Imprimir Republish


A world of possibilities

Establishing literary authorship poses a complex challenge to researchers

Léo Ramos Chaves

More than 100 years after his death, the name Joaquim Maria Machado de Assis (1839–1908) remains very much alive on the Brazilian literary scene, judging from the excitement created by the recent article “M´achado biógrafo: Da investigação de uma revista a um texto inédito” (M´achado as biographer: An investigation of a magazine reveals new findings), published by Cristiane Garcia Teixeira in ArtCultura, an academic journal of the Federal University of Uberlândia. In the article, the historian suggests that a biographical sketch about Dom Pedro II (1825–1891), published anonymously on November 6, 1859 in the magazine O espelho: Revista Semanal de Literatura, Modas, Indústria e Artes, may have been written by Machado de Assis at the age of 20.

Teixeira found the text when doing research for the master’s dissertation she defended at the Federal University of Santa Catarina (UFSC) in 2016. “The first thing that caught my eye was the biography’s first-page placement in the magazine,” says the researcher, now a doctoral student in the Graduate Program in History at the same institution. “The five-column text not only was on the first page, but used the same format as used for other novels and articles by Machado, the periodical’s main contributor.”

There were also other elements suggesting to Teixeira that the text may have been written by the future author of classics such as Dom Casmurro and The posthumous memoirs of Brás Cubas. In issue 6 of the same magazine, dated October 9, an anonymous notice announced: “We will soon begin publishing a gallery with biographies and corresponding portraits. The photographer is Mr. Gaspar Guimarães, and the biographer is Mr. Machado de Assis.” Garcia also identified distinguishing attributes of Machado’s work, such as the use of the first person as was common in other contemporaneous texts by the now-celebrated author. “Contradiction, which was very characteristic of Machado, is another element that appears in the text. In the opening sentence, the author warns that he will not talk about politics, but then proceeds to do the very opposite,” she notes. “But we can’t know for sure, as the text is unsigned.”

Hélio de Seixas Guimarães, a professor of Brazilian literature at the School of Philosophy, Languages and Literature, and Humanities at the University of São Paulo (FFLCH-USP), sees fewer similarities. “It is not implausible that Machado wrote about the emperor—he admired him and in the 1860s dedicated poems to both Dom Pedro II and the wider imperial family,” says the expert. “In my opinion, the most compelling evidence of Machado’s authorship is found at the beginning of the biographical sketch, in the separation that the anonymous author makes between the role of a cronista from that of a historian, and between the subjects that are in the domain of each. This is a topic that Machado did in fact explore elsewhere in his writing career.” However, Guimarães finds it difficult to identify any resemblance of Machado’s style in the biographical sketch. “The text ends abruptly with no closure, whereas Machado was known to end his texts with some food for thought for the reader. I also don’t see the use of contradiction in the biographical text,” observes Guimarães. “In my view, the text is report-like, laudatory, and perhaps written to win the emperor’s favor and financial patronage for the magazine, as was common at the time. But I also wouldn’t rule out the possibility that the text was indeed written by Machado.”

The authorship of anonymous literary texts has long intrigued researchers. “There is a lot of uncertainty surrounding some of Machado de Assis’s writings, and even today we don’t know exactly how many short stories he wrote, in addition to the 200 for which his authorship is well established,” says Guimarães. “The dubious ones are largely from his early career, published anonymously or with initials or pseudonyms now attributed to Machado, some without positive evidence. Even Brazilian poet and writer Carlos Drummond de Andrade [1902–1987] ventured to identify the pseudonyms Camillo da Anunciação, used in the short story ‘A vida eterna’ (Eternal life), and Marco Aurélio, from the short story ‘Possível e impossível’ (Possible and impossible), as belonging to Machado de Assis.”

In his master’s research under Guimarães, Fernando Borsato dos Santos is investigating the wide range of pseudonyms and initials that Machado de Assis used in his writings, and how they varied throughout his career and across different genres. “He is analyzing a total of 83 signatures, of which 53 are pseudonyms and the rest are different initials that Machado is thought to have used in his texts, such as M. A., M-as, and J. M.,” says Guimarães. Uncertainty over authorship is a source of controversy among Machadian scholars. A case in point, says Guimarães, is the short story “Felicidade pelo casamento” (Happiness through marriage), published in Jornal das Famílias, in 1866. Authorship was first attributed to Machado de Assis by biographer Raimundo Magalhães Júnior (1907–1981), then contested by another biographer of Machado, Jean-Michel Massa (1930–2012), only to be reasserted more recently, in 2015, by researcher Mauro Rosso.

Léo Ramos Chaves

A profusion of pseudonyms, initials, and anonymous texts was customary in nineteenth-century periodicals, explains Tania Regina de Luca, a professor in the History Department at São Paulo State University (UNESP). “These publications typically only had a handful of contributors, and many used pseudonyms or variant initials to suggest to readers they had a larger writing staff. In addition, many contributors—who could either be established authors or aspiring literary writers—wrote for multiple publications and saw journalistic work as a means to supplement their income at a time when writing as a profession was in its infancy,” says de Luca, who authored the book A ilustração (1884–1892) – Circulação de textos e imagens entre Paris, Lisboa e Rio de Janeiro (The Illustration (1884–1892) – The interchange of texts and images between Paris, Lisbon, and Rio de Janeiro; Editora Unesp, 2018). “Legal protection of copyright was also non-existent, and was only established internationally at the end of the nineteenth century with the Berne Convention, in 1886. So many texts were then published unsigned.”

Joaquim Marçal Ferreira de Andrade, head of Brazil’s National Digital Library, concurs. “If even as late as the 1980s major Brazilian newspapers didn’t feel the need to credit photographers for images, imagine what things were like in the nineteenth century. The upshot is that today, researchers are often unsure as to the authorship of texts and images in old publications,” says Andrade, author of the book História da fotorreportagem no Brasil: A fotografia na imprensa do Rio de Janeiro de 1839 a 1900 (The history of photojournalism in Brazil: Photography in Rio’s press from 1839 to 1900; Editora Campus/Elsevier, 2004). “Establishing authorship is an enterprise that requires extensive knowledge, research, insight, and perseverance. It’s detective work.”

It was while doing research on nineteenth-century periodicals at the National Library that independent researcher Felipe Pereira Rissato discovered the first pseudonym used by writer and journalist Euclides da Cunha (1866-1909). “Cunha’s debut in the press was thought to have been in 1884, in the newspaper O democrata, but he had previously written for the newspapers Evolucionista and Espectador in 1883, at the age of 17, under the pseudonym Ícaro,” says Rissato, who shares credit for the discovery with his colleague, Leopoldo Bernucci, a professor at the University of California at Davis. Another of Rissato’s discoveries is an anonymous text published in the second edition of Revista Luso-Brasileira, in 1860, under the title Lembranças de minha mãe(Memories of my mother), which he credits to Machado de Assis. “I have recently discovered texts signed with the initials M. A., but I’m still investigating whether they belong to Machado de Assis or perhaps to Moreira de Azevedo, a co-contributor to many of the same periodicals.”

Sílvia Maria Azevedo, a professor at the School of Sciences and Languages and Literature at UNESP, has no doubt as to Machado de Assis’s authorship of the roughly 300 crônicas written under the pseudonym Dr. Semana in Semana Ilustrada, a magazine featuring political cartoons, caricatures, and political satire.  “He took over the column in 1869, when it was renamed Badaladas, remaining there until the magazine closed its doors in 1876,” says Azevedo, who last year launched the book Badaladas – Dr. Semana (Nankin Editorial), a two-volume compilation of footnotes, onomastic indexes, and tables with textual evidence of Machado’s authorship.

This has long been the subject of controversy among scholars. “José Galante de Sousa, in a 1950s book about Machado’s bibliography, wrote that the pseudonym had been used by several of the magazine’s contributors, and authorship of the crônicas could only be established after a careful examination of these materials,” says Azevedo. At that’s precisely what she undertook to do. “I think that we, as scholars, need to question sources and embrace challenges. This enriches the debate. But I only ventured to assert these claims because I’ve been studying Machado since the 1980s.”

Her book is the result of research she started in 2012. Azevedo analyzed each of the 300 crônicas over a period of three years. “I looked for internal evidence—aspects of the text itself and writing style—as well as external marks, such as references in other works by Machado in different genres, including theater criticism and fiction,” she explains. Finally, Azevedo’s study was critiqued by Machadian scholar Valentim Facioli, a retired professor of Brazilian literature at USP, and owner of the book’s publisher, Nankin Editorial. “Assigning authorship is a laborious and painstaking task,” says Azevedo.

For historian Denise de Almeida Silva, manager of the Archive Service at the Brazilian Studies Institute (IEB) at USP, authorship attribution is a multidisciplinary effort. “To determine the authorship of a document, I would need to consult specialists in fields such as philology, paleography, and diplomacy, and possibly scholars of a particular subject matter or person,” she says. “I would also consider any relationship between the relevant document and other documents in the archive.”

IEB houses the collections of Brazilians writers, artists, and intellectuals such as João Guimarães Rosa (1908–1967), author of The Devil to Pay in the Backlands. During research on Guimarães for a doctoral thesis she defended in 2013 at FFLCH-USP, Mônica Gama came across “small enigmatic manuscripts, resembling poems in prose” that turned out to be the dust jacket flaps of the third edition of the novella series Corpo de baile (Backland nights; Livraria José Olympio Editora, 1956). “Immediately after Guimarães’s death in the late 1960s, his publishers, José Olympio Editora, released a book in tribute to the author [Em memória de Guimarães Rosa (In memory of Guimarães Rosa; 1968)], in which they revealed that Guimarães himself had written the flap texts for his books. But there was no evidence that this was indeed the case,” recalls Gama, now a professor of letters at the Federal University of Ouro Preto (UFOP).

Léo Ramos Chaves

Following this clue, Gama searched the archives at IEB as well as collections at Fundação Casa de Rui Barbosa and the National Library in Rio de Janeiro. “Through these manuscripts, I confirmed that the anonymous flap texts of all of Guimarães’s books had in fact been written by the author himself,” she says. “It’s interesting to note how he first created an image of himself as an author who told stories inspired by his own memories—perhaps his rural background might attract readers who enjoyed the regionalist prose that was in vogue at the time—but gradually he waxed more poetic in introducing his works.”

Saulo Cunha de Serpa Brandão, a retired professor in the Graduate Program in Letters at the Federal University of Piauí (UFPI), believes Brazil is lagging when it comes to authorship attribution methodologies. “Computer-aided stylometry, for example, can be used to detect an author’s writing patterns using software. We can determine how many times a given word is repeated, or whether an author uses the word “bigger” more often than “greater”, for example,” explains Brandão, who has applied these methods since 2003. The software output, however, is not easy to use and interpret, and requires a command of advanced mathematics and statistics that researchers in the humanities do not always possess. Ideally, researchers should work in multidisciplinary teams.”

And software has other limitations, as Brandão learned in his research. In 2003 he began a study on the poem series Cartas chilenas (Chilean letters), whose authorship is attributed to the poet Tomás Antônio Gonzaga (1744–1810). “It has a relatively small amount of text, less than 30,000 words. But these software systems are designed to analyze larger volumes, such as the 900,000 words in the complete works of William Shakespeare [1564–1616]. A large corpus is needed to find patterns,” he notes. “But a computer can’t operate on its own: behind the screen is a researcher devising strategies to capture the author and discover the characteristic traits of their writing. In the poem series in question, we wanted to determine whether other contemporary poets who lived in or around Vila Rica may have participated in writing the text.”

On whether a software system of this kind could have aided her doctoral research, Mônica Gama says: “In my case there would have been little use in looking for style patterns as the texts on Guimarães Rosa’s dust jacket flaps differed completely from is known writing style.” In her analysis Gama used genetic criticism, a method that emerged in France in the 1960s and was introduced in Brazil in the 1980s, which attempts to describe the meanders of an author’s creative process based on written traces or marks left by the artist in the process. “It is not the goal of genetic criticism to prove the authorship of a document, but because we analyze collections in-depth, we are able to evaluate the probability of authorship.”

Technology was recently used in the production of a Brazilian edition of the writings of Antonio Gramsci (1891–1937), in a project started in 2007 and sponsored by the Gramsci Institute Foundation, in Italy. The goal of the project is to establish a definitive and comprehensive Portuguese version of the Marxist thinker’s body of work. One of the major challenges in the project is establishing the authorship of journalistic texts attributed to the Italian from the 1910s and 1920s, most of which were published anonymously. “This is a very important part of Gramsci’s work, and questions of authorship have fueled a decades-long debate among scholars around the world,” says political scientist Alvaro Bianchi, director at the Institute of Philosophy and Human Sciences at the University of Campinas (IFCH-UNICAMP). The republishing of his works is being aided by philological methods, including the use of software. “Philology provides important tools for authorship attribution, but unfortunately these methods are little known in Brazil,” notes Bianchi, who previously managed the Edgard Leuenroth Archive at the same university between 2009 and 2017.

Marcelo Módolo, a professor of philology and the Portuguese language at FFLCH-USP, concurs: “In Brazil, even among academics, philology is often mistakenly associated with linguistic and literary studies of the classical and medieval periods,” he says. But part of the philologist’s task is establishing the authorship of texts, whether they are handwritten or printed, contemporary or old. “This task involves, among other procedures, a careful comparison of the vocabulary and syntax used in texts of unproven authorship with other works for which authorship is well established. This methodology could be particularly useful in analyzing the recently discovered text by Machado, for example, but it’s important to remember that authorship attribution is a world of possibilities, not of certainties,” says Módolo.