidential values is therefore stronger than the expression of epistemic modal values (of uncertainty). However, the epistemic reading of alegado, in particular, seems to be related to the fact that the author does not mean to compromise him/herself with the veracity of the proposition, rather than convey a certain degree of uncertainty.

A “língua desportuguesa”. Próclise no português angolano e no português moçambicano

The expansion of proclisis in contexts that typically exclude it in European Portuguese has been described as a syntactic feature that characterizes Angolan Portuguese. This article studies the presence of this feature in the literary language, based on a corpus of texts by Angolan and Mozambican authors, representing two generations: authors born in the colonial period (the Angolan Pepetela and the Mozambicans Mia Couto and Paulina Chiziane) and authors born after the independence of their countries (the Angolan Ondjaki and the Mozambican Lucílio Manjate). The study results show that proclisis has a stronger presence in the works of Angolan authors than Mozambican authors, suggesting that the shift towards the generalization of proclisis is more advanced, socially widespread and accepted in Angolan Portuguese than in Mozambican Portuguese, although it is visible in both. It is in non-finite domains that the contrast between the two African varieties is more evident. Not only is the frequency of proclisis to the infinitive higher in the Angolan Portuguese corpus, but only there is proclisis to the past participle attested. Comparing the two generations of writers, we see a significant rise of proclisis to the infinitive between Pepetela and Ondjaki and it is in Ondjaki’s works that proclisis to the past participle occurs. Regarding the Mozambican Portuguese corpus, however, there seems to be a regression in the advance of proclisis between Mia Couto and Manjate. A closer look at Rabhia, by Manjate (2017), suggests that young speakers with a high level of education may perceive the spread of proclisis as a socially marked feature.

Grau de auxiliaridade da microconstrução [deixar+de+V2inf.] no português arcaico

Propomos investigar, com base na abordagem construcional (TRAUGOTT; TROUSDALE, 2013), a microconstrução aspectual [deixar+de+V2inf.] no português arcaico (séculos XIII-XV). Para análise, utilizamos o Corpus do Português (DAVIS; FERREIRA, 2006, 2016) e cotejamos parâmetros fundamentados nos estudos clássicos de verbo auxiliar. No período estudado, a microconstrução, sob a forma [leixar+de+V2inf.], apresenta índices de auxiliaridade e marca aspecto final (terminativo e cessativo). Além disso, ela é parcialmente esquemática, pela recorrência de elementos entre seus componentes, parcialmente produtiva, por ainda estar expandindo seus arranjos colocacionais, e não composicional, pela codificação de seus elementos como um chunk procedural.

Intensification in Portuguese: A Cross-dialectal Study of Muito and Bem

Intensifiers have been the focus of a number of studies over the past decade, with considerable interest in their meaning and variability. Several scholars have discussed the use of such forms, particularly in English (Ito and Tagliamonte 2003, Tagliamonte 2008) and Spanish (Brown and Cortes-Torres 2013; Kanwit et al. 2017), exploring their structure and sociolinguistics. The objectives of the current analysis are 1) to explain the use and behavior of the canonical intensifier muito and its more colloquial counterpart bem; and 2) to understand this dichotomy in four varieties of Portuguese, namely Angolan, Brazilian, European, and Mozambican. Previous studies have shown that intensifier variation in English (e.g., very versus really) and Spanish (e.g., muy versus bien) is sensitive to a variety of language-internal and -external factors. Using a Digital Humanities-informed approach with corpus data from Corpus de referencia do portugues contemporâneo (CRPC) and Spoken Portuguese Corpus (SPC), we show that the distribution of Portuguese muito and bem, while semantically parallel to that Spanish muy/bien, displays a wide range of dialectal nuances. This work is the first to provide a comprehensive account of the behavior of intensifiers in Portuguese, both within and across dialects.

Construção e Avaliação para 15 Listas de Palavras Baseadas no Paradigma Deese-Roediger-McDermott

Resumo Com base nos estudos conduzidos de acordo com o paradigma Deese-Roediger-McDermott (DRM), em quais palavras de uma lista tendem a evocar outros itens relacionados, utilizado em pesquisas de falsas memórias, neste estudo elaborou-se 15 listas de palavras semanticamente relacionadas para o português. Para a elaboração das listas, foram considerados critérios como frequência de palavras no corpus de português brasileiro e extensão de palavras. A pré-seleção guiada por características linguísticas visava controlar possíveis vieses por parte dos voluntários. Testes de associação semântica e concretude de palavras foram realizados para apurar a conveniência dos itens selecionados, e um teste de recordação livre foi conduzido para avaliar características de evocação. Sugerimos a utilização das listas de palavras em pesquisas que avaliem falsas memórias.

Distributional Models with Syntactic Contexts for the Measurement of Word Similarity in Brazilian Portuguese

The similarity between words constitutes significant support to tasks in natural language processing. Several works use Lexical resources such as WordNet for semantic similarity and synonym identification. Nevertheless, words out-of-vocabulary or missing links between senses are perceived problems of this approach. Distributional-based proposals like word embeddings have successfully been used to meet such problems, but the lack of contextual information can prevent the achievement of even better results. The distributional models that include contextual information can bring advantages to this area, but these models are still scarcely explored. Therefore, this work studies the advantages of incorporating syntactic information in the distributional models, fostering for better results in semantic similarity approaches. For that purpose, the current work explore existing lexical and distributional techniques regarding the measurement of word similarity in Brazilian Portuguese. Experiments were carried out with the lexical database WordNet, using different techniques over a standard dataset. The results indicate that word embeddings can cover words out of vocabulary and have better results in comparison with lexical approaches. The main contribution of this article is a new approach to apply syntactic context in the training process of word embeddings to a Brazilian Portuguese corpus. The comparison of this model with the outcome of the previous experiments shows sound results and presents relevant complementary aspects.

Obituaries in translation: a corpus-based study

Mais cedo ou mais tarde, a morte afetará todas as pessoas, em todos os lugares. No entanto, essa dura realidade é enfrentada de formas diferentes em culturas distintas. Os obituários podem ajudar a revelar algumas dessas diferenças e seu impacto para a tradução (LOOCK & AMP; LEFEBVRE-SCODELLER, 2014). Usando a Linguística de Corpus como metodologia (BAKER, 1993; LAVIOSA, 2002; McENERY & AMP; HARDIE, 2012; PHILIP, 2009; ZANETTIN, 2012), investigamos se e até que ponto um corpus comparável de obituários em inglês estadunidense e em português brasileiro pode ajudar na tarefa de conscientizar os aprendizes sobre as peculiaridades culturais observadas em um mesmo gênero textual escrito em diferentes idiomas, e suas consequências para a identificação de equivalentes. Para atingir nossos objetivos, selecionamos textos publicados em jornais brasileiros e estadunidenses entre 2015 e 2017. Apesar de abordarem um assunto cotidiano, os obituários são pouco explorados academicamente. No entanto, essa negligência não é proporcional nos dois países, e resulta da popularidade do gênero. Enquanto nos Estados Unidos os obituários são lidos regularmente, no Brasil eles são raros, quase exclusivamente dedicados a falecidos famosos. Análises qualitativas e quantitativas mostraram que a ausência de termos em português se deve a diferenças culturais que permeiam o tema, e que a falta de familiaridade com o gênero, além das peculiaridades dos rituais praticados nos dois países / culturas, pode ajudar a explicar as dificuldades enfrentadas pelos aprendizes brasileiros na tarefa de traduzir obituários estadunidenses para o português brasileiro.

Recognizing Textual Entailment: Challenges in the Portuguese Language

Recognizing textual entailment comprises the task of determining semantic entailment relations between text fragments. A text fragment entails another text fragment if, from the meaning of the former, one can infer the meaning of the latter. If such relation is bidirectional, then we are in the presence of a paraphrase. Automatically recognizing textual entailment relations captures major semantic inference needs in several natural language processing (NLP) applications. As in many NLP tasks, textual entailment corpora for English abound, while the same is not true for more resource-scarce languages such as Portuguese. Exploiting what seems to be the only Portuguese corpus for textual entailment and paraphrases (the ASSIN corpus), in this paper, we address the task of automatically recognizing textual entailment (RTE) and paraphrases from text written in the Portuguese language, by employing supervised machine learning techniques. We employ lexical, syntactic and semantic features, and analyze the impact of using semantic-based approaches in the performance of the system. We then try to take advantage of the bi-dialect nature of ASSIN to compensate its limited size. With the same aim, we explore modeling the task of recognizing textual entailment and paraphrases as a binary classification problem by considering the bidirectional nature of paraphrases as entailment relationships. Addressing the task as a multi-class classification problem, we achieve results in line with the winner of the ASSIN Challenge. In addition, we conclude that semantic-based approaches are promising in this task, and that combining data from European and Brazilian Portuguese is less straightforward than it may initially seem. The binary classification modeling of the problem does not seem to bring advantages to the original multi-class model, despite the outstanding results obtained by the binary classifier for recognizing textual entailments.

A quantitative survey of N Prep N constructions in Romance languages and prepositional variability

The distinction between syntagmatic compounds of the type N Prep N, such as Fr. jouet d’enfant, and nominal syntagms of the type N Prep N, such as the partially equivalent Fr. jouet pour enfants, remains unclear and vague. This is mainly because the lexical and syntactic status of syntagmatic compounds still is controversial. In some cases, as in jouet d’enfant and jouet pour enfants, partial equivalent syntagmatic compounds and nominal syntagms may coexist and underlie a specific variation and alternation. In other cases, such as Pt. bracelete de aço and bracelete em aço, two variants of a syntagmatic compound may alternate and coexist. The first part of this paper provides an overview of the current discussion on these two types of constructions. The second part addresses the alternation and variation of syntagmatic compounds and nominal syntagms by means of analysis of large-scale corpus data, the French, Spanish and Portuguese corpus of the TenTen family. Here, the focus lies on the variation of the prepositional internal element of these constructions as well as on a comparison of different word formation patterns.

Mood choice in complements of Spanishcomprenderand Portuguesecompreender(‘understand’) – distribution and meaning

AbstractThe present paper analyzes the occurrence of indicative and subjunctive complements of the verbscomprender(Spanish) andcompreender(Portuguese) in European Spanish and European Portuguese. A quantitative analysis based on 400 occurrences of the complements randomly selected from the newspaper genre shows that the indicative mood occurs more frequently than the subjunctive mood in both languages, although the subjunctive mood is more frequent in the Portuguese corpus than in the Spanish one. The analysis also shows that the occurrence of the subjunctive complement is highly restricted to contexts in which the subject of the main clause verb is either 1st person or 3rd person singular. From the theoretical perspective of Cognitive Grammar, the mood alternation is explained by the concept ofdominion, i.e. the indicative complement designates an event that is located within the conceptualizer’s epistemic dominion, whereas the subjunctive complement designates an event that is located outside the conceptualizer’s dominion of effective control.

BrAgriNews: Um Corpus Temporal-Causal (Português-Brasileiro) para a Agricultura

There has been a recent sharp increase in interest in academia and industry in applying machine learning and artificial intelligence to agricultural problems. Text mining and related natural language processing techniques, have been rarely used to tackle agricultural problems, and at the time of writing there was a single project in the Portuguese language. It is possible that the failure of researchers to use text mining techniques to analyze Portuguese texts to resolve agricultural problems may be due to a lack of freely available corpora. To correct the lack of a Portuguese language agriculture centric corpus we are releasing a Brazilian-Portuguese agricultural language resource, which is described by this paper. The corpus is partially non-contiguous and spans a time period from 1996 to 2016. It consists of news stories that have been scraped from Brazilian News sites that have been annotated with the following information types: causal, sentiment, named entities that include temporal expressions. The corpus has additional resources such as a: treebank, lists of frequent: unigrams, bigrams and trigrams, as well words or phrases that have been identified by journalists as either: ``important'' or domain specific. It is hoped that the release of this corpus will stimulate the adoption of text mining in agriculture in the Lusophonic research community.

Consequências da composição corporal na produção do Português Europeu: estudo de resultados preliminares

We found that production parameters of a European Portuguese corpus are influenced by body composition: more prominently in the frequency parameters (f0, DP f0 and formants), perturbation parameters and noise parameters, for vowel segments [a] and [i]. We concluded that extralinguistic patterns of voice quality were significantly correlated with body composition and somatotype. The vocal quality characterization of a speaker can allow the identification of his particular phonetic settings and the use of his anatomic mechanism, during speech.