Epistemic work in Portuguese parliamentary education debates: Externalisation to world situations as a source of epistemic capital

This article seeks to analyse how epistemic work (Alasuutari, 2018; Alasuutari and Qadir, 2019) in Portuguese parliamentary education debates develops through externalisation to world situations (Schriewer, 1990) whose references are used as epistemic capital (Alasuutari, 2018). The study explores debates occurring during Legislature X (2005–2009), in which the number of external references in plenary education debates significantly peaked compared with previous and subsequent legislatures. The analysis demonstrates that there was a change in the (de-)legitimation strategies policymakers used during this legislature. In addition to using the traditional ideological principles associated with each party, deputies often opted to use international organisations and their instruments, alongside reference societies, as authoritative sources to strengthen arguments and persuade their audience of their validity. This suggests that parliamentary policymakers assume that these external elements are useful in earning support for their policy ideas and promoting reform proposals.


Introduction
The growth of international organisations (IOs) and their focus on education's performance and improvement makes it difficult to sustain the view that education policymaking is solely a nation state issue. The existence of a 'global education policy field' (Rizvi and Lingard, 2010) must be acknowledged. Nevertheless, despite the recognition of global-national intertwinements in education policymaking, the policy process remains highly localised. Although IOs and practices and interesting findings. The lack of interest in PISA results may be related to the fact that Portugal accepted the limitations of its education system more than 150 years ago, when comparative statistics among European countries began to systematically reveal the country's continuing low literacy levels, poor school infrastructure and insufficient education funding (e.g. Antunes, 2004;Gomes, 1999;Mendes, 2015;Nóvoa, 2005). The poor results in PISA and other international large-scale assessments (ILSAs) were therefore expected and simply taken as confirmation of what was already known.
I performed a qualitative content analysis of 115 parliamentary education debates occurring in plenary sessions. I also applied rhetorical analysis specifically to Legislature X, when the quotations identified with external references showed an unusually high peak (from 136 in Legislature IX to 354 in Legislature X). In a previous study (Santos and Kauko, 2020) we identified all the external references used in Portuguese parliamentary education debates between 2001 and 2018, and the tone and functions of their use (see list of the ten most frequent references by legislature in Appendix 2), arriving at the understanding that most external references were used positively to promote legitimacy. In that earlier paper we demonstrate that while international influences certainly occurred, Portuguese parliamentary policymakers strategically instrumentalised external references to feed arguments where problems were framed in certain ways, and policy solutions were aligned with each party's ideological and political agenda. This intensifies the controversial dynamic that characterises the Portuguese political system. Ultimately, the goal is to manage conflict between the different political forces and eventually enable policy reform. This article aims to deepen that early analysis by exploring how processes of externalisation to world situations bring to policymakers' arguments useful resources of epistemic capital used in the epistemic work developed in Portuguese parliamentary education debates.

Epistemic governance framework: analysing the art of affecting the audience's views
The EGF, as described by Alasuutari and Qadir (2019) and adopted in this analysis, focuses on social changes in modern societies. It is described as an additional analytical layer in the study of how power is developed and handled, and it specifically analyses the dynamics of governance and policymaking (Alasuutari and Qadir, 2014, 2016. It defines governance as 'the efforts to bring about change (and maintain status quo) in any given social system' (Alasuutari and Qadir, 2019: 3) from small organisations to the world system. Governance therefore consists of the ability to intentionally, but through conscious, semi-and unconscious actions and arguments, affect others' 'conceptions of reality', thereby producing social change (Alasuutari and Qadir, 2014: 71).
The EGF is strongly influenced by Foucault's concepts of power, order and especially governmentality, as well as his premise that the art of government involves several non-hierarchical strategies of control and power (Alasuutari and Qadir, 2014: 67-68). One such strategy consists of the ability of political actors to influence others' views and behaviour through discourse. As individuals can control their own actions and constructions of reality and consequently steer social change at different levels -especially in democratic nations -power often depends on policymakers' ability to convince others that the situation under debate is problematic and needs to be addressed through specific policy reforms. This rhetorical work is often done using different forms of constructed knowledge and authority (Alasuutari, 2018).
The EGF is a complex framework that integrates several components, which I shall describe below. The framework focuses on the analysis of the epistemic work developed in one or several specific contexts (e.g. in comparative studies). Epistemic work is described as a set of 'particular techniques used by actors engaged in affecting views and hegemonic definitions of the situation at hand' (Alasuutari and Qadir, 2019: 22). To fully understand how epistemic work develops, one needs to identify which sources of authority -also called epistemic capital -are used and which imageries they help to build. The broader idea is that epistemic work strategically develops simultaneously around three objects: environmental ontology, actor identifications, and norms and ideals. 'These are constructions of what the world is, who we are and what is good or desirable' (Alasuutari and Qadir, 2019: 21). The epistemic work on these three objects simultaneously occurs in two dimensions: paradigmatic, constituting broad ontological assumptions of reality; and practical, through the production of knowledge concerning specific paradigmatic assumptions, for example (Alasuutari andQadir, 2014, 2019; see Figure 1).
According to the EGF, imageries play a fundamental role in epistemic work. They consist of 'pictures' through which individuals and institutions see the world. Constructed by the 'combination of root metaphors', 2 they are strategically used as illustrations by policymakers seeking to affect people's views and behaviour (Alasuutari andQadir, 2016, 2019: 37-57). The EGF identifies three different imagery narratives used by policymakers and other political and social actors: functional requirements of modernisation -new policies are presented as required to maintain or improve the country's or region's development; society as hierarchy -a view of a hierarchically organised society with the individuals or institutions at the top retaining more power, making the decisions and having the responsibility; and society as competing blocs -a conceptualisation of society as divided into units with their own interests that compete with each other (Alasuutari andQadir, 2016, 2019: 37-53).
To construct these imageries, the EGF highlights that political and other social actors need different sources of authority that work as epistemic capital (Alasuutari, 2018): a cumulative collection of all kinds of resources that can be used as knowledge and validation bases for the arguments presented (Alasuutari, 2018; Alasuutari and Qadir, 2019: 50-61). It is important to note that authority is always relational; it depends on others' recognition of them as such . Alasuutari and Qadir (2019: 59-77) identify four different forms of authority: capacity-based, referring to people, organisations or objects seen as respectful within their social system who are recognised as able to do what they promise legally or illegally (e.g. the idea that IOs help to change reality); ontological, referring to a person, text or institution used to 'accurately' present the world and the debated situation (e.g. research reports, IOs, religious texts); moral, alluding to wellaccepted principles, rules and standards, and the people who represent them (e.g. laws, religious books, judges or clergy); and charismatic, referring to a person or institution because of their talents, acts, blood or rights of office. They are constructed as sacred through narratives, public emotions and rituals, and they can influence or be used to influence others, even outside their expertise and abilities (e.g. artists' and social movements' involvement in policymaking).
In accordance with the EGF all the above aspects of epistemic work are cumulative . Different imageries and sources of authority (epistemic capital) are often combined in arguments that seek to affect people's views and emotions, consequently originating (or limiting) change.
Because of its focus on how discourse affects others, the EGF is a suitable theoretical and methodological framework for analysing plenary parliamentary debates. 3 The parliament's work is strongly epistemic in nature, because the deputies intend to affect other deputies' views and decisions -as well as the broader national audience -using epistemic capital in the construction of imageries that illustrate arguments (de-)legitimising the debated reforms. In the construction of such arguments, external references are often invoked as epistemic capital. My analysis focuses on which external references are most frequently used and their role as epistemic capital in the construction of the imageries that policymakers use when developing their epistemic work in Portuguese parliamentary education debates. The suitability of the EGF for the study of externalisation to world situations is related to both frameworks' focus on the analysis of the reception and the translation of global ideas and practices in local contexts.

Data and methods
The study starts with the overall data of 115 parliamentary education debates happening in the Portuguese parliament's plenary between December 2001 (when PISA's first results were announced) and December 2018 (when the data was collected). It includes debates from Legislatures IX to XIII, 4 with all the available 81 general or specific education debates and 34 debates on education-related bills (first reading and discussion). While this data was explored in a previous article (Santos and Kauko, 2020), I use this large dataset in this article as background/contextual information and focus only on the 41 debates of Legislature X between March 2005 and October 2009 (all the available 31 general or specific debates on education and 10 debates on education-related bills). This narrowing of focus is related to the identification of an evident change in the discourse's strategy during this period, with an important accentuation of the use of external references, from an average of 46 quotations with external references per year during Legislature IX to 89 during Legislature X, returning to fewer external references in subsequent legislatures. The data was collected from the online Diários da Assembleia da República (DAR) 5 Kauko, 2020).
The analysis of the debates developed on two different levels. First, I developed a qualitative content analysis (Schreier, 2014) with the support of the Atlas.ti8 software for the entire dataset (115 debates, 2001-2018), developing an inductive and deductive coding system. Eight categories 6 were created beforehand, based on the research question and the basic information that would identify references, their tones and functions, speakers, topics discussed and so on. These categories were further complemented by the sub-categories emerging from the reading of the debates. The method was helpful for familiarisation with the data, facilitating the reduction of the amount of data to the essential elements required for the study, and giving a general idea of who the speakers were, who and what were used as external reference targets, the tone (positive, negative, neutral) and function (legitimation or de-legitimation) of the references (e.g. references to other countries: Spain; speakers: PS deputies). To better understand the political context and the government agenda, the same content analysis was performed for the XVII government programme's sections on education.
A second level of analysis, rhetorical analysis (Edwards et al., 2004;Leach, 2011), was performed for the debates of Legislature X, which are this article's analytical focus. The decision to apply rhetorical analysis in addition to content analysis was connected with the use of the EGF, because rhetorical analysis promotes an understanding of the strategies used to influence or persuade others through discourse, text or images (Edwards et al., 2004;Leach, 2011). The rhetorical analysis focused on the allusions to world situations as a rhetorical device used to strengthen the speaker's credibility (ethos) and to appeal to the audience's rational thinking, logic (logos) and emotion (pathos) (Leach, 2011). This analysis enabled an understanding of the parliamentary policymakers' assumptions concerning their audience's a) perception of reality, b) identification of the entities with authority and c) perception of what was positive or negative, acceptable or unacceptable; and ultimately, how the speakers considered that specific external references could contribute persuasively to constructing the debated issue in ways that suited their party's agenda.

The Portuguese political system and externalisation to world situations
To better understand the parliamentary debates, it is helpful to understand their context. Portugal has a semi-presidential political system. Power is divided between the president, parliament and government. Parliament is unicameral, has strong legislative powers and holds the government to account (e.g. Freire, 2005;Goes and Leston-Bandeira, 2019). The 230 deputies are selected by each party internally and elected by the d'Hondt method, forming party parliamentary groups (e.g. Leston-Bandeira, 2004;Lobo et al., 2015). During the legislature under scrutiny six parties were represented in parliament: three left-wing parties, one right-wing party, one centre-left (also government) party, one centre-right party and one independent deputy.
In examining externalisation during the five legislatures, an accentuated opening of the system during Legislature X can be observed ( Figure 2). Figure 2 shows an average of 46 quotations with external references every year during Legislature IX (3 years -2002-2005), 35 during Legislature XI (2 years -2009-2011), 39 during Legislature XII (4 years -2011-2015), and 38 during Legislature XIII (3 years -2015(3 years - -December 2018. During Legislature X (4 years -2005-2009) an average of 89 quotations with references to world situations was found in each year.

The externalisation process as a source of epistemic capital in the XVII government programme and parliamentary education debates
Legislature X (and simultaneously the XVII Portuguese constitutional government) commenced in a national and international context characterised by an increasing focus on improving educational equality, efficiency and quality, accompanied by significant interest in international articulation, cooperation and competition. In Portugal this race to improve efficiency and quality at all levels of the education system coincided with the growth of assessment mechanisms for students, schools and teachers, along with curriculum changes and reforms in school and higher education management (Barroso, 2009(Barroso, , 2016Carvalho et al., 2019), to name the most relevant. However, these issues are not new. They belong to the modernisation of public systems in western countries that started during the 1960s. Portugal followed slowly, with various incremental and frequently conflicting reforms implemented as governments changed (Barroso, 2006(Barroso, , 2016. What appears to have changed is the type of argument used to discuss these topics. In addition to traditional ideological arguments, during this legislature there was an increase in the use of external references as a tool to sustain rhetorical evidence claims. This externalisation peak is in line with previous research (e.g. Afonso and Costa, 2009;Costa, 2011), suggesting that during the XVII constitutional government (and Legislature X) 'There is a change in the conceptualization of the decision-making process and a renewal of the foundations of political legitimacy, more based on scientific reasoning, and less on ideological logics' (Afonso and Costa, 2009: 53).
When reading the XVII government programme, one notices that it is replete with terms such as 'equality', 'quality', 'efficiency', 'innovation', 'development', 'assessment' and 'qualifications' and their relationship with 'employment', 'quality of life', 'economic development' and 'internationalisation' of 'people' and 'organisations'. The policy plans for education focus on the reorganisation of the school system, its curriculum and the assessment of its different elements at all levels, including higher education. References to the 'European Union' (EU) and its countries (as a broad, regional reference society -'European countries') are the main targets of these externalisations, feeding arguments about the need to fulfil EU agreements or advice and to catch or keep up with other 'European countries'. It is therefore clear that the organisation is seen by the XVII government as a key source of authority. External references are also strongly used in parliament. They are used as sources of authority to build arguments that, although dressed up as based in external evidence, ultimately aim to gather the audience's support for proposals that follow each party's ideology and that have been longstanding agenda items. During Legislature X, 91 external targets were identified. The most frequently used (and further explored in this analysis) were 'European countries' (1st), 'Bologna Process/Declaration' (2nd), 'EU' (3rd), 'OECD' (4th), 'PISA' (joint 5th) and 'several countries' (joint 5th). 7 These main external targets were generally used positively, serving to legitimise speakers' own arguments. However, PISA was mostly used by opposition parties to de-legitimise and criticise government decisions (Figures 3 and 4).
In addition, in the analysed debates external references are mostly used as authorities with an ontological or capacity-based character. They are less often used as a charismatic or moral authority. The same reference often accumulates different kinds of authority.
The next section of the article explores the nuances of the use of the five most frequent external targets used by different policymakers. The aim of this analysis is to identify the similarities and Union (51) OECD (37) PISA (22) Several/other countries (22) LegiƟmisaƟon De-legiƟmisaƟon  differences in how these references were used during Legislature X, who used them and for what purposes (Appendix 1 offers an overview of which party refers to each of the five top targets).

Performing epistemic work in the Portuguese parliament during Legislature X
The following description of how the five main external targets are used as epistemic capital by the different party representatives is important for demonstrating how and why these targets become epistemic capital, and how their use is nuanced according to the speaker and her/his aims. This information helps to identify the different components of the epistemic work, such as the epistemic assumptions made by the speakers, the imageries they build and how these components function together in the arguments presented by the speaker, leading to an understanding of the epistemic work itself.
References to EU countries. The most frequent external reference during Legislature X was 'EU countries'. Of the 356 quotations identified with external references, 97 referred to 'European countries' (27%). As the reading of the government programme has already suggested, most references to 'EU countries' were made by government members (43 quotations, 44%) in legitimising their programme's ideas and bills concerning diverse education issues, from student and teacher assessments to higher education management. References to 'EU countries' were joined by many other references, 'Bologna Process/Declaration' and 'OECD' being the most frequent. For example, a (left-wing) PEV deputy uses OECD information to support an argument against higher education tuition and the inequalities it will generate compared with other European countries: Indeed, these are the cases that prove what the OECD data says -namely, that Portugal is one of the European countries where socioeconomic inequalities are most felt in terms of access to and the frequency of higher education. (PEV deputy, 2008-06-11) 8 This quotation demonstrates that references to 'EU countries' frequently serve as benchmarks for comparison, with an implicit expression of the need to belong to the group of 'successful' EU countries.
References to the Bologna Process. 'Bologna Process/Declaration' was the second most used external reference during Legislature X, present in 92 quotations (26%). This was directly linked to the government's intensive focus on reforming higher education to accommodate the measures defined in the Bologna Declaration. It is therefore unsurprising that the policymakers who referred most frequently to the declaration were government members (30 quotations, 33%). Discussions of these reforms were long and problematic. The centre-right and rightwing parties highlighted the benefits of adopting the declaration, using it as a framework for the country's development, or its adoption as a means to improve the percentages of qualifications, the quality of human capital and therefore people's quality of life and the country's economy. Meanwhile, the left-wing parties argued against most of the reform ideas legitimised by the declaration, negatively framing the declaration and its focus on the marketisation of education among other issues.
Higher education reforms were debated with the use of several external references besides the Bologna Process/Declaration. The OECD and especially the 'European countries' were frequently used as a symbiotic source of authority to feed arguments related to the importance of 'fully' belonging to these organisations, following their guidelines, or the need to catch or keep up with their countries' practices and policies, as the quotation below from a government member underlines: Mrs [deputy X], feeding any illusion in this joke that Bologna is a kind of 'tailoring' for European accreditation that does not serve us -that is, that Europe doesn't serve us . . . I apologise, but this is unacceptable. (Government, 2005-05-12) References to the EU. The EU was the most referenced IO in Portuguese parliamentary education debates. It arose in 51 quotations (14%) of the total of 356 quotations with external references. Government members used this reference most (22 quotations, 43%). Like references to 'EU countries', the government's frequent referencing of the EU accorded with the government programme, in which the EU also featured strongly. The EU was often described and referred to as common ground, a 'society' to which the country proudly belonged, as the speech of a PS deputy (government party) demonstrates: The internationalisation, Mr and Madam Deputies, belongs to the genetic material of the university, Europe and Portugal's past. It will certainly belong to the genetic material of the Europe we are building . . . a cluster of wealth and social rights, to which Portugal today belongs. (PS deputy, 2007-05-09) References to the OECD. During Legislature X the OECD was a frequently used external reference (37 quotations, 10%). Although the OECD (as an institution) was only mentioned in the government programme once (and five times as 'OECD countries'), it is noteworthy that most references to this organisation were made by government members (15 quotations, 41%). It was frequently cited as an expert, a provider of trustworthy data on which policymakers could rely for information in advocating certain policies. For example, in arguing for the need to invest more in vocational and adult education to bring people with low literacy levels back to school, the government legitimised its arguments with the organisation's data: The OECD recently published a study showing that if all countries increased their average level of education by one year, this would have an effect of a 1.2% to 1.7% increase in employability. (Government, 2005-09-21) As the example above effectively illustrates, references to the OECD often concerned the organisation's core values and aims, such as its focus on workers' qualifications, its relation to the quality and qualifications of human capital and its implications for economic growth.
References to PISA. Unlike the other top-five external targets, PISA's results were used at very specific moments during Legislature X: in only 4 of the 41 analysed debates (22 quotations, 6%). It first earned sporadic status as a (de-)legitimation tool in May 2005. PISA was mostly used by opposition parties, especially CDS-PP (right-wing) (8 quotations, 36%). Most references to PISA were made in a 2007 interpellation debate requested by the CDS-PP party to question and criticise the government on the state of Portuguese education. In his opening speech the CDS-PP deputy explained the reasons for requesting the interpellation, supporting his arguments by referring to the poor Portuguese PISA 2006 results. During the debate all the other parties also cited the recently launched PISA results while criticising the government for the education system's poor quality and inefficiency, high levels of inequality and students' low literacy and performance.
Furthermore, as was the case in other references, the use of PISA was often accompanied by other frequently used external references, very often to 'EU countries' or 'OECD countries', but also others as in the example below, in which a PEV deputy focused on the growth of social and economic inequalities during the XVII government: The 2006 data, which came first through the UNESCO study and then through the PISA Report, is very clear and shows again that we occupy, at almost all levels of performance, the tail end in relation to our European partners and the OECD's. (PEV deputy, 2007-12-06) Like other references, PISA helps to construct arguments on the need to develop an education system that promotes equity and students' learning, leading to improvement in their performance.
References to 'several/other countries'. Besides these more specific references, a broad and vague reference to 'several/other countries' is also common (22 quotations), especially in arguments used by government members (9 quotations, 41%). Like the references to 'EU countries', this reference is routinely used as a benchmark associated with a sense of belonging to the group.
The government's policymaking role means it is unsurprising that it is the actor that most often resorts to external references as epistemic capital (see Appendix 1). References to the EU or 'EU countries', the 'Bologna Process/Declaration', the OECD, or PISA, as well as to other external elements, are accompanied by transversal (and globalised) ideas such as the importance of improving the equality, quality and efficiency of the education system. Although these external references are mostly used by the government and its supporters, they are also referenced by other parties in the parliament to highlight problems that are described as requiring attention if the country is to develop into a functional society in which every citizen has opportunities to succeed. In a country like Portugal, which has shown the ambition and experienced the struggle to attain the levels of development and quality of life of other countries (Antunes, 2004;Gomes, 1999;Nóvoa, 2005), mentioning these values and ideals is highly strategic and appeals emotionally and rationally to the audience, foregrounding common wishes for and visions of future possibilities of a better life and, ultimately, of belonging to groups of 'others' perceived as performing better.

Discussion: Policymakers' assumptions in the epistemic work developed in the Portuguese parliament
Epistemic work departs to a large extent from the assumptions policymakers make about their audience. On the one hand, the analysis of which external references are used and how they are narrated in relation to shared ambitions and dreams reveals the assumptions that policymakers make about their audience. On the other, assumptions guide the selection of the external elements that are used as epistemic capital, because if the audience does not recognise these elements as authoritative, references to them cannot assist the speaker to strengthen his/her argument. Like other components of epistemic work, the assumptions have a cumulative character: in a single argument two or more types of assumption are identified.
The assumptions are directly connected with the objects of epistemic work ('ontology of the environment', 'actor identifications' and 'norms and ideals' -see Figure 1). Table 1 presents the main assumptions identified in the analysed parliamentary debates. It was possible to organise these into three categories: 'ontological assumptions', 'identification assumptions' and 'normative and ideological assumptions'.
'Ontological assumptions' concern the country and the government's performance -Portugal is seen as lagging behind other reference societies, mostly the 'EU countries', but also the 'OECD countries' or 'several/other countries'. These assumptions are unsurprising: they reflect the historically developed views of the country's development compared with others, especially European countries (e.g. Antunes, 2004;Nóvoa, 2005), as previously mentioned in this article. These countries are used as benchmarks in discussions focusing on the development of the education system and for a comparison and assessment of Portuguese students' performance (Nóvoa et al., 2014).
Arguments are therefore frequently built around this assumption, whereas the audience is assumed to consider them essential to achieving or sustaining policies and practices that converge with those of EU and OECD partners.
In relation to 'identification assumptions', the frequent use of both the EU and the OECD shows that it is assumed by the speakers that the audience identifies these organisations as relevant, and that their data and guidelines should be taken into consideration, despite these organisations' lack of regulatory functions for the education systems of their member countries. This applies to centre, right-wing and, occasionally, left-wing parties. It seems it is assumed that EU membership is viewed today as it was during the early years of the new democracy, as fundamental for Portuguese democracy and modernisation (Royo and Manuel, 2003). Their ideas, tools and advice serve as frameworks supporting diverse policy arguments. In addition, it is frequently assumed that the audience sees the OECD as an education expert. Statements from the organisation's studies, reports and recommendations are carefully selected by each party and used as a form of scientific evidence in support of policy arguments.
'Normative and ideological assumptions' are directly related to party ideology. While the rightwing party (CDS-PP) presents more conservative and traditional proposals, revealing assumptions directly related to values such as the importance of economic growth and a focus on meritocracy, the left-wing parties express ideas centred on the value of social and economic equity, and the need to respect the characteristics and needs of each person and nation beyond international trends or views, while defending the state's role in providing quality services accessible to every citizen. The centre parties assume, as expected, that their audience sees the world more moderately, presenting arguments combining arguments of left-and right-wing ideologies, such as the importance of having a public education system that provides equal opportunities to everyone, along with the relevance of each person's academic development, the improvement of their living conditions and therefore the country's social and economic development.

Ontological assumptions
-Portugal needs to improve and develop.
-Portugal lags behind other reference countries.
-The government is improving/worsening the country's development.

Identification assumptions
-The EU is an important organisation; Portugal's membership is relevant (or irrelevant depending on which party is speaking). -The OECD is a reliable expert.
-PISA data is reliable and trustworthy.
-The government is (in)efficient.

Normative and ideological assumptions
Left-wing parties -Importance of social equity.
-Importance of respecting each person's and nation's characteristics, needs and independence. -Fundamental duty of the state to provide quality services such as education to all citizens.

Right-wing party
-Importance of economic growth.
-Importance of free choice. Centre parties -Equal opportunities.
-Inclusion and economic growth as the means to improving living conditions for all citizens.

Assumptions and the construction of imageries
Assumptions are of great importance, not only in the selection of the external targets used but also, and in consequence of that selection, in the construction of the imageries that illustrate the reality as it is hoped the audience will perceive it. In essence, policymakers' assumptions concerning their audience's beliefs and interpretations are the basis on which policymakers choose certain externalisation targets when building the imageries used in the attempts to convince others of what the best line of action is. The analysis of the Portuguese parliamentary debates on education led to the identification of the same categories of imageries already introduced by Alasuutari and Qadir (2019) -'functional requirements of modernisation', 'society as competing blocs' and 'society as hierarchy' (see Figure 1). Two interesting features of these imageries also became clear in the analysis that are relevant to an understanding of the dynamic of epistemic work developed by policymakers in the Portuguese parliament and these actors' interpretations of their audience, both inside and outside the parliament.
Negative and positive imageries. The use of imageries was identified positively or negatively. For example, when attempting to persuade others of the significance of a proposed law, in the following quotation a government member positively uses external references to demonstrate the importance of such a law if Portugal is to fulfil the 'functional requirements of modernisation': [t]he international network [ENQA] prepared . . . the standards and guidelines for the quality assurance of higher education in the European area . . . that came to be adopted by all the signatory governments of the Bologna Declaration. This new law . . . is, in our opinion, a fundamental piece of reform for the qualification of Portuguese higher education. (Government, 2007-05-09) In contrast, in the next quotation, one of the opposition parties also uses external references to argue that Portugal is not developing sufficiently, failing to fulfil the 'functional requirements of modernisation', lagging behind other 'European countries' and losing the battle of 'society as competing blocs': Unfortunately, the numbers show we are behind the rest of Europe: the low education level of the Portuguese population . . . the persistence of high rates of early dropouts which compromise qualifications, the persistence of low levels of learning and success, and the existence of high social inequalities which schools have been shown incapable of countering. (PVE deputy, 2007-03-29) Imagery of national competing blocs. The analysis also showed that the imageries of society as competing blocs need to incorporate national competitions. For example, external references are used to feed arguments concerning the 'battles' between parties, while promoting their policy ideas and criticisms, or attempting to prove the efficiency of their actions. A PS deputy refers to an OECD report to accuse an opposition deputy of being distant from the government's actions and the country's development during Legislature X: In his initial speech the deputy said that he did not know the work of this government in matters of higher education. It revealed a great distance from the answers that have been given to the problems of higher education in Portugal . . . But it would be enough, I don't even say to read, but it would be enough to examine the OECD report to understand what the Portuguese government has already done. (PS deputy, 2006-12-21) The enduring national competition concerning the quality and performance of private versus public schools is another interesting example of national competing blocs. In the following quotation right-wing deputies invoke PISA results to argue for parents being allowed to choose their child's school, leading to more equality:  -PP deputy, 2007-12-06) Like the character of the epistemic capital and the assumptions, imageries are also cumulative. In the analysed debates arguments using the functional requirements of modernisation imagery arguably frequently identify imageries of society as competing blocs or society as a hierarchy.

Conclusions
The parliamentary plenary debates are rich in research possibilities, of which this article embodies one. Its premise is that governance is always epistemic and that 'actors try to steer others' thoughts and conducts' with the aim of initiating social change (Alasuutari and Qadir, 2019). In identifying and exploring externalisations to world situations as a provider of epistemic capital which is used in the epistemic work developed in Portuguese parliamentary education debates, I have brought to light a) Portuguese policymakers' assumptions about their audience, b) how and why specific external targets are chosen and c) the imageries they construct in their arguments. These clarifications help to understand the epistemic work developed in this context. This study therefore demonstrates that the two frameworks, externalisation to world situations (Schriewer, 1990;Steiner-Khamsi, 2004) and the EGF (Alasuutari and Qadir, 2019), can complement each other in a constructive analysis of policymaking and governance processes, promoting a deeper understanding of them. First, the analysis demonstrates that during Legislature X, in addition to the traditional ideological arguments, the strategy for political (de-)legitimation was based strongly on externalisation to world situations. The increased use of specific external references indicates that they constituted powerful tools of (de-)legitimation during this legislature. However, party ideology and cohesion remain as strong as they have been since the restoration of democracy in 1974 (Goes and Leston-Bandeira, 2019;Leston-Bandeira, 2004). The continuity of the strength of party ideology is expressed in the entanglements of the topics raised by different policymakers. Although the debates include globally shared ideas like the need to reduce inequality and improve the quality and efficiency of the education system, these ideas seem to take the shape of transversal goals or justifications for issues which, as Barroso (2016) argues, are topics that are repeatedly present in Portuguese education debates and on each party's agenda. It therefore becomes clear that global ideas are strategically selected and rhetorically used as sources of authority precisely in support of these enduring reform proposals.
Furthermore, in this analysis I demonstrate the complexity of the political interactions in the parliamentary context and the strategic character of the use of claims of knowledge and evidence in political contexts. Knowledge becomes mouldable: the same entity or practice is often referenced to accurately describe opposing realities. The epistemic work developed in the analysed debates foregrounds the mouldability of such claims. During Legislature X policymakers needed to find sources of authority beyond the traditional party-related ideological arguments, often resorting to externalisations as sources of epistemic capital in their attempts to convince their audience that their policy ideas were the best solutions for improving the education system, while their opponents' ideas were of no use or inappropriate.
Second, by examining in detail how the same external references are used by different policymakers in the Portuguese parliament, I bring to light the assumptions ('ontological', 'identification', 'normative and ideological' -see Table 1) these policymakers make about their audience, highlighting an element of analysis that has yet to be central in the studies developed through the EGF lens. This focus promotes a more exhaustive comprehension of the epistemic work process. The identification of policymakers' assumptions reveals that neither the imageries used to illustrate the presented arguments nor the external references used to construct these imageries are randomly chosen. The choice of these external elements and their use depends on the speaker's assumptions of how the audience understands the debated issues, their perceptions of which IOs and international tools the audience considers reliable, the international practices the audience appears to consider worth pursuing and the ideas they see as valuable or acceptable. In sum, the assumptions are the foundational guide for the speakers' understanding of which references their audience will accept as valid authorities and which will help garner support for the advocated policy ideas.
In the case of the analysed parliamentary debates in education in Portugal these assumptions are strongly related to historically, culturally and socially constructed representations of the EU and the OECD, and their perceived relevance for the country's continuing educational development, as already reported in other studies (Antunes, 2004;Gomes, 1999;Mateus, 1999;Santos and Kauko, 2020;Teodoro, 2019). In examining the entire 2001-2018 timeframe, it is noteworthy that they are the only external elements always present in the list of the ten most used references (Appendix 2). During Legislature X the top five external targets are these international organisations, their associated countries (as a regional reference society) and their tools of guidance and assessment -the Bologna Process/Declaration and, later, PISA. These constitute the main epistemic capital used by policymakers in developing epistemic work in Portuguese parliamentary education debates. This observation aligns with earlier studies focusing, for example, on the choice of reference societies (e.g. Waldow and Steiner-Khamsi, 2019), demonstrating that historically constructed relationships are a decisive element in the selection of the countries or regions used as references.
Third, the identification of the nuanced use of external elements provides information about the strategic intertwinements of the imageries constructed by each policymaker with their party's ideology. For example, while the OECD was often portrayed by all parties as an expert producing reliable information and evidence, it also caused some controversy in the debates. It is clear that the OECD's advocacy of economic aspects, such as the increase in people's qualifications with the aim of generating economic growth, were adopted by centre and right-wing parties but were contested by left-wing parties, which referred to the organisation's data in arguments focusing on the improvement of qualifications as a path to increased social equity.
While epistemic capital can be anything seen by the audience as authoritative (not necessarily external references), by utilising the EGF in the empirical analysis of externalisation as a source of epistemic capital my analysis attempts to contribute new theoretical ideas to the research into externalisation to world situations (e.g. Schriewer, 1990;Steiner-Khamsi, 2002, 2003Takayama, 2008Takayama, , 2009Waldow, 2017). The two frameworks can fruitfully benefit from each other. In this study the epistemic governance framework deepened the understanding of the 'sociologic', which in the Portuguese parliamentary context leads to the choice of certain world situations as symbolic elements of (de-)legitimation in the epistemic work that aims to earn support for specific policy ideas, particularly regarding controversial topics. This analysis therefore buffers answers not only to the questions of which external references are used and why but also how they are used in the work of convincing others of the viability of certain policy ideas. It emphasises the complexity of global-local intertwinements: while IOs, their tools and practices in other countries/ regions influence policymaking processes at the national or local levels, policy actors in specific national contexts (such as in the parliament) utilise the prominence of these external elements as epistemic capital when their traditional sources of authority are insufficient to persuade their audience of the validity of their arguments.
Future studies may continue to expand this analysis by exploring the use of externalisation to world situations as a source of epistemic capital in other governance arenas at the global, national and local levels and their interplay, broadening its coverage and allowing a comparison of the roles that externalisations play in the epistemic work developed in a wide and diverse range of contexts. Target