Has the concept of censorship gone astray? How to operationalize muddy waters? Half a century of censorship in Portugal

ABSTRACT Censorship studies are currently facing an important conceptual change towards enlarging the phenomena deemed to be censorial (class, race, gender), while risking losing sight of the conceptual specificity of what censorship means. If everything is censorship, then the concept loses explanatory power, but if we restrict the concept it fails to account for the multiple forms censorship takes to be powerful. By mapping censorship studies about the Portuguese dictatorship (1926–1974), we propose an overreaching framework to study the censorship phenomenon in general that articulates its institutional-regulatory and socio-structural dimensions. This articulation enables us to understand the dynamics, productivity and mutations of the regulatory dimension, without losing the conceptual specificity of censorship by assuming that only when this dimension manifests itself can structural constraints be considered part of the censorship process. This framework allows to identify which scientific areas, their preferential objects and methodologies, facilitate an articulated approach to censorship.

nature is built, we choose the Portuguese dictatorship Estado Novo (1933Novo ( -1974) ) as a case study, since this is an example of a long-lasting censorial and Imperial state, as we shall see below.
Censorship became difficult to define.What was simply termed "censorship" is today best defined as "regulatory censorship:" the one decreed by legislation, based on formal rules, stemming from institutions with a religious or secular nature, and executed by official agents.By making explicit the existence of structural forms of exclusion and constraints exerted on freedom of expression, whose existence goes beyond repressive political systems, the New Censorship Theory (NCT), strongly influenced by Pierre Bourdieu and Michel Foucault, redefined the concept by placing it in opposition to regulatory censorship (Burt 1994;Holquist 1994;Post 1998)."Constitutive" censorship, understood as omnipresent, since it is constitutive of the speech act itself, highlights the structural role certain constraints have in the formulation of discourse.
Even if often criticised, both for introducing relativism (equating different degrees of violence under the same signifier) and for proposing a concept whose operationality is problematic (Darnton 2014;Müller 2004), the NCT has, nevertheless, contributed to breaking the dichotomous thinking that limited the understanding of censorship.The dichotomic approach is anchored in the liberal tradition, which understands censorship as a self-evident concept, considered synonymous with prohibition, persecution and intended to foster fear and silencing, that is, a gesture that impedes freedom of expression.NCT questioned the concept of censorship, revealing its multiple, productive, and tentacular nature.Understanding the criticism that can be levelled against this trend does not mean considering this new conceptualization should be completely rejected.Yet, we do not believe the constitutive dimension should replace, and thus relativize and make inoperative, what is meant by censorship.That is why we will try to operationalize the contributions of NCT by focusing on a dictatorial regime.
When choosing a dictatorial regime as a case study, we know that the regulatory dimension will be the one prominently addressed, given it has a more expressive role.We choose this path because it is our understanding that a conceptual operationalization, without rejecting the contributions of the NCT, needs a clear articulation between the regulatory and constitutive dimensions.It seems more useful to think the regulatory dimension can become a richer analytical phenomenon if it incorporates the contributions that understand censorship as structural, that is, constitutive of social reality itself.Observing censorship as a constitutive phenomenon allows us to understand it as something distributed throughout society, counteracting a tendency to frame it as an exclusive monopoly of (repressive) institutions.Yet, it should be added that even the regulatory dimension, read in the light of its constitutive dimension, can become a denser object of study, as we can understand aspects of its power that would not be possible otherwise.We are anchored, therefore, in the NCT's breaking gesture against dichotomous thinking about censorship to give greater analytical and empirical breadth to the regulatory meaning of the concept.
Our aim is to map the studies on censorship imposed by the military dictatorship in 1926 and enhanced by the Estado Novo , as a case study to operationalize the articulation of censorship's regulatory and constitutive dimensions.The Portuguese case is useful because: 1) the recent nature of the field of censorship studies about Estado Novo allows us to map the most relevant studies, which provides an account for how a particular field has dealt empirically and theoretically with censorship's constitutive and regulatory dimensions; 2) Its long duration and imperial dimension makes it a case study subjected to various media transformations and international contexts (Spanish Civil War, WWII, Cold War, etc.); 3) Finally, the relative absence of this case-study in the international field of censorship studies undermines the possibility of cross-national comparisons.

Our framework
How can we operationalize an articulation between the two censorship dimensions?Building on the contributions of NCT (Bunn 2015), and considering the objections raised by scholars who favour the study of regulatory censorship, we propose an original way of articulating both dimensions: the relationship between constitutive and regulatory censorship dimensions can be seen as if they were the simultaneous "substrate" for each other.So, the constitutive power of society (e.g.: moral and social norms) creates the conditions from which regulatory power (e.g.: laws and institutions) can emerge.This means that there is an "articulation" between them: the regulatory dimension "leans on" existing relations of power constitutive of society, which allows regulatory censorship to "model," with the necessary adaptations, its "ideal society," through its own institutional rules; the constitutive dimension, for its part, is a negotiated "integration" of the institutional rules in "naturalized" social relations of power over time.
How to articulate both dimensions in research practice?
(1) Expanding the range of agents to be investigated as contributing to censorship processes.For instance, a film producer may adapt the product according to market demand, which is an economic force that should be articulated with the regulatory power exerted upon the same product.If the constitutive dimension distributes censorship power differently, it is then useful to expand the analysis of agents related to it, thus going beyond regulatory censorship institutions and their agents.
(2) The social fields, to use a Bourdieusian concept, where the censorship phenomenon can be investigated, are not only externally regulated (e.g.: literary state censorship or the Index Librorum Prohibitorum), but also self-regulating (e.g.: the literary field has its own constitutive rules).By knowing their structure and internal rules, we can understand how regulatory censorship becomes dynamic in its relationship with the constitutive dimensions, identifying "new" sites where censorship should be investigated.
This articulation results in a broader concept, which allows for a more thorough analysis of censorship processes.Regulatory censorship becomes productive because it is composed of a series of exclusionary mechanisms that are simultaneously in a permanent negotiation with the social norm of a given society.Thus, this articulation operationalises the analysis of constitutive censorship, identifying, through the regulatory trail, other types of structural powers/interdictions that influence regulatory power.It also allows to explain the mutable character of censorship: the regulatory dimension must respond to existing power relations in society (gendered, economic, racial, etc.) when it tries to translate these social power relations efficiently into explicit regulatory apparatuses; or when regulatory censorship must revisit its own rules because the evolution of social power relations calls into question the effectiveness of those rules.This framework helps to explain the durability of censorship, considering its mutable and adaptive behaviour in relation to the powers already at play in society.Society is constituted by changing power relations, and if censorship wants to be successful, it must learn how to dialogue with society's powers to silence it effectively.By using this framework to read studies that have as a common object censorship applied during the Estado Novo, we seek to map: 1) how censorship has been conceptualised, operationalised and applied; 2) whether both dimensions of censorship have been articulated, or not, 3) how this articulation has been operationalised, or not, (and suggest some hypotheses why this happens).

Theoretical and methodological criteria for corpus selection
Our goal is to map the studies published both in Portugal, and abroad, in several formats: books, articles, conference proceedings, dissertations, or theses. 1 The following items were included: i) publications regarding censorship exercised by the Estado Novo and/ or military dictatorship, published between 1926 and 2022.ii) publications in Portuguese, Castilian, Catalan, Italian, French and in English, languages which are mastered by the authors of the article.From this corpus were excluded: 1) texts published in Portugal before 1974, as they were produced under censorship, except two texts (Arons de Carvalho 1973; Arons de Carvalho and Monteiro Cardoso 1971), because they are frequently "recovered" in studies conducted after 1974; 2) texts published in non-academic outlets.
This corpus was selected by searching the terms "censura," "censorship" and "cens/," coupled with the words "Estado Novo" (New State), to circumscribe the dictatorial regime, and the term "Portugal" to isolate it from the Brazilian dictatorship (1937)(1938)(1939)(1940)(1941)(1942)(1943)(1944)(1945) with the same name.The search was made on the following databases: Porbase (National Bibliographic Database), RCAAP (Portuguese Scientific Open Access Repository), Web of Science, 2 Google Scholar, Google Books, Renates (the Portuguese database of officially registered PhD theses), B-On (the Portuguese academic search engine), Portuguese National Library database and BOCC (Portuguese Online Communication Sciences Library), resulting in 2896 items.
Repeated references and those that did not fit the defined criteria were discarded, namely 1) publications that use the word censorship, but do not analyze it; 2) whistleblowing texts, namely some articles published in the Index of Censorship.Finally, we included publications that, albeit not focused on Estado Novo, have as their subject the lasting effects of 48 years of censorship (Bragança de Miranda 1985;Paquete de Oliveira 1988;Rodrigues 1985).The definitive list is composed of 443 3 items.

Analytical criteria
The material was analysed through categories resulting from the operationalisation of our framework: (1) Type of research: The first criterion is the academic literary "genre" of the publication, which is divided in empirical (uses archives, interviews or other types of corpora), theoretical (elaborates arguments and establishes a debate about censorship as a concept, without associated empirical work) or a mix of the two.Theoretical approaches signal attempts to problematize the concept, not taking for granted its self-evident nature.
(2) Type of approach: The second criterion is the type of censorship researched.Since it is suggested that censorship may operate beyond its repressive and regulatory (i.e.institutional) dimensions, usually associated with dictatorial regimes, it can also be observed as a phenomenon not always controlled by institutions, but supported by social norms, moral imperatives, economic constraints, intersectional power relations (gender, class, race, disability).As such, Censorship would be understood as a phenomenon with productive effects, being thus constitutive, for example, of discourses, practices (rules and rituals) and modes of subjectivation (rules subjects impose on themselves, consciously and/or unconsciously).Bering this distinction in mind, are the studies focused on the regulatory or constitutive dimension, or they articulate both?(3) Implicit theory: We want to understand whether the research formulates an "implicit theory" in interpreting the empirical material, which allows us to include original conceptualizations about censorship, even if they do not engage with previous studies or with explicit conceptual work.This criterion allows us to include original theoretical proposals on censorship even when there is no reference to the New Censorship Theory or an explicit conceptual and theoretical work.(4) Archive: We want to understand if studies use archival material (including newspapers, but not the consultation of laws) and/or resort to testimonies (i.e. the use of interviews, life histories and questionnaires, constituting thus an "extended archive" of censorship).The goal is to understand if (and how) the constitution, organisation and availability of archives relate to the type of studies produced.(5) Funding: What studies are funded and by which institutions.This category includes master's and doctoral scholarships, individual projects (researcher contracts) and collective research projects.(6) Object: What objects are privileged in the study of censorship?Through an abductive approach of going back and forth between our theoretical framework and our empirical material, we arrived at the following variables to code the objects of study, which can be cumulative: agents (individual actors); cinema; colonial cinema; (book) publishing; censorship as a concept (theoretical approaches); censorship institutions; journalism (professional class); film subtitling; literature; children's literature; music; periodicals (newspapers and magazines); colonial periodicals; women's periodicals; radio; colonial radio; theatre; "popular theatre;" television; translation; other.(7) Author's career path: We will give an account of authors' academic career to know how different disciplines approach censorship.The authors' background is characterised in two ways: disciplinary area(s), which may include more than one discipline, and occupation: tenured professor; researcher; student; non-academic.This allows to understand the labor conditions and sustainability of the field.
(8) Relevant bibliography: Through the bibliography cited, we will try to understand if and how the study interacts with the field, both nationally and internationally.

Number of publications, funding, careers and archive
Studies on Estado Novo's censorship have been increasing since 2006 (Figure 1), which can be associated to the eased access to archives and public funding policies, since close to a third of studies (29%) have been funded by the Portuguese national agency (FCT) (Figure 2), created in 1997.Funding from other entities is scarce (5%), yet the Brazilian agency CAPES stands out (2%).Despite this, 66% of the studies are not associated with direct funding.Almost half of the studies (49%) are performed in precarious working conditions, with one third of them being conducted by students (masters and doctoral students) 4 and 6% by non-academics, where the journalist profile is high.The authors' professional status changed in the first decade of the 2000s (Figure 3).It ceased to be a field dominated by non-academics to become a field highly represented by professors, students and researchers funded by the FCT, either through individual funding or collective projects. 5 Despite the "academicization" of the field, the organization and access to the archive also influenced both the studies' nature and quantity, which increased considerably from the mid-2000s (Figure 4).
The studies growth may be related to the accessibility of the archive, which starts to be possible to consult in the second half of the 1990s.Yet, whilst no longer access restrictions apply, there are two moments (2005-2011 and 2015-2018) (Figure 4) when studies without use of the archive are dominant, suggesting that, besides access, there might be other issues at stake, such as its size and organisation.
This archive is composed by materials from the Estado Novo's Council Presidency, the Ministry of the Interior and other structures of the regime, for instance the political police (PVDE/PIDE/DGS) or the single party (União Nacional).In 1977, these were entrusted to the Black Book Commission on the Fascist Regime, created by proposal of the then PM,  Mário Soares.This commission, which is not recognized as a truth commission (Morais 2022), although it aimed to collect, organize and publish information on crimes and legal abuses committed during the dictatorship (25 volumes were published), was dissolved in 1991 by the government of António Cavaco Silva (Portuguese prime minister from 1985until 1995).The António Salazar Archive (AOS), which had been deposited in the National Library (BNP) in 1981, coming from the Palace of São Bento, where it was stored after the dictator's death in 1970, was transferred, following the dissolution of this commission, to the National Archive (ANTT) in 1992.However, until at least 1995 it was difficult for researchers to access this archive.Only by their pressure, having met the 25-year deadline after Salazar's death (on 25 July 1995) and taking advantage of a change of government that year, was it possible from then on to investigate this period by freely consulting the documentation deposited at the ANTT (Loff 2014, 88-90).Marcelo Caetano's Archive was only deposited, by his descendants, in 1999 and is still, according to legislation, conditioned.The SNI funds (National Information Secretariat) were gradually incorporated afterwards: in 1997 was transferred the material that had been deposited in the BNP since 1975, this material was collected by order of Oliveira Marques (director of the BNP between October 1974 and April 1976) after the invasion of the Censorship Services following the overthrow of the regime.In the year 2000, it was incorporated what was stored in the Pendão warehouse.The funds of the Directorate-General for Public Performances (DGE), which were more dispersed, were incorporated between 1997 and 2005 (Tremoceiro 2014, 5).The archive has quite a few gaps and is relatively disorganized, due to the way it was constituted, having also suffered a purge of documentation produced before 1955 in 1970 (Cardoso Gomes 2014, 18), and an interruption of guardianship in 1974, which makes it a somewhat challenging archive.In sum, the type of archive available limits what can be read and how it is read.
Prior to this date, studies resorted to privileged access to certain archives, such as the Jornal de Notícias archive (Forte 2000;Príncipe 1979), or public archives in a period of transition, such as the archive of the DGE (António 1978).Between the second half of the 1990s and 2006, when funding was rare, a set of works are published that are indebted to easier access to this archive (Leitão de Barros 2005, Santos 2005;Cristo 2005;George 2000;Ramos do Ó 1999;Ramos do Ó 1996) and taking advantage of the publishing boom (Azevedo 1997(Azevedo , 1999;;Castrim 1996) triggered by the twentieth anniversary of April 25 (Loff 2014, 114-117).
The scant use of testimonies, either as the sole methodology (7%) or with archival work (8%), shows that the notion of the censorship archive is still very much related to the use of documental sources.Using testimonies, which are not so dependent on archival policies, could stimulate the integration of constitutive dimensions of censorship.This is a pressing matter, in view of the ageing of the subjects who lived through the dictatorial period.

Type of investigation and disciplinary areas
Approaches to censorship are mostly empirical (71%), with less emphasis on combining empirical work with theoretical reflection (19%), or on specifically theoretical approaches (10%).From our inquiry, three types of disciplinary path resulted dominant: literary studies (22%), history (20%) and a multidisciplinary profile (22%).Other relevant areas are communication (11%) translation (6%) and, finally, sociology (4%) or professional journalists (4%).Considering the four more relevant paths (corresponding to 75% of the total number of studies): literary studies, history, multidisciplinary profiles and communication, the preponderance of purely empirical studies remains (71%).In less significant fields from the statistical point of view, theoretical work, associated or not with empirical approaches, is equivalent to empirical work, as in the case of translation studies (3%) and sociology (2%).Only in the case of law scholars, theoretical reflection (1%) overcomes an empirical approach (0.4%).Finally, the reconversion of ideas related to other dimensions of the Estado Novo is rare, but we should highlight the re-appropriation of José Gil's concept of invisibility (Gil 1995), or Hermínio Martins's idea that the Estado Novo employed an "economy of violence" (Martins 1998).
Within the four aforementioned relevant paths, the combined theoretical-empirical approach is only significant in the case of literary studies (29%) (Table 1).A possible explanation is that the field is more familiar with the type of research and literature that supports the notion of the constitutive censorship (i.e.literary field structure and membership, processes of self-censorship), so much so that the NCT emerged within it (Burt 1994;Holquist 1994).Another value that stands out is the theoretical approach of the multidisciplinary profile (19%), which suggests that the encounter of disciplines favours conceptualization work, since different fields of research approach the objects and workings of power in different ways.

Type of approach
Most studies address regulatory censorship (55%), focusing residually on the constitutive dimension (5%).Yet, the intention of combination of both dimensions has a considerable expression (40%).An expression that signals the need and/or relevance of investigating censorship framed in its social and economic context, although explicit forms of articulation, as the one we propose, were not identified.Bearing in mind that 78% of the studies are performed by authors inscribed in a singular disciplinary domain, a greater multidisciplinarity might facilitate an increase of an explicit articulation.This could be stimulated through a conceptual approach to areas such as sociology, political philosophy, gender studies or communication studies, where different types of power are typically explored.
On another note, the integration of both dimensions stood out in the 1970s and 1980s (Figure 5).It should be underlined that these authors lived through the censorship exerted by the Estado Novo and did not reduce their analysis to its regulatory dimension, something that started to happen considerably from 2006 onwards, coinciding once again with public investment and access to the archive.Considering, again, the four major paths, it is in literary studies that the combination of dimensions stands out: 55%.As suggested earlier, this field might be more open to address censorship constitutive dimension.If taken as a research area, 45% of the studies conducted by authors with a multidisciplinary profile seek to combine both dimensions, suggesting the hypothesis that the ability to look at censorship through different disciplinary perspectives increases the possibility of addressing the constitutive dimensions of the phenomenon.On the contrary, both in history and communication, the regulatory dimension is the one that prevails as an object of study, 60% and 76% respectively, followed by studies performed by journalists (70%).There are fields dominated by the combination of both dimensions (information sciences − 100%, or sociology − 60%), but they represent very few studies.

Relevant bibliography
Most works have relevant bibliography (73%) referring to national or international studies on censorship.This use is more expressive in areas such as sociology, translation studies, literary studies, communication and history (Table 2).On the contrary, studies carried out by journalists tend not to resort to relevant bibliography (90%) Using relevant bibliography is related, moreover, to the type of approach to censorship: 63% of the studies addressing the regulatory dimension refer the state-of-the art.This percentage increases with studies that address censorship constitutive dimension (79%).The more a study distances itself from the notion of censorship as something selfevident, the greater the possibility of doing so from the state-of-the art.A significant part of the bibliography used gives access to legislative compilations or primary sources, thus serving as an agent to mitigate the above mentioned archival difficulties.

Implicit theory
Most studies contain an implicit theory about censorship (63%), formulating a theory when interpreting the empirical material, without necessarily explaining and/or referring to key authors in censorship.This way of working was dominant until the second half of the 1990s, possibly given the difficult access to the archive, a time when a type of empirical work began to increase without a theoretical interpretation, something that becomes more evident from 2006 on (Figure 6).This more descriptive work may be related with the growing archival work and the need to map it, handling large amounts of uncatalogued documents, which may limit the readiness to interpret them.It's noteworthy that implicit theory is more prevalent compared to descriptive approaches.

Archive
A censorship archive is always challenging, and in the Portuguese case the challenge increases given the degree of dispersion and disorganisation.Although these studies, as expected, use this archive (49%), and even combine it with other ways of obtaining information (8%), more than half do not (51%).Seven percent of these studies resort exclusively to other non-documental ways of obtaining information (interviews, life stories, etc.).The latter includes testimonies of prominent actors (cf.Azevedo 1997Azevedo , 1999)), or forms reception of censorship in everyday life (Carvalheiro and Silveirinha 2015).The questionnaire is rarely used (Paquete de Oliveira 1988), though it has been used in similar cases (Abellán 1980) to obtain information about individuals with the same profession (writers, journalists).Diversifying the sources used may mean an attempt to enrich the readings of the regulatory dimension of censorship.As 40% of the studies attempt to combine the two dimensions and, in doing so, this may be reflected in a search for other sources or a different use of those available.
The non-use of the archive can be related to its unavailability and its incommensurability, something that can be mitigated by resorting to bibliography that provides primary sources (António 1978;Arons de Carvalho 1973;Franco 1993;Príncipe 1979) as well as the volumes published by the Black Book Commission on the Fascist Regime (i.e.Comissao 1980,1981).This type of bibliography is cited until recently, which might reveal a sustained scholarly dialogue, facilitated by book format, but might also be a sign of the difficulties in accessing the archives (or organized legislative information, made available by these studies).Some books are repeatedly cited because they grant access to a once unavailable archive (António 1978;Forte 2000;Príncipe 1979).Incidentally, recent studies (Leitão de Barros 2022; Castanheira 2009; Gonçalves 2012) continue to benefit from privileged access to private archives (Newspaper's own Archives or the project Ephemera Archive), perhaps demonstrating that the material preserved by the public archive may fall short of research needs.

Object
Following the tendency of the historiography on repression during the Estado Novo, the metropolitan dimension has preponderance over the colonial one, with only 37 studies.It was equally expected that the study of censorship privileges writing, since the selected materials, in part, reflect what is available in the archives and translate the easiness of access to these sources (Table 3).
Concerning the audio-visual archive, except for cinema, is much scarcer in terms of documentation and more demanding in terms of conservation and consultation, namely requiring reproduction equipment.In fact, even with audio-visual objects, a substantial amount of research is based on written documents (cinema resorts to the Censorship Commission's files, including the commission's minutes; radio and television have less paper trail, and music can be researched through the censorship of lyrics).The focus on written documents also biases the study towards the regulatory dimension, although this is not an insurmountable limit if we take the example of literary studies.The reduced concern with the idea of censorship (7%) is congruent with the reduced presence of theoretical studies (10%).Finally, only 10% of the studies address more than one medium, which limits the understanding of the cross-media action of censorship.

From combination to articulation
The framework used allowed us to identify a significant number of studies (40%) that have sought, over time, to combine the regulatory dimension of censorship with constitutive ones.The expression "combine" is essential, given that few studies effectively "articulate" its various dimensions.We performed a "solidary" reading to identify whenever the author approached censorship, empirically or theoretically, beyond its regulatory dimension, which can explain why we got a high number of combinatory approaches.Albeit with varying degrees of conceptualisation, this intuition has given rise to nomenclatures that try to capture this combination like "hidden" (Paquete de Oliveira 1988), "parallel" (Azevedo 1999;Cardoso Pires 1977), "panoptic" (Melo and Costa Dias 2022), "informal" (Areal 2013), "internal censorship" (Rosa 2009;Ventura 2014), "economic" (Cabrera 2006;Cavaco 2012;Melo 2016;Rebello 1977;Vargues 2007), class (Vilhena and Gomes Ferreira 2014) or gender-based censorship (Carvalheiro and Silveirinha 2015), among other options.Yet, even these studies remain, mostly, without formalizing an operative articulation.Within a field dominated by the need to account for the regulatory dimension, there is a concern to integrate it into a long-term historical perspective, emphasizing a cumulative regulatory experience (i.e.Garcia 2009;Brangança de Miranda 1985;Rodrigues 1980;Vargues 2007), as well as, in the 1980s, an attempt to dissociate dictatorial regime and censorship, underlining the elements that spill-over to democratic regimes (i.e.Paquete de Oliveira 1988, 149-162).Some attempts have been made to introduce economy and gender as constitutive dimensions in the study of censorship.These elements also correspond to two of the pillars of the regime, namely class and gender domination (Brasão 1999(Brasão , 2012)), which hinder social mobility and promote female subjugation.Despite some relance given to economic structural constraints (Cavaco 2012;Gonçalves 2012;Rebello 1977), it is gender, nonetheless, he dimension without which censorship cannot be studied, specially within literary studies, a field dominated by authors with a name typically female (74%).Within this field, authors, especially female (if, problematically, we assume a name corresponds to a feminine or masculine gender) that study censorship from a gender perspective, resort more frequently to its constitutive dimension (e.g.: focusing on women journalists and writers such as Natália Correia, Maria Archer, Maria Lamas or the court case of Novas Cartas Portuguesas).If we consider the entire corpus, female names are more associated with approaches combining the regulatory and constitutive dimensions of censorship (22% compared to 16% of male names).This data shows the importance of gender diversity in academia, which directly translates into the type of knowledge produced (Harding 1986, 90-91), something that could be extended to other forms of diversity, namely racial, clearly underrepresented given the scarcity of studies devoted to the colonial dimension.

Conclusions
The problematic of the censorship exercised during the Estado Novo was not articulated around a transitional justice programme, but mainly as an academic field, whose growth was aided by public policies of investment in research and constitution of archives.The political role of the Black Book Commission on the Fascist Regime, responsible for the availability of material that is still frequently used currently, should not be forgotten.The reverse side of this role was a limitation on direct access to the archive whilst it was under its jurisdiction.A restriction that remained even after its extinction.The archive's availability was central, especially in articulation with funding policies, to the shift in the field: more studies, predominance of the empirical perspective, greater emphasis on the strictly regulatory dimension.This shift affected the type of researcher as well: the scholar replaced a non-academic profile.Studies conducted precariously (by students and non-tenured researchers) have a relevant role, thus impairing consolidation of lines of research.
Literary studies have a relevant role in defining the field, as well as history, communication, and a multidisciplinary profile.A clear preference is noted for textual content, in the form of multiple objects (mainly books and periodicals), in detriment of audio-visual objects or crossmedia perspectives.Also, the focus on certain materials, linked to artistic endeavors (cinema, art, literature), limits an approximation to censorship in the everyday life of anonymous subjects, where the constitutive dimension of censorship operates.Using our framework, we observe a tendency for studies to seek a combination between regulatory censorship and other dimensions, even creating original concepts to designate them ("internal," "social" or "economic" censorship).However, this articulation is not yet operationalised.
By resorting to our framework, studies seeking to broaden the concept of censorship, without losing sight of the articulation between its regulatory and constitutive dimensions, may include agents that participate in the ramification of power by the social body within different social fields, going beyond a strict focus on agents and official institutions of censorship.This broadening allows to understand how regulatory censorship informs social norms, and how these negotiate what is prescribed by censorship.It means that a field of studies too focused on the regulatory dimension of power, and rarely relating empirical work with theory, may not be studying censorship in all its richness.Studies that adopt our framework may add to this regulatory focus an articulation with agents and social fields performing censorship's constitutive dimension.This articulation has not yet been operationalized, perhaps precisely because the pressing need to map the archive, a task covered by roughly half of the studies, does not stimulate the exploration of other dimensions of censorship.
It also appears that the disciplinary path informing most of the studies favours a regulatory approach to censorship, especially in history and communication.We put forward two speculative hypotheses for this. 1) The field of communication, which studies Estado Novo's censorship, is the recent field of communication history that, on Portugal, performs relevant work, though often more descriptive.2) History has an intrinsic relationship with the archive, where the regulatory files are stored.In contrast, both a multidisciplinary profile and the field of literary studies seem to be more permeable to work on non-regulatory dimensions of censorship.In the first case, the intersection of multiple disciplines fosters several perspectives on power, which densifies the censorious phenomenon.In the second case, it is a field that traditionally incorporates the constitutive dimension of censorship and is not so conditioned by the archive, being also the field where the perspective proposed by NCT was outlined.

Notes
1. Undergraduate monographs were excluded as they do not need to present original contributions.2. Scopus was not included since Web of Science covers the publications of the former, avoiding thus duplication of references (Singh et al. 2021, 5137).
3. 30 items were not available.However, these absences do not produce a significant bias in our results, as they are a minority, and some of its authors are already represented.4. A total of 104 doctoral and master's theses.The dominant areas are history (16 MA and 10 PhDs),communication (15 MA and 7 PhDs), literary studies (9 MA and 12 PhDs), sociology (6 PhDs) and translation (5 MAs). 5. 8 collective projects between 2002 and 2020: "Living Memories of Journalism" ( 2002

Disclosure statement
No potential conflict of interest was reported by the author(s).

Figure 1 .
Figure 1.Number of publications by year.

Figure 2 .
Figure 2. Publications with FCT funding by year.

Figure 3 .
Figure 3. Academic path by year.

Figure 4 .
Figure 4. Number of publications and archival work by year.

Figure 5 .
Figure 5. Regulatory and combinatory approaches to censorship by year.

Table 1 .
Type of investigation inside each dominant area.

Table 2 .
Area's use of relevant bibliography.

Table 3 .
Selection of the most studied objects of censorship.