A Turning Point in Oral History: From the Digital Recorder to the Times of Artificial Intelligence through Hermeneutics

ABSTRACT

This article analyses the impacts of the digital turn on oral history, from the popularization of the digital recorder in the mid-2000s to Artificial Intelligence technologies. It argues that digital transformations have opened up new methodological, ethical and hermeneutic perspectives. Digitization has enabled advances such as remote interviews and automated transcriptions, but has prompted proposals for advances beyond these practices. Integration with generative AI challenges conventional interpretations, suggesting hybrid approaches. The conclusion is that the digital turn has broadened the horizons of the field, while maintaining a commitment to oral memory and its complexities.

Keywords:
Oral History; Digital Turn; Artificial Intelligence

RESUMO

Este artigo analisa os impactos da virada digital na história oral, desde a popularização do gravador digital, em meados dos anos 2000, até as tecnologias de Inteligência Artificial. Argumenta-se que as transformações digitais abriram novas perspectivas metodológicas, éticas e hermenêuticas. A digitalização possibilitou avanços como entrevistas remotas e transcrições automatizadas, mas suscitou propostas de avanços além dessas práticas. A integração com IA generativa desafia interpretações convencionais, sugerindo abordagens híbridas. Conclui-se que a virada digital ampliou os horizontes do campo, mantendo o compromisso com a memória de expressão oral e suas complexidades.

Palavras-chave:
História Oral; Virada Digital; Inteligência Artificial

Introduction

[...] digital technologies are transforming so many aspects of our work as oral historians [...] how people remember and narrate their lives - that they will, over time, also change the way we think about memory and personal narrative, about telling and storing life stories, and about sharing memories and making stories. This Digital Revolution - the fourth transformation of the oral history paradigm - is still in process, and life on the brink of change in the face of an ever-changing horizon can be uncomfortable. The future of oral history and the role of the oral historian have never been more exciting, or more uncertain (Thomson, 2007, p. 70).

Alistair Thomson’s excerpt shows the beginning of the digital turn in oral history in the first decade of the 21st century in the USA, which was noticed in Brazil a little later (Alberti, 2013, position 1159). Ricardo Santhiago listed some of the multiple turns in the field, considering that

[...] each of these movements is recognizable in the practice of oral history, in its processes and in its products. However, the idea that they constitute “turns” is not only inaccurate, but also contributes to clouding their roles as openings, emphases, inclinations, inaugurations, which add to the accumulation of the practice itself and integrate new elements into a system [...] (2023, pp. 638-639).

Santhiago then mentioned two important articles by Linda Shopes (2014; 2015), stating that Thomson’s understanding of “revolutions” in the “paradigms” of oral history “theory and practice” is methodologically “impoverishing” (Santhiago, 2023, p. 639). Santhiago’s article is full of qualifiers, but it is worth arguing that the term “turns” is “imprecise” and contributes to “clouding”, in the author’s words, the reflection on oral history only when there is still a need to better understand what a turn is and what its impacts have been¹.

Christian Delacroix has argued that historiography is overwhelmed by the avalanche of multiple turning points or different turns that reveal the internal intellectual developments and contexts that can serve to rethink its status (2014, p. 220). It is therefore possible to consider that there is nothing “impoverishing” about turns because they signify self-reflexivity, just as there is nothing impoverishing about the turns mentioned by Thomson.

If Delacroix deduced the turns in French historiographical currents (1999), Zoltan Boldizsár Simon analyzed what these turns meant in the theory and philosophy of history (2019, p. 19). Simon said that if someone intends to “announce a ‘turn’ in historical studies or contribute to one”, it is possible to infer that they are traveling along theoretical paths (2019, p. 20). Whether in historiography as in the case of Delacroix (2014), in the history of historiography (Araújo, 2013), in the theory and philosophy of history (Simon, 2019), or even in oral history, the turns are complex because they occur in the movement of the field with the epistemological or contextual liveliness of the subject itself.

In the American Historical Review forum: Historiographic “Turns” in Critical Perspective, the turns were described by their sophistications: “how should we now understand the various ‘turns’ that have marked the recent history of history writing and theorizing?” (Introduction, 2012, p. 698). The notion that turns are “mere historiographical markers, denoting definable movements or changes in the way some of our professionals have understood, approached and explained their work” has been brushed aside by analysts.

These history professionals who contributed to the debates in the American Historical Review refused to understand the turns as a “simple or unproblematic feature of our profession’s recent past”, submitting them to a “comprehensive and insightful critique that simultaneously reminds us of what can be gained from this historiographical retrospection and also what can be lost by thinking of them in definitive and categorical terms” (Introduction, 2012, p. 698). In this way, announcing a turning point, engaging in it, recognizing it or even acting to support a set of new circumstances that intend to change something means “committing oneself explicitly or implicitly to a ‘theory of history’ in one sense or another” (Simon, 2019, p. 47).

The hypothesis of this article is that (I) a digital turn in oral history was initiated as the digital society advanced for more than two decades of the 21st century, bringing with it (II) the conditions of possibility for the insertion of oral expression memory as a hermeneutic phenomenon rather than a human one by (III) enabling research in the expanding environment of AIs.

For Daniel Woolf, “periodization is both inescapable and indispensable” even though one can “take into account the warnings against fixing periods, which are neither absolute nor phantasmatic” (2023, p. 56). With a sense of unfinished business, the beginning of the digital turn in oral history is established in the mid-2000s with the popularization of the digital recorder and its continuation to the present day due to the expansion of AIs. This is a periodization that imposes the provisionality of the present, estimating the beginning of a movement and not its end.

To assume that there has been a digital turn within Brazilian oral history is not, however, to support the abolition of the ways in which oral history was and still is done, even before the almost ubiquitous relationship of technology in the present. Oral history is still done by getting one’s shoes dirty in villages, quilombos or facing the challenges of radically listening to traumas or contradictory positions that are later criticized. If the dimensions of online life do not have the power to revoke the dimensions of the analog, just as we try to avoid dualisms, we must also frustrate those who look to the digital turn of oral history as an excluding absolute.

IN ADDITION TO REMOTE INTERVIEWS

Oral history research no longer necessarily involves traveling from one university or museum to another. [...] Scholars [...] can access the Internet to obtain information about available recordings. [...] Networked information allows researchers to connect oral history resources to other relevant ones. All of this results in very different expectations of access than just twenty years ago (Boyd; Larson, 2014, p. 4).

What Douglas A. Boyd and Mary A. Larson said can be found in fieldwork and in dealing with oral memory sources, as well as in the constitution of archives or characteristic story banks (Meihy; Seawright, 2020, p. 145). Digital humanities, digital history and digital public history have made similar observations about the changes that have taken place since Web 2.0. The digital society can be delineated as one that “emerges globally from the digital revolution that began in the last decade of the 20th century - and it should be noted from the outset that we can only consider that we have effectively entered” it “when the technological and computer resources disseminated by the digital revolution begin to reach most of the planet’s populations and at all social levels in different ways and in a widespread and decisive manner” (Barros, 2022, p. 11).

If the digital revolution was operationalized in the mid-1990s (Almeida, 2022, p. 101; Barros, 2022, p. 11), in oral history it is possible to trace the beginning of the digital turn to the popularization of digital recorders in the mid-2000s. After appearing between the late 1980s and early 1990s, with the Digital Audio Tape, these devices offered superior sound quality and greater durability compared to analog ones. The miniaturization of the devices, their increased storage capacity and the advent of the MP3 format brought greater technology into the daily lives of researchers. The advantages and concerns surrounding digital recorders, as well as recording media and all the risks regarding the preservation of materials, were discussed from the outset by Donald. A. Ritchie - among others (2003, p. 166).

Verena Alberti said that until the 1990s, the standard was the use of cassette and reel-to-reel tapes. Since then, it has been recognized that digitization is essential for established collections and for new recordings (2013, position 1159). As Boyd and Larson understand, professional work in oral history has changed and has established the possibility of mediation technology through tools that can guarantee the recording of remote interviews. In other words, recordings can be made using AIs to transcribe interviews using transcription software or speech recognition tools. These technologies use Natural Language Processing (NLP) and Automatic Speech Recognition (AVR).

Rather than evaluating each tool by means of a long commute, it is worth reflecting, with Thiago Lima Nicodemo, Alesson Ramon Rota and Ian Kisil Marino, that “digital technologies in people’s lives is a global phenomenon”: “smartphones that fit in the palm of your hand are just a small sample of the place they occupy in everyday life” (Nicodemo; Rota; Marino, 2022, p. 7). All without accessing the marvelous, tracing the charms of the (not) so New Information Technologies, NIT; access to information, real-time communication, automation, cloud storage, big data and data analysis; digital education or history teaching (Costa, 2021). There is a tendency to recognize both the anteriority of digital transformations in relation to the COVID-19 pandemic and the measures aimed at mitigating its effects, in addition to the accelerationism of the pandemic environment, which is often marked by “actualism” (Pereira; Araujo, 2021).

In addition to the seminal work of Daniel Cohen and Roy Rosenzweig (2006) in the United States, it is possible to infer that: “since the epochal experience of the pandemic, this ‘encounter with the new’ can no longer be postponed: the virtualization of the historian’s craft has become an imperative” (Pereira; Marques; Ramalho, 2022, p. 290) - which does not mean neglecting the analogical procedures of the “opération historiographique” (Certeau, 1975, p. 64; Ricoeur, 2000, p. 257).

As can be seen in previous texts (Seawright; Maceno, 2023), the digital turn in oral history did not depend on the COVID-19 pandemic, while it is not possible to ignore its effects or maintain contradictory arguments such as the one that proposes the idea that the documentation of the pandemic would represent the continuity of conventional practices in the field, but at the same time there would be certain ruptures through the intensive use of technology. The proposal to approach these elements as mere recontextualizations - due to empiricist adaptability - shows how the reconfiguration of the digital world can still be underestimated.

In turn, the emphasis on “documentary collecting” demonstrates that the “production of sources” in oral history can relegate closer examination to the bottom of ontological realism’s trunk. Ethan Kleinberg has argued that the “conjuring trick of ontological realism is that it allows the historian to exert his or her own influence on the past by restricting its light to all that it reveals [...]” (2021, p. 47). The selectivity about what is most “urgent” among the urgencies through the metaphor of light employed by Kleinberg indicates that choices remain a realistic “trick” that deals with the abundances of the digital age from the same analog vortex (Gil, 2024, p. 12).

If documenting the COVID-19 pandemic is “urgent”, it is also urgent to treat it with maturity, and then consider the digital conditions that gave rise to it. This requires moving beyond the initial phase of the heuristic of oral sources with a realistic appeal. Everything invites us to overcome the sense of realist or empiricist “urgency” and to become more mature in dealing with traumas stemming from memory attacked by political, economic or health circumstances. After all, is there any documentation that is more “urgent” than other urgent documentation? Is it possible to provoke a competition of catastrophes to measure the pain of others, between wars, pandemics and other traumas that characterize the genre of testimonial oral history (Meihy; Seawright, 2020, p. 75)?

Before the COVID-19 pandemic, there were differences of opinion about conducting remote interviews. There were those who advocated remote interviews in all circumstances and those who advocated remote interviews when face-to-face interviews could not be carried out. There were also those who did not accept remote interviews at all because of the “warmth of the encounter” as the raw material for verbalized history. Finally, there were those who wanted - in the case of multiple interviews with the same narrator - the first meeting to be face-to-face and the others to take place remotely.

During the COVID-19 pandemic, adherence to remote interviews has increased on the grounds that this is the way to make research projects feasible. Adriena Casini da Silva set out to understand the professional drama that struck oralists. How to conduct interviews in times of social isolation, under degraded political circumstances, with few financial resources and under “vaccine denialism” when immunizers were available for purchase/application?

According to Silva, his research was “based on the methodological stages that constituted the writing of a doctoral thesis” on “educational and institutional policies that motivated the expansion of Colégio Pedro II (CPII), in the 2000s, starting from an investigation of the History of Education in the present time”: the “interviews were carried out remotely [...] considering the challenges imposed by social isolation, in the context of the Covid-19 pandemic, since March 2020” (Silva, 2022, p. 154).

Experiencing the need for “technological adaptations”, Silva asked: “how can we do Oral History online without losing nuances and details?” (2022, p. 154). The researcher adjusted her schedule to digital dynamics, especially in the context of pandemic temporality in the “relationship between what is proper to history” (in this case, oral history), “what is proper to computing” and how these domains interact (Silveira, 2022, p. 222). The distance imposed by the remote circumstance required the use of digital tools such as Zoom, WhatsApp, Google Meet, among others (Silva, 2022, pp. 159-160).

If it is true that “conducting Oral History interviews during the period of social isolation imposed by the pandemic led us to break down hitherto invisible barriers in the use of technology and the suitability of methodology” (Silva, 2022, p. 169), it is unreasonable to neglect previous experiences demonstrated by Anita Lucchesi - a pioneer of Digital Public History in Brazil.

Lucchesi mentioned three projects: “Memoro - La Banca della Memoria”, in Turin, which collected life stories of people born before 1950, highlighting the “oral tradition” and elders as narrators with audio or video interviews uploaded to the web and organized on an interactive platform (2014, p. 40). The second, “Herstories”, is an online archive of Sri Lankan women’s stories, with spoken and visual narratives, such as “trees of life”, and a space for “venting” about the civil war (Lucchesi, 2014, p. 43). The third is the Museu da Pessoa, in Brazil, which combined oral history and digital collaboration, with more than 16,000 stories to create a “social technology of memory” (Lucchesi, 2014, p. 45).

Agreeing with Lucchesi on the importance of the three initiatives, as well as others, and highlighting the heuristic element as a key to oral history, it remains to ponder methodological improvements in the era of Generative AI. Is it feasible to overcome in the field of oral history the initial and quite obvious understanding that remote interviews are a possibility? The answer is affirmative, above all because it is a recurring practice among different oral history professionals, but also because there seems to be a certain saturation with archival excesses and empiricism that calls for reflection.

BEYOND THE HUMAN

It is in the interaction between humans and computers - “human-computer interactions”, or HCI - that I seek to understand the emergence of relational experiences that constitute what I call a “more than human” history [...] the growing autonomy of artificial agents as mediators of social relations [...] provided by machine learning, has suggested the problematization of a more than human condition [...] (Bonaldo, 2023, p. 3).

In addition to the text by Nicodemo and Oldimar Cardoso entitled Metahistory for (Ro)bots: Historical Knowledge in the Artificial Intelligence Era (2019), Rodrigo Bragio Bonaldo contributed to thinking about the temporal cadences of humans and computers: “human-computer interactions”. From the popularization of the digital recorder to AIs, part of Bonaldo’s proposal may also be possible in the methodological field of oral history. This proposal refers, from one perspective, to the transposition of a certain fetish for the production of sources, and, from another, to reflection on hermeneutics in the times of AIs.

It is clear that discussions within oral history are still in their infancy when it comes to AI. Some members of the International Oral History Association (IOHA) and the Brazilian Oral History Association (ABHO) are waiting for greater openness to the digital theme. At the 22nd IOHA International Conference, which took place in Rio de Janeiro between July 25 and 28, 2023, the important theme was chosen: Oral History in a Digital and Audiovisual World. For several speakers, however, the “digital world” would be an expression for what is current and what provokes some practitioners in the field. It is therefore significant that the topic of AI is somewhat absent from the work proposals, even though, it should be noted, several initiatives have addressed elements of the digital society.

One of the central aspects of the program was the roundtable Oral History in a Digital and Audiovisual World, the theme of the IOHA Conference held in July 2023. Participants included Elizabeth Visser (University of Cape Town/South Africa), Jun Oguro (Doshisha University/Japan) and Sumallya Mukhopadhyay (International Institute of Research and Studies/India), moderated by Mayara Jucá (FGV CPDOC/Brazil). The focus of the discussion was to reflect on “ways of doing oral history in the face of the digital and imagetic challenges of the first decades of the 21st century”, based on “trajectories of researchers around the world” (22nd IOHA International Conference, 2023)².

Established journals such as The Oral History Review reveal a certain vacuum in discussions about artificial intelligences - somewhat understandable given the emergence and improvement of this tool, with the use of algorithms or computer models in recent years. Noting the excessive caution with which the Associazione Italiana di Storia Orale, AISO, has treated digital material, one can see that this dynamic is not exclusive to ABHO or the international community of oralists. In the Italian case, there is an attachment to empiricist ways of doing oral history, as shown by most of the publications in the Il de Martino periodical. Storie, voci, suoni.

Alessandro Casellato pointed this out by treating the so-called “positivism” in relation to oral sources with less conceptual precision, but also by revealing a practice in oral history that is still dependent on the first phase of something that resembles the aforementioned historiographical operation. This was shown by the recognition that “outside the university” they carry out “campaigns to record interviews” in a “naive and positivist way, without the tools to interpret them properly” (Casellato, 2023, p. 157).

If the public experience of oral history is something that can be verified in Casellato’s argument, as is also the case in Brazil, it is possible to understand the exhaustion of the idea that oral history only serves to produce sources: archive, archive, archive. This perception, which gained momentum with the dissemination of Pierre Nora’s work (1996; 1997; 1998) - based on heritage and commemorative motivations - needs to be investigated when we look at the conceptions of verbalized memory. Criticism can be made in different ways, taking for granted the effectiveness of Ricoeur’s phenomenology of memory, especially in La Mémoire, l’Histoire, l’Oubli (2000, p. 534).

Of the desire for the remains of the past through the production of the archive as a “product of the time”, Ricoeur said: “it is somewhat in a tone of imprecation that Nora exclaims: Archive, archive, there will always be something left!” (2000, p. 213). The archive as a “voluntary secretion” (Ricoeur, 2000, p. 525), however, has its virtues, its powerful social dynamics and, equally, its limits as a way of saving data, information or even “redeeming the past”. Harald Weinrich’s expression in Lete: Art and Critique of Forgetting - “stored, that is, forgotten” (2001, p. 281).

From the philosopher de Valence’s sophisticated critique of Nora’s environment, it follows that the obsession with accumulation in oral history archives reflects nothing other than the desire for alternative redemption of the past. The nobility of demonstrating variable, dissonant and less heard voices in oral history is sometimes lost in the archival appeal to the ontology of accumulation, which is a form of realism. Assuming that it is impossible to capture the past is not something new. Just like the past in its passability (Ricoeur, 1998), memory is not just gain. It is also loss. And it is because of this loss and the forms of forgetting that we can archive and then doubt the sense of sufficiency of our accumulations. It is the sense of incompleteness that qualifies the archive, the interview, the library, the memory; not the glamor of sufficiency. We can then paraphrase: interview, interview, there will always be someone left!

It is not a recent realization that oral history has outgrown the initial phase in which it claimed to “give voice” to the “voiceless”; its progress towards “giving ears” to the unheard has been significant. While recognizing these inclinations, it was necessary to question the saturation that invited the qualitative task. Methodological maturity went beyond the mere filling of “documentary gaps” or the hackneyed “memory retrieval”.

Boyd and Larson said of oral history over the first two decades of the 21st century that the internet: “blew the hinges off the doors of archives, and access took on a whole new meaning” (Boyd; Larson, 2014, p. 4). Blowing the “hinges off archive doors” and spreading oral histories is remarkable, as it is part of “democratizing oral histories”. It seems unreasonable to speak only to peers, because peers are invited to do more than “consume stories”: they are invited to feel attracted to their significant cores: they are invited to tell good stories.

People are interested in people’s stories. Some of these people are interested in narratives about the weather, animals, viruses, plants and inhuman artifacts like computers. Even machines are dialoguing with anthropomorphic expressions, according to Bonaldo: “computer memory”, “machine learning”, “hallucination” and, above all, “artificial intelligence”. For Bonaldo, this could be an example that “possibly reinforces the suspicions of neopositivism that arise among historians” (Bonaldo, 2024, p. 144). In any case, the mathematical foundations of AI programming and engineering point to the extrapolation of human temporality in the same way that the tendency of oral history is to overcome the accumulation of stories as if they were just scraps like any other scraps: the ontological accumulation of life, of the rest of human life.

Oral history has never needed to declare something similar to what Marc Bloch wrote: “science des hommes dans le temps” (Bloch, 1952, p. 29). It didn’t need to because it emerged in an environment centered on the human; on overcoming wars, dictatorships and political censorship (Gattaz; Meihy; Seawright, 2019, p. 11). Oral history would therefore be nothing other than sensitive and humanizing. For oral history, homocentrism - as Dipesh Chakrabarty put it - has almost always prevailed, since human beings obviously speak, and if they speak, they have the primacy to tell their own story or the story of others.

Speech would not be possible without its “other” and the “other” of its “other”: the various forms of life and things. It wouldn’t even be possible to think of oral history without the projection of speech, which is nothing other than a form of prospection: from the inside out. In other words, a prospection that becomes a prospection because, before speech, there is an expectation of speech. Being a talking animal, however, shouldn’t mean neglecting other animals; things that don’t speak or technology that “speaks differently”, at the risk of provocatively reinventing the unlikely formula: I speak, therefore I am. Although AIs can generate images, sounds and speech itself, human speech seems to indicate other existences that are more than human, such as its artificial reproduction.

How to interpret them?

If temporal and mathematical quantification are lapidary for the informational society, the quality of the analysis has the potential to instruct more robust proposals.

In this regard, Stephen M. Sloan analyzed that

With the push towards quantitative analysis [...] qualitative understanding is more necessary than ever. Even researchers who defend the vast potential of Big Data research, such as Mayer-Schönberger and Cukier, warn of the dangers of a “data dictatorship”. [...] Quantitative analysis offers the opportunity for every voice to be heard, to find the exception, the nuance of existence and the texture of reality (Sloan, 2014, p. 184).

The “dictatorship of data” or even the “primitive accumulation of data” are elements that imply a critique of digital colonialism or its power of profitability/extractivism (Lippold; Faustino, 2022, p. 63). Argumentative critiques such as those by Walter Lippold and Deivison Faustino, or even problems with the ethics of algorithms (Mittelstandt et al., 2016; Floridi, 2022) depend on interpretative parameters that can hybridize what is considered here to be a certain human and more than human hermeneutics.

The attitude of neo-Luddism in the 21st century, of attacking machines, doesn’t seem to apply when what is proposed has a qualitative slant, including in the case of oral history. Several texts were written - at the time of the institutionalization of oral history in Brazil - about the importance of daydreams, night dreams or waking dreams; worldviews, beliefs and even lies or inaccuracies.

Such questionable materials for a number of historians locked in the cult of written sources were virtuous for oralists who intended to embrace subjectivities. Because memory is almost always malleable and spiral, it is unreasonable to criticize it with the same parameters as those used to criticize written sources from the outset. Operators in oral history have employed hermeneutic parameters, even under the more usual name of “story analysis”; they question impostures, denials, bad faith, hoaxes. But the feeling of welcoming imperfections seems to be provoked in the current environment of consolidation of Generative AI.

Generative AI can be understood as a new media or as a tool for integrating social relations. Bonaldo analyzed these models, emphasizing their virtues and potential for improvement in the face of failures. On this subject, he presented the oral communication: “Three cheers for the shipwreck: on the rise and fall of digital humanities”, at the XII Brazilian Seminar on Theory of History and History of Historiography (2024). The metaphor of the shipwreck, which symbolizes the failure of AIs, is not seen as annihilation, but as a stimulus to “double the bet”: instead of just declaring the “technological limitations in the face of the superiority of human interpretation”, we can improve models with parameters more adapted to our data (2024).

Just as in oral history narratives go through daydreams, distortions and deviations from the teleological meaning of verbalizations, machine learning does not prevent “hallucination” - which happens when “the connection between the data and its conclusions may not be obvious, or even accessible, if not mistaken” (Bonaldo, 2023, p. 13). The possible encounter, yet to be made in a more specialized way, between oral history and Generative AI does not ensure a territory of large blocks of perfection in which inerrancy is sought as confirmation of an ideal god-like model.

TWO BRIEF DETOURS AND A PROBLEM

The symbol makes you think.

(Ricoeur, 1969, p. 46)

NLP allows computers to understand, interpret and respond to human language, bypassing the temporal relationship between humans and machines (Caseli; Nunes, 2023, p. 12). With the use of Word Embeddings it is possible to mathematically represent words, allowing the machine to understand expressions with similar but not identical meanings (Mikolov et al., 2013). The Transformer architecture improves these processes by contextualizing words within sentences and improving accuracy in interpreting or generating texts (Vaswani et al., 2023). In the aporia between the representation operated by AIs, and the world returning to things, the possibility of another hermeneutic is established that tensions phenomenology and the world of things themselves.

Ewa Domanska, Simon and Dominick LaCapra, among others, have considered the category of “non-anthropocentric humanities”, “post-anthropocentrism” or even “post-humanism” to designate something more than human (Domanska, 2010; Simon, 2021; LaCapra, 2018). The “non-anthropocentric humanities” are like an “institutionalized set of topics, techniques and research interests”, from which an ethos is derived from the “intellectual movement and ethical stance called posthumanism”. (Domanska, 2010, pp. 118-119).

It doesn’t mean that human beings are superseded as if they were machines when they reach obsolescence, nor does it mean that the dystopia of machines dominating people in a kind of hell without redemption is reiterated. As LaCapra explained, this means that the humanities have become “a field or conglomerate of fields that is not exclusively centered on the human being” (2018, p. 62), but which may not exist without him. For LaCapra, this conception includes “artificial intelligence (AI), robots, genetic engineering and the advent of futuristic possibilities” (2018, p. 64).

In an attempt to overcome anthropocentrism by opening up to the relationship between humans who tell stories about other humans and non-humans; things, animals, plants, objects or AIs, it is appropriate to raise the issue of hermeneutics. In other words, with temporal cadences that are more than human and the aforementioned return to things, as in Domanska (2006, p. 338), would it still be feasible to employ hermeneutic theoretical-methodological foundations?

Digital hermeneutics was addressed by Murilo Gonçalves dos Santos in his doctoral thesis and in scientific papers (2022). The author argued that digitalization has altered the ontological and epistemological foundations of history, changing the conditions in which we can know. In addition, he went through the thinking of the classic schools of hermeneutics - from Droysen and Dilthey to the thinkers of digital hermeneutics (2023, p. 138). The author therefore pointed to Heideggerian thought as a turning point for the sciences of interpretation. For him, hermeneutics operates through the need for human interpretation, at which point it can be articulated with digital technologies - which occurs, according to the author, through research and technological ubiquity as an “active agent of human experience” (2023, p. 146). The correlation between the “active agent”, cognitively coded or decoded, and the meaning of the temporalized “human experience” stands out (Pimenta, 2024, p. 157).

When examining authors of digital hermeneutics and their limitations, Murilo Gonçalves dos Santos highlights the stance of “de-emphasizing narrative” (Hayles, 2012, p. 47) and considers the human creation of programs, as well as data processing. Referring to Manovich (2013), he argues that the digital hermeneut can subject code to critical scrutiny, including software, algorithms and cultural artifacts. Thus, more than the product of Generative AI, the coding and programming process is part of digital hermeneutics.

It is worth proposing two questions that serve as brief Ricoeurian detours as contributions to oral history, but also to the theory of digital history. The next question is the epistemological limits of hermeneutics, i.e. the problem of whether it is possible to invest in the limits of interpretation, just as the limits of narrative are known. In order to briefly appreciate the limits of interpretation and narrative, two short detours are needed.

The first deviation in Ricoeur’s analysis (1990, p. 20) refers to the return to Friedrich Schleiermacher’s concept of the “divinatory method”, which defended the idea that the interpreter could understand the author better than he understood himself. In this sense, could AI, through machine learning, invert this process and examine the programmer or narrator more deeply than the oral history professional himself could understand their stories?

However, as Ricoeur pointed out, the emphasis should be shifted from subjectivity to the meaning and reference of the text itself (Schmidt, 2014, p. 214). The memory of oral expression, although it is often transcribed or textualized (Meihy; Seawright, 2020, p. 139), has a distinct codification from conventional textual manifestations. Generative AI, on the other hand, generates texts from an autonomous process which, although it requires supervision, criticism and ethical assumptions, generates texts that can be considered as independent objects.

An example of this last case can be seen in models such as the GPT (Generative Pre-trained Transformer), which, when generating texts, not only follows learning patterns from large volumes of data, but responds to commands in a way that may seem autonomous. This type of AI, for example, generates responses in natural language without the need for direct intervention, but it still depends on criteria established by its programmers to guarantee relevance and values on the generated content.

The second deviation demonstrates Ricoeur’s re-reading of Dilthey’s work (1990, p. 23). According to Dilthey, there is a bifurcation between explanation and understanding - which Ricoeur contested with a “disastrous move” (Schmidt, 2014, p. 215): “[...] Dilthey needs to abandon the psychological notion of transference and interpret the text from its own meaning” (Schmidt, 2014, p. 215). The relocation and continual questioning of the strictly psychological aspects of understanding led Ricoeur to re-evaluate the hermeneutic circle - the understanding of the part by the whole and the whole by the part - by moving away from the primacy of psychological interpretation. Ricoeur did not give up on the meaning of the text itself, the textual other, as well as the unveiling of the open world in critical reading.

Against this argument, Ricoeur changed the relationship between the producer and the text by arguing that “instead of looking for the author in the text, it is a matter of exploring the meaning and the world discovered by the text itself” (Ricoeur, 1990, p. 26). Although oral history narratives and generative AI products, when put into relation, are linked to authorship, co-authorship when it comes to interviews, or rather programming, research does not always return to the author’s experience. Conversely, just as the text becomes autonomous and distinct, the stories are worthwhile in and of themselves, and can be criticized or pondered by those who analyze them.

Similarly, the hermeneutic task of the generative character of AI has intrinsic value insofar as AI itself acts in symbolic places (Bonaldo, 2024), but can then be made available as another type of support. It is worth realizing that an interpretation does not have to be probable at all costs, but more probable than another interpretation (Ricoeur, 1976, p. 78). Ricoeur said that, in scientific language, a “model is essentially a heuristic procedure that serves to demolish an inadequate interpretation and open the way to a new and more adequate interpretation” (1976, p. 78). Thus, an oral history interview and an AI model can take the place of a text in Ricoeurian theory, above all by acting as mediators “between explaining and understanding” (Ricoeur, 1990, p. 11).

Verbalization in contact with mathematical structures of the more-than-human world, machine training data, technological generalizations, statistical models, neural networks, learning supervisions, challenge the binarism between what is explained, “Erklären”, and what is understood, “Verstehen” (Ricoeur, 1990, p. 11). There is a hermeneutic relationship of collaboration between humans and more than humans.

Finally, the problem that can be posed in the aporia between a hermeneutic with a phenomenological matrix and a return to the world of things in more than human terms, unfolds in two ways. On the one hand, it should be recognized that Heidegger redefined hermeneutics as an ontological investigation. By introducing the concept of Dasein - as being-in-the-world - Heidegger centered the human in its historicity, since it doesn’t just act as an interpreter, but is linked to the world it proposes to interpret. Interpretation, for Heidegger, “does not consist in acquiring knowledge of what is understood, but in the elaboration (Ausarbeitung) of the possibilities projected in understanding” (2014, p. 421).

Heidegger inaugurated the ontological turn by shifting the question of how we understand to the analysis of the meaning of being. Focusing on being and the being that understands, he proposed that cognition is part of the structure of being. For Ricoeur, Heidegger’s central question is: “Instead of asking how we know, we must ask what is the mode of being of this being that exists only by understanding” (Ricoeur, 1990, p. 30). Despite the advances in ontology, however, Ricoeur pointed out that Heideggerian philosophy, by turning to the fundamentals, prevented a return to the fundamental epistemological question (Ricoeur, 1990, p. 36).

If there are evidently ontological concerns for oral history in the era of generative AI, there are also basic problems linked to narratability in the time of algorithms; of a touch memory marked by multiple forms of learning. Machine learning itself, and the autonomization of stories through textualization or transcreation, call for methodological reinforcements. Ricoeur was afraid of not returning to the foundation of things, although he recognized the virtues of Heideggerian ontological thinking.

From another angle, the projects between oral history and AI can lead to the question of the limits of hermeneutics. Heidegger’s Origin of the Work of Art was “fundamental to Gadamer’s reflections in Truth and Method” (which Ricoeur never ceased to problematize), and “also in The Actuality of Beauty”, according to Hans Ulrich Gumbrecht and Thamara de Oliveira Rodrigues (2021, p. 38). Koselleck approached Gadamer, his teacher, in such a way as to assume that the “notion of ‘fusion of horizons’ and that of ‘strata of time’ share the understanding of the coexistence of diverse temporalities”, but distanced himself from him when he valued the “pre- and extralinguistic structures that condition history, which means that his philosophy of time could not be understood exactly as a part of hermeneutics” (Gumbrecht; Rodrigues, 2021, p. 39).

Without rejecting hermeneutics, Koselleck identified experiences that could not be fully grasped by language; by tradition or historiography or because the “non-verbal” element imposed itself through the limits of expression (2014, p. 91). Contrary to what one might think, in oral history the silences matter as much as the bursts of speech: memories are imperfect and wandering, allowing themselves to be constituted by subjectivities or sinuosities. More than being stopped by language, Koselleck’s astonishment at the dreams of Jews during the Third Reich (2006, p. 247), for example, imposes not only possible limits from a reading of the Kantian sublime or Freudian psychoanalysis (LaCapra, 2020, p. 29). It also imposes the terror that instructs situations triggered by the absurd, as in Sediments of Time (Koselleck, 2018, p. 177). Koselleck set limits on the absurd, but - by setting limits on language - he didn’t give up the narration that makes it at least comprehensible. After all, as Ricoeur says: “the symbol gives us something to think about” (1969, p. 46).

FINAL CONSIDERATIONS

The interaction between oral history and digital society reveals innovative perspectives, demanding a theoretical-methodological balance. Digital technologies make listening, sharing and analysis more dynamic, connecting narratives to broader contexts. Ethical and hermeneutical challenges remain. Hermeneutics in this field values subjectivities, silences and mnemonic contradictions or dimensions that are irreplaceable by AI. Meanwhile, NLP and machine learning offer support, but may require human supervision and deontological standards. Even so, they optimize processes, explore complex relationships and broaden the scope of less obvious narratives.

Despite “shipwrecks” (Bonaldo, 2024), the critical interaction between AI and oral history points to a time of hybridization. Listening and ethics remain essential, while AI asserts itself as a mediator and enhancer of protocols. However, the human dimension of memory remains central to ensuring that, even in the age of algorithms, oral history remains a space for empathy and social transformation.

Almost two decades on from Thomson’s publication, perspectives have changed. Bringing it to a close seems appropriate: “artificial intelligence projects” will allow “anyone, anywhere, to make extraordinary and unexpected creative connections within and between oral history collections, using sound and image in addition to text” (2007, p. 68). Furthermore, “computer-assisted qualitative data analysis can already be used to support, extend and refine the interpretation of large sets of oral history interviews” (Thomson, 2007, p. 68). Acknowledging Shopes’ critical sophistication, it is worth affirming, however, the unfinished turn within oral history - from the digital recorder to AI: there is nothing “impoverishing” about it, as Santhiago (2023, p. 639) said.

References

• ALBERTI, Verena. Manual de história oral. 3ª Ed. Rio de Janeiro: Editora FGV, 2013. E-book - versão Kindle.
• ALMEIDA, Fábio Chang de. Internet, fontes digitais e pesquisa histórica. In: BARROS, José D’Assunção (Org.). História Digital: a historiografia diante dos recursos e demandas de um novo tempo. Petrópolis: Vozes, 2022. pp. 101-120.
• ARAÚJO, Valdei Lopes de. História da historiografia como analítica da historicidade. História da Historiografia, Ouro Preto, v. 6, n. 12, pp. 34-44, ago. 2013.
• BARROS, José D’Assunção. Revolução digital, sociedade digital e História. In: BARROS, José D’Assunção (Org.). História Digital: a historiografia diante dos recursos e demandas de um novo tempo. Petrópolis: Vozes , 2022. pp. 11-135.
• BOYD, Douglas A.; LARSON, Mary A. Introduction. In: BOYD, Douglas A.; LARSON, Mary A. Oral History and Digital Humanities: Voice, Access, and Engagement. New York, NY: Palgrave Macmillan, 2014. pp. 1-18.
• BLOCH, Marc. Apologie pour l’histoire ou métier d’historien. Paris: Librairie Armand Colin, 1952.
• BONALDO, Rodrigo Bragio. História mais do que humana: descrevendo o futuro como atualização repetidora da Inteligência Artificial. História, São Paulo, v. 42, pp. 1-28, 2023.
• BONALDO, Rodrigo Bragio. Três vivas ao naufrágio: sobre a ascensão e queda das humanidades digitais. In: SESQUIM, Ilda Renata Andreata et al. (Orgs.). Hoje, Teoria da História. Belo Horizonte: Fino Traço, 2024, pp. 139-152.
• CASELI, Helena de Medeiros; NUNES, Maria das Graças Volpe (Orgs.). Processamento de Linguagem Natural: conceitos, técnicas e aplicações em português. São Carlos: BPLN, 2023.
• CASELLATO, Alessandro. História oral na Itália: trajetórias e desafios. História Oral, v. 26, n. 3, pp. 155-167, set./dez. 2023.
• CERTEAU, Michel de. L’Écriture de l’histoire. Paris: Gallimard, 1975.
• COHEN, Daniel J.; ROSENZWEIG, Roy. Digital History: A Guide to Gathering, Preserving, and Presenting the Past on the Web. Philadelphia: University of Pennsylvania Press, 2006.
• COSTA, Marcella Albaine Farias da. Ensino de história e historiografia escolar digital. 1ª Ed. Curitiba: Editora CRV, 2021.
• DELACROIX, Christian. “Écoles”, “paradigmes”, “tournants”, “ruptures”: les embarras de la périodisation en historiographie. ATALA Cultures et sciences humaines, n. 17, pp. 219-231, 2014.
• DELACROIX, Christian; DOSSE, François; GARCIA, Patrick. Les courants historiques en France: XIXe-XXe siècle. 1ª Ed. Paris: Armand Colin, 1999.
• DOMANSKA, Ewa. Beyond Anthropocentrism in Historical Studies. Historein: A Review of the Past and Other Stories, v. 10, pp. 188-130, 2010.
• DOMANSKA, Ewa. The Material Presence of the Past. History and Theory, v. 45, n. 3, pp. 337-348, 2006.
• FLORIDI, Luciano. Etica dell’intelligenza artificiale: Sviluppi, opportunità, sfide. Torino: Raffaello Cortina Editore, 2022.
• GATTAZ, André; MEIHY, José Carlos Sebe Bom; SEAWRIGHT, Leandro. A História oral nos vãos da democracia. In: GATTAZ, André; MEIHY, José Carlos Sebe Bom; SEAWRIGHT, Leandro (Orgs.). História oral: a democracia das vozes. São Paulo: Pontocom, 2019. pp. 9-16.
• GIL, Tiago Luís. Sobre Big Data e Neopositivismo Digital na Pesquisa em História. Almanack, Guarulhos, n. 36, pp. 1-19, 2024.
• GOMES, Angela de Castro (Org.). História Oral e Historiografia: questões sensíveis. São Paulo: Letra & Voz, 2020.
• GONÇALVES, Murilo. The Hermeneutic Problem Posed by Digital Humanities. Revista de Teoria da História, v. 25, n. 1, pp. 136-150, 2022.
• GONÇALVES, Murilo. Towards a Historical Hermeneutics of Digital Interface. Expedições, Morrinhos, v. 16, pp. 118-139, jan.-jun. 2023.
• GUMBRECHT, Hans Ulrich; RODRIGUES, Thamara de Oliveira. Reinhart Koselleck: Uma latente filosofia do tempo. São Paulo: Editora Unesp, 2021.
• HAYLES, Nancy Katherine. How We Think: Digital Media and Contemporary Technogenesis. Chicago: University of Chicago Press, 2012.
• HEIDEGGER, Martin. Ser e Tempo. Campinas, SP: Editora da Unicamp; Petrópolis, RJ: Editora Vozes, 2014.
• INTRODUCTION. The American Historical Review, v. 117, n. 3, pp. 698-699, 2012.
• KLEINBERG, Ethan. Historicidade Espectral: Teoria da história em tempos digitais. Vitória: Editora Milfontes, 2021.
• KOSELLECK, Reinhart. Estratos do tempo: estudos sobre história. Rio de Janeiro: Contraponto: PUC-Rio, 2014.
• KOSELLECK, Reinhart. Futuro passado: contribuição à semântica dos tempos históricos. Rio de Janeiro: Contraponto: PUC-Rio , 2006.
• KOSELLECK, Reinhart. Sediments of Time: On Possible Histories. California: Stanford University Press, 2018.
• LACAPRA, Dominick. Traumatropismos: do trauma ao sublime pela via do testemunho?. In: FREDRIGO, Fabiana de Souza; GOMES, Ivan Lima (Orgs.). História & Trauma: Linguagens e Usos do Passado. Vitória: Editora Milfontes , 2020. pp. 29-70.
• LACAPRA, Dominick. Understanding Others: Peoples, Animals, Pasts. Ithaca, New York: Cornell University Press, 2018.
• LIPPOLD, Walter; FAUSTINO, Deivison. Colonialismo digital, racismo e a acumulação primitiva de dados. Germinal: Marxismo e Educação em Debate, v. 14, n. 2, pp. 56-78, ago. 2022.
• LUCCHESI, Anita. Conversas na antessala da academia: o presente, a oralidade e a história pública digital. História Oral , v. 17, n. 1, pp. 39-69, jan./jun. 2014.
• MANOVICH, Lev. Software Takes Command. New York: Bloomsbury Academic, 2013.
• MEIHY, José Carlos Sebe B.; SEAWRIGHT, Leandro. Memórias e Narrativas: História Oral Aplicada. São Paulo: Contexto, 2020.
• MIKOLOV, Tomas et al. Efficient Estimation of Word Representations in Vector Space. arXiv:1301.3781v3 [cs.CL], pp. 1-12, 7 set. 2013.
• MITTELSTADT, Brent Daniel et al. The Ethics of Algorithms: Mapping the Debate. Big Data & Society, pp. 1-21, Jul./Dec. 2016.
• NICODEMO, Thiago Lima; CARDOSO, Oldimar Pontes. Metahistory for (Ro)bots: Historical Knowledge in the Artificial Intelligence Era. História da Historiografia: International Journal of Theory and History of Historiography, Ouro Preto, v. 12, n. 29, pp. 17-52, jan./abr. 2019.
• NICODEMO, Thiago Lima; ROTA, Alesson R.; MARINO, Ian Kisil. Introdução: das humanidades digitais à história digital. In: NICODEMO, Thiago Lima; ROTA, Alesson Ramon; MARINO, Ian Kisil. Caminhos da história digital no Brasil. 1ª Ed. Vitória, ES: Editora Milfontes, 2022. pp. 6-29.
• NORA, Pierre (Dir.). Les lieux de mémoire. Tome 1: La République. Paris: Gallimard , 1996.
• NORA, Pierre (Dir.). Les lieux de mémoire. Tome 2: La Nation. Paris: Gallimard , 1997.
• NORA, Pierre (Dir.). Les lieux de mémoire. Tome 3: Les France. Paris: Gallimard , 1998.
• PEREIRA, Mateus Henrique de Faria; ARAUJO, Valdei Lopes de. Atualismo: pandemia e historicidades no interminável 2020. Estudos Ibero-Americanos, Porto Alegre, v. 47, n. 1, pp. 1-16, jan./abr. 2021.
• PEREIRA, Mateus; MARQUES, Mayra; RAMALHO, Walderez. O atualismo chega à História? Virtualização do ofício do historiador durante a pandemia de Covid-19. In: NICODEMO, Thiago Lima; ROTA, Alesson Ramon; MARINO, Ian Kisil. Caminhos da história digital no Brasil. 1ª Ed. Vitória, ES: Editora Milfontes , 2022. pp. 277-302.
• PIMENTA, João Paulo. Vivências e experiências: uma nota acerca dos tempos da história e suas hierarquias. In: MARQUESE, Rafael de Bivar et al. (Orgs.). Sistemas, tempos e espaços: o Lab-Mundi em dez anos de fazer historiográfico. São Leopoldo: Casa Leiria, 2024. pp. 156-162.
• RICOEUR, Paul. Le conflit des interprétations: essais d’herméneutique. Paris: Éditions du Seuil, 1969.
• RICOEUR, Paul. Interpretação e ideologias. Rio de Janeiro: F. Alves, 1990.
• RICOEUR, Paul. La marque du passé. Revue de Métaphysique et de Morale, n. 1, pp. 7-31, jan./mar. 1998.
• RICOEUR, Paul. La mémoire, l’histoire, l’oubli. Paris: Éditions du Seuil , 2000.
• RICOEUR, Paul. Teoria da interpretação: discurso e o excedente de significado. Rio de Janeiro: Imago, 1976.
• RITCHIE, Donald A. Doing Oral History: A Practical Guide. Oxford: Oxford University Press, 2003.
• SANTOS, Murilo Gonçalves dos. A História (de)codificada: prolegômenos para uma hermenêutica digital. Tese (Doutorado em História) - Faculdade de História, Universidade Federal de Goiás. Goiânia, 2022.
• SANTHIAGO, Ricardo. De volta ao “para uso futuro”? História oral, pandemia e a documentação urgente do presente. História Unisinos, v. 27, n. 3, pp. 636-648, set./dez. 2023.
• SCHMIDT, Lawrence K. Hermenêutica. Petrópolis: Vozes , 2014.
• SEAWRIGHT, Leandro; MACENO, Lucas. História oral e sociedade digital. Projeto História, São Paulo, v. 78, pp. 503-533, set./dez. 2023.
• SHOPES, Linda. After the Interview Ends: Moving Oral History Out of the Archives and into Publication. The Oral History Review, v. 42, n. 2, pp. 300-310, Summer/Fall 2015.
• SHOPES, Linda. Insights and Oversights: Reflections on the Documentary Tradition and the Theoretical Turn in Oral History. The Oral History Review , v. 41, n. 2, pp. 257-268, Summer/Fall 2014.
• SILVA, Adriena Casini. Um zoom nos desafios metodológicos de fazer História Oral em tempos de pandemia: confluências e adaptações tecnológicas de uma investigação de História da Educação. História Oral , v. 25, n. 2, pp. 153-172, 2022.
• SILVEIRA, Pedro Telles da. O que é uma ferramenta historiográfica? História da Historiografia , Ouro Preto, v. 15, n. 40, pp. 219-231, 2022.
• SIMON, Zoltán Boldizsár. Os teóricos da História têm uma teoria da história?: Reflexões sobre uma não-disciplina. Tradução de Arthur Lima de Avila. Vitória: Editora Milfontes , 2019.
• SIMON, Zoltán Boldizsár; TAMM, Marek. Historical Futures. History and Theory , v. 60, n. 1, pp. 3-22, 2021.
• SLOAN, Stephen M. Swimming in the Exaflood: Oral History as Information in the Digital Age. In: BOYD, Douglas A.; LARSON, Mary A. Oral History and Digital Humanities: Voice, Access, and Engagement. New York, NY: Palgrave Macmillan , 2014. pp. 175-186.
• THOMSON, Alistair. Four Paradigm Transformations in Oral History. The Oral History Review , v. 34, issue 1, pp. 49-70, 2007.
• VASWANI, Ashish et al. Attention is all you need. 31st Conference on Neural Information Processing Systems (NIPS 2017), Long Beach, CA, USA, 2017. arXiv:1706.03762v7 [cs.CL], pp. 1-15, 2 Aug. 2023.
• WEINRICH, Harald. Lete: arte e crítica do esquecimento. Rio de Janeiro: Civilização Brasileira, 2001.
• WOOLF, Daniel. Historical Periodization: An Exploration and Defence. In: ZWIERLEIN, Cornel; MAHLER, Andreas (Eds.). Zeiten bezeichnen: frühneuzeitliche Epochenbegriffe: europäische Geschichte und globale Gegenwart. Wolfenbüttel: Herzog August Bibliothek, 2023. pp. 29-58.

• 1
  The Oral History and Dimensions of the Public collection, containing 17 volumes (Almeida, 2016-2025), stands out as an inflection point in the field of oral history (Gomes, 2020). The themes were the most diverse and derived from public debates, such as: cultural heritage, women’s history, the right to the city, media, migration, art, community, social movement, rural conflicts, archives, nature, transsexuality, soccer, health and science, as well as anti-racist education and historiography.
• 2
  Among the approved papers, some focused directly on digital oral history, such as Jillian Lucy de Fresnes: From wax cylinders to digital devices: the use of technology in oral history and education, and Lucas Maceno: Encounters and disagreements: a comparative analysis of oral history between Brazil and the United States in the digital age.

Publication Dates

• Publication in this collection
  19 Sept 2025
• Date of issue
  2025

History

• Received
  09 Dec 2024
• Accepted
  20 July 2025