A note on the scope of this article

The purpose of this article is not to point fingers or apportion blame; rather it is about us collectively taking responsibility, both for the current state of the discipline and the actions we can each take to bring about positive change. The driving force behind this article is, predominantly, race and ethnicity. I recognise, however, that there are numerous intersectional issues at play here relating to gender, sexuality, class, religion, disability, and precarity; where possible, these intersections are drawn out. I would encourage readers to conduct more studies of this nature, so that we can begin to draw out these intersections with more nuance and understand in more depth what a fully inclusive Roman Archaeology might look like.

Acknowledging bias

Before analysing the data in depth, I would like briefly to acknowledge my own biases and privileges. Firstly, and for me most crucially, I am British Iraqi and grew up in the UK (on being of mixed heritage in the UK, see Hirsch 2018). I identify as a woman of colour, but I am ‘white-passing’, which gives me immense privilege in my day-to-day life as I can walk down the street, go shopping, etc. without being subject to many of the microaggressions and racist abuses that are a sad, difficult, and seemingly unavoidable part of life for many people of colour. As will be seen below, this has not protected me entirely from racism and microaggressions, much of which I experience due to my name, which is clearly not English. My family background is complex and tainted by the many conflicts Iraq, and my family in Iraq have suffered; again I have been subject to privilege by growing up in the UK. Furthermore, I have had the benefit of an excellent education, with undergraduate and postgraduate degrees from the University of Oxford.

My experiences as a woman of colour in academia have moulded my position strongly on this topic, such that I believe that diversity, inclusivity, and equity can only ever be positive in academia. In addition, my consciousness of my own biases leads me to believe that no piece of academic work takes places outside of our social, cultural, and political milieux and that to operate otherwise is misleading (of oneself and others) and potentially damaging (e.g. in not being able to acknowledge that one’s own background has influenced one’s work and so may have, inadvertently or otherwise, excluded other perspectives).

WHO participates in RAC/TRAC conferences?

All session organisers and paper presenters have been assigned one of the following categories in the dataset:³

• B: BAME

• D: affiliated with an institution in a development-assisted country (DAC); not BAME⁴

• ?D: probably affiliated with an institution in a DAC country; not BAME

• I: Israeli⁵

• N: not BAME and not from a DAC country

• ?N: probably not BAME and not from a DAC country

• §: unknown⁶

As noted above, data regarding a participant’s ethnicity are difficult to extract from the programmes alone. In order to err on the side of caution, many people have been put into the ‘?N’ category.

The benchmark used for these data is the UK 2011 census, which demonstrated that the white ethnic group made up 86% of the population of the UK (Office for National Statistics 2012). We should, then, expect to find a minimum of 15% BAME scholars attending conferences to reflect the wider UK population.

Figure 2 and Table 2 show the percentage and count (respectively) of session organisers and paper presenters by ethnicity. These figures are stark. Only 2.4% and 2.2% of session organisers and paper presenters are BAME. No one from a DAC country (and not BAME) has ever organised a session and a meagre 0.6% have presented papers. This falls far short of the 15% that we would hope to see if these conferences reflected UK society. It is possible that additional BAME participants are hidden in the ?N category. In order to meet the 15% threshold, we would need to find 78 additional session organisers and 365 additional paper presenters. Given the RHS data and the low DAC representation, this number of unrecognised BAME people in the dataset seems highly unlikely. Furthermore, these patterns also reflect those seen at other conferences, for example the International Congresses of Papyrology (Blouin et al. 2019).

Figure 2

Session organisers and paper presenters by ethnicity (%).

Looking in more detail at session organisers, of the identified 15 BAME organisers, 3 have organised 2 sessions, which lowers the total number of BAME individuals further. In addition, only 12 sessions have an identified BAME organiser.⁷ Only six sessions have an identified BAME person named as first organiser, five of whom are men; the only woman is me. The data suggest that there are some intersectional issues at play here as identified BAME men are c. three times more likely to have been a session organiser than identified BAME women, though the numbers are small (11 men; 4 women). Two identified BAME organisers were based in DAC countries (possibly three, but affiliation not given). In contrast, seven were based in the UK, three in mainland western Europe, and two in the US, which suggests that there is an advantage to being based in a university in the Global North. There have never been more than three identified BAME session organisers at a single conference. There is a marginal improvement in identified BAME session organisers in recent years (Figure 3; Appendix 1.1), especially in 2019 with 10.9% (3 by count), but still well below the 15% threshold and no-one identifiable in 2014 and 2015. The first identified BAME man to be first organiser was in 2006 and the first woman was in 2010; it was me. I am also the first BAME person to deliver the keynote lecture at a TRAC conference since it started 29 years ago.

Figure 3

BAME and DAC (not BAME) session organisers and paper presenters over time (%). Count = Appendix 1.1 and Appendix 1.2.

As with the session organisers, there has been a marginal improvement in representation in identified BAME paper presenters over time (Figure 3; Appendix 1.2). In the first ten conferences in the dataset, only one person identified in the dataset as BAME presented a paper: a man in 1999. The first BAME woman identified in the dataset presented seven years later in 2006. Only one conference has identified BAME presenters in double figures: 12 presenters in 2018, but this is only 3.5% of the overall total for that year. The highest percentage of identified BAME presenters is only 7.1% in 2011; most fall well below the 15% threshold. Some identified BAME people return as presenters to the conferences: three people have presented twice; one person three times; one person four times (me) and one person five times from 2006 to 2010. Although I have also been a session organiser, it is notable that the person who came five times never organised a session and is a woman. This raises two issues. First, why was this pattern of regular attendance not spotted and this person, seemingly, not encouraged to organise a session? Second, the fact that this person is a woman of colour also seems to bolster the potential negative influence of intersectionality on session organisers noted earlier. That said, it is pleasing to note that there is gender parity in identified BAME paper presenters (31 women and 31 men).

WHERE are participants based?

In order to draw out these patterns most clearly, the institutional affiliations of session organisers and paper presenters were grouped by country and then by synthesised country group categories. The main level of analysis is at synthesised country level with the individual country level used to draw out relevant underlying patterns. As is clear from Figure 4 and Table 3, session organisers and paper presenters based in the UK dominate the conferences both as session organisers (47.9%) and as paper presenters (44.3%). When added together with participants based in mainland western Europe the dominance of all western Europe is even more emphatic (session organisers: 68.7%; paper presenters: 72.1%). A strong reason for this, of course, must be proximity to the conference location, given that most conferences have been held in the UK. This, for example, might explain the under 2% of session organisers coming from Australasia. It does not, however, explain the disparity between session organisers from eastern Europe (1%; NB: none are from DAC countries in eastern Europe)⁸ and MENA (1.3%) compared to session organisers from North America (7.7%). This begins to suggest that we are looking instead at larger structural issues around access to funds and ease of obtaining visas (discussed further below). This case is further supported by looking in more detail at the MENA category for session organisers, where of eight people, five are from Israel, the person from Turkey is western European, and only two are BAME (from North Africa). It is also worth noting that significantly large parts of the world are missing from session organisers: south and central America, China, the Indian Sub-continent, sub-Saharan Africa and Russia. Although some of these gaps are filled by paper presenters, the picture is not vastly improved. Nearly 50% of MENA paper presenters, for example, are from Israel (29/61), which is highly disproportionate in favour of one of the few countries in the region that is not DAC. Indeed, even when all DAC countries are added together across the various categories (1.8%),⁹ they are demonstrably and substantially less well-presented than North America (7.6% including 2009 Michigan; 5.3% excluding 2009 Michigan). Although the numbers are small, the affiliations of BAME people again suggest a slight advantage to being a BAME person in a university in the Global North (32/62), in comparison to being based in DAC and MENA countries (23/62).

Figure 4

Session organisers and paper presenters by synthesised affiliation (%).

Over time, we do see some shifts in the institutional affiliation of session organisers, though the UK dominates overall and only loses its dominance when the conference is held outside of the UK, when session organisers from mainland western Europe or North America (in Michigan 2009) dominate (Appendix 1.3). From 2015 onwards, session organisers from mainland western Europe have consistently made up over 20% of the session organisers, with the exception of 2019 where the (unrealised) possibility of Britain leaving the EU around the time of the conference may have served to discourage attendees from mainland Europe. Holding the conference outside of the UK also has a slight effect on MENA inclusion with the first MENA-based session organisers appearing in Michigan 2009 (two from Israel). Startlingly, and demonstrating that this problem goes beyond ‘decolonising’ to ‘diversifying’ more generally, the first session organisers based in eastern Europe do not appear until 2016 in Rome. Neither MENA nor eastern Europe go above 6% of session organisers at any single conference and each only approaches these levels once. It will be interesting to observe the extent to which this pattern changes at RAC/TRAC 2022 in Split (TRAC 2020 (postponed 2022) n.d.).¹⁷

Similarly, in paper presenters we do see a gradual lessening of UK dominance over time, though mostly this involves increasing participation from mainland western Europe and, to a lesser extent, North America (Figure 5; Appendix 1.4). While mainland western Europe and North America are almost always represented in paper presenters,¹⁸ this is not the case for DAC countries: MENA DAC is not represented at 13 conferences; eastern Europe DAC not at 19; and world other DAC also not at 19. In terms of numbers, MENA presenters reach double figures only once in 2016 (14), where it equals North America (13), otherwise representation is usually under 5 people. Participants from MENA DAC countries, i.e. not including Israel, never reach double figures. We may also be seeing the toll that conflict takes on scholars and scholarship as conflict in the former Yugoslavia (1991–2001) may well account for why we do not see a paper presenter from an eastern European DAC country until 2005. By percentage, MENA, World other DAC and eastern European DAC countries are never over 10% (let alone 15%…) of paper presenters at a single conference. MENA paper presenters reach over 5% three times in 2009 Southampton (5.3%), 2013 (5.8%), and 2011 (7.1%); without Israel only once in 2009 Southampton (2013: 3.8%; 2011: 4.8%). In contrast, the UK – a single country – only goes under 40% in years when RAC/TRAC conferences are not held in the UK and by just 0.6% in 2018 in Edinburgh.

Figure 5

Paper presenters by synthesised affiliation over time (%). Count = Appendix 1.4.

WHAT topics and which places are we discussing at RAC/TRAC conferences?

Themes for both sessions and papers have also been synthesised to facilitate analysis, so any discussion of themes here refers to synthesised themes. The spreadsheet allows for sessions and papers to have more than one theme, so total theme numbers are higher than total numbers of sessions and papers.¹⁹ In both cases, the sets at the top end are predictable and largely traditional (Figure 6; Appendices 1.5 and 1.6). At the bottom end – under 5 for sessions and under 35 for papers – we find the themes where a progressive agenda might most obviously be under discussion: heritage, museums, migration, beyond frontiers, and decolonising. This is disappointing as it suggests that opportunities to discuss these important topics are not being taken regularly. It is also noteworthy, if not staggering, that diversity (2017) and ethics (2008) have only had one dedicated presentation each in all the conferences.

Figure 6

Sessions and paper by theme (count). Count = Appendix 1.5 and Appendix 1.6.

Of course, it is also important that we begin to see these issues as central to all our work and therefore do not slip into the potential trap of only discussing them in specifically-tailored sessions. Topics in the top categories, such as identity, military, theory, and cultural change, would all lend themselves to such work; personal experience of attending papers and sessions with these themes suggests that while a post-colonial lens may now be being applied on a regular basis, a decolonial perspective is rarely taken. Likewise, it is instructive to note that BAME people have organised sessions on a wide-range of themes, including, but definitely not limited to a progressive agenda, so we must also be wary of assuming that all BAME people want to dedicate their time to decolonial research (see Atkinson et al. 2018: 89 on tokenism).²⁰ Indeed, the burden of labour must not be placed on the shoulders of those in the minority, but rather needs to be shared with allies, in order for positive change in this area to be effective and sustainable.

The strongest biases in our current research practices come to the fore when looking at what regions are most commonly discussed at RAC/TRAC conferences. Again, these regions are synthesised to facilitate analysis, and sessions and papers may span more than one region.²¹ As demonstrated in Figure 12 and Table 4, most sessions (69.3%) are ‘open’ i.e. the organisers did not specify a particular region for discussion. Drilling down into the Europe data, Britain and Italy are the two highest named places and the only two places that make over 5% of the total: Britain = 28 sessions, 7.3%; Italy = 21 sessions, 5.5%. Over time, Europe is the only specified region that regularly has more than one dedicated session in a conference year. Everywhere else usually has one or none and never over three; for example, MENA had three dedicated sessions in 2009 Michigan and in 2018; regions beyond the frontiers had two dedicated sessions in 2017 and three in 2018. In contrast Britain and Italy, almost always have dedicated sessions and occasionally break the three-session threshold: four sessions on Britain in 2007 and six on Italy in 2016 (Figure 7a). Although specific sessions relating to MENA are increasingly included in conferences, given that Britain and Italy are significantly smaller in size than MENA, which is approximately half of the Roman Empire, this is a strong disparity.

Figure 7

a) Sessions that specify Britain, Italy, and MENA over time (count). b) Papers discussing Britain and Italy in comparison to the rest of the Roman Empire (%).

This pattern becomes stronger, and more concerning, when looking at papers where it becomes clear that an ‘open’ session does not really mean ‘open’: effectively, ‘open’ means Britain or Italy (Figure 7b; Table 5).²⁵ In papers we see a large bias towards Europe with 68.3% of papers discussing Europe alone and in combination with other places. Papers discussing MENA, in contrast, make up a meagre 9.6%. Within Europe, Britain and Italy are overly dominant. The total of all papers discussing Britain is 668 (41% of Europe papers and 28% of all papers); the total for Italy is 446 (27.4% of Europe papers and 18.7% of all papers). Together, then, Britain and Italy make up 46.7% of all papers and are massively over-discussed at RAC/TRAC conferences in comparison to the rest of the Empire (Figure 7b).

Over time, Europe’s dominance in papers is, to all intents and purposes, unchanged, dipping below 50% only once in 2013 and up to 80% four times, most recently in 2014 (Figure 8; Appendix 1.7). For the past five years, papers on Europe have been in the 60–70% range. Within Europe, papers on Britain were very dominant until 2008, when Italy began to rise in prominence at the conferences (Figure 9; Appendix 1.8). Britain has only dipped below 20% of all papers three times (2012, 2013, and 2016) and of these only 2016 (Rome: 7.8%) is below 10%. Over the past five years papers on Britain have been in the 20–30% range, which is marginally less dominant than the first five years of 30–60%. Britain and Italy combined are extremely dominant across time with only a very slight downwards trend (Figure 9). In the past five years, papers on Britain and Italy combined have been in the 35–60% range and have never been below 50% of all papers on Europe. So, while Britain’s dominance might have declined, this has not seen a concomitant increase in the rest of the Empire, other than Italy, which is disappointing.

Figure 8

Papers by region over time (%). Count = Appendix 1.7.

Figure 9

Papers on Britain and Italy (% of all papers). Count = Appendix 1.8.

A small amount of increasing attention has been given to areas beyond the frontiers from c. 2012. MENA has been consistently present in papers since 2001, but has not increased in that time, usually being within the 5–15% range; papers on MENA have been over 20% once in 2009 Michigan (22.7%). Again, this clearly does not represent the size of the region (approximately half of the Empire), nor the amount of time it was under Roman rule. Breaking down MENA into discrete sub-regions, the Middle East (98) and North Africa (99) are each twice as popular as Asia Minor (48), with Egypt making up 38.3% of papers on North Africa (Appendix 1.9); there is no particular pattern over time. It is also striking, negatively, that not a single sub-region within MENA has been over 20% of all papers, unlike Britain which has only been under 20% three times.

When we combine these data on papers by region with affiliation data, further patterns arise that begin to reveal biases within research cultures (Figure 10; Appendix 1.10). Regardless of affiliation, most people (50–75%) are presenting on Europe, with the notable exception of people based in MENA countries, for whom papers on the MENA region equal 68.9%. If the conferences are perceived as Euro-centric, this may be a further barrier against participation from people based in MENA countries. The highest percentage of people presenting on Europe are UK-based academics (74.5%). Furthermore, the responsibility for the Britain and Italy bias does seem to lie at particular doors (Appendix 1.11): 50% of people based in the UK present papers on Britain; scholars based in North America and mainland western Europe are more biased towards Italy (31.5% and 31.3% respectively). It is also noteworthy that eastern European scholars are not contributing to this bias, and also that people based in North America are slightly more likely to present research on MENA.

Figure 10

Papers on Europe, Eastern Empire, MENA, and beyond the frontiers by scholars from MENA, North America, UK, and western Europe (%). Count = Appendix 1.10.

Taking action

Teaching by regions of the Roman world

Respondents were teaching on a wide range of programmes from specialised archaeology degrees to broader liberal arts programmes at undergraduate and postgraduate level (Qs 2, 2a and 2b). Q5 asked respondents to tick which places in the Roman world they teach in their Roman Archaeology courses, including an option for places beyond the frontiers. The list provided was not comprehensive, but provided coverage across most parts of the Roman world to see whether some regions were taught more than others; an ‘other’ option was provided for respondents, who wished to flag particular provinces or regions not listed and was mostly used for Germany, Pannonia, the Rhine, and the Danube. The data show a much more even distribution of places taught than was seen in the conference data, though it is worth noting that the top four places taught are, perhaps, predictable: Rome (116; 85.9%), Pompeii (104; 77%), Italy (not including Rome, Pompeii, or Herculaneum: 101; 74.8%), and Britain (99; 73.3%) (Figure 12; Appendix 3.2).

Figure 12

Places taught in general Roman Archaeology teaching and specialised regional courses in the RATS survey (%; total responses = 135). Count = Appendix 3.2 and Appendix 3.3.

In the specialised regional courses (Q5a), however, we begin to see the regional biases that also exist in the conference data with 25.4% of respondents who teach specialised regional courses (16/63) teaching Britain and another 25.4 % teaching Italy (Figure 12; Appendix 3.3). Indeed, as many people are currently teaching a specialised course on Pompeii (i.e. just one city) as are teaching the Middle East (six respondents each). Again, the dominance of Britain and Italy is almost to the exclusion of the rest of the Roman world. Furthermore, we see a similar relationship between the specialised region being taught and country of affiliation as was seen in the conference data between research region and country of affiliation. Namely, 13 out of 16 specialised courses on Roman Britain are being taught in Britain and 7 out of 16 specialised courses on Italy are being taught in the US, with a further 5 in Britain (Appendix 3.3). There are currently no specialised courses being taught in Britain on Iberia, the Balkans, the Mediterranean, Egypt, and Germany and only two (each) for the Middle East, North Africa, and Asia Minor. Again, even with account being taken for potential employability reasons for teaching Roman Britain, this is a disproportionate emphasis on a very small part of the Roman world.

When respondents were asked why they were not teaching certain places (Q5d), a range of responses were given (Table 8). The most common response (19/66) was that they did not have enough knowledge. There was some laudable honesty about personal biases e.g. respondent #7 (Asst Prof; US) says: ‘I’m more comfortable in my knowledge of the western Mediterranean than the eastern, and I’m aware of that skew too.’ I would, however, like to see a further push to address that skew. One of the most illuminating comments came from #127 (Senior Lecturer; UK), who noted that ‘I wasn’t taught those areas as an undergraduate’. Here we begin to see the underlying issues at play: a circle of influence whereby what we learn as undergraduates becomes our research interests, which in turn becomes what we teach. This is a natural circle, but unless we begin to change the input, i.e. what we teach, this circle will only reproduce itself and never expand to redress the current imbalance in our teaching and research.

Furthermore, some respondents noted that one of the barriers to making change in their teaching came from senior colleagues. Respondent #19 [postdoc; US], for example, stated that ‘several senior scholars in my department told me that I was teaching courses that are “too niche” and I should try to do something “more normal” (meaning Pompeii or Rome or Latin), so that I will be more appealing on the US job market’. This is extremely disappointing to hear. Such attitudes serve only to entrench disciplinary inertia, strongly discouraging upcoming scholars from being innovative and pushing our discipline forward.

Teaching by topics

The spread of topics taught is also not as stark as at the RAC/TRAC conferences, but a similar set of topics can be found at the lower end, with slavery, migration, museums and ethics, and politics all featuring in the bottom five (Figure 13; Appendix 3.4). It is also interesting to note that while marginally more people are teaching multiculturalism and the history of Roman archaeology, they are doing so without also teaching slavery, migration, and ethics/politics, which suggests that these topics are not being taught at their fullest. Question 4d asked survey participants to explain their reasons for not teaching two or more of history of Roman archaeology, ethics and politics, museums, gender, multiculturalism, identity, migration, and slavery. There were 53 responses to this question, of which 17 were supportive of these topics. A range of responses were given by the 36 people who responded to explain why they were not teaching these topics (Table 9).

Figure 13

Topics taught according to the RATS survey (%; respondents = 133). Count = Appendix 3.4.

A small, but nevertheless worrying, number of respondents was blatant in their disregard for the value of these kinds of topics. One respondent (#120; Prof.) replied by saying this was ‘bullshit political correctness’ (see further on reading lists below). This was the most extreme response, but echoed respondents e.g. #38 who were worried that their institution was ‘heavily involved in these topics’, so felt they needed to make sure their students ‘get a diverse education’ and ‘have access to other visions/subdisciplines’ [US; Visiting Assistant Prof.; responses to 4c and 4d combined]. This is a curious response as it implies they feel their institution has gone too far in the opposite direction and instead wants diversity through a more traditional set of topics, which seems somewhat incompatible.

The majority of responses to Q4d claimed that there was not enough time (n = 11), these topics are outside the syllabus (n = 9) and are taught by other people (n = 5). This response was typified by #11’s comment: ‘No time for history of archaeology, ethics and politics, museums (we barely have time to install the essentials)’ [Lecturer; Germany]. The implication in comments of this kind is that there is some kind of canon that we must teach on Roman archaeology courses. This begs the question of who created that supposed canon, whether it is time for that supposed canon to change, and who makes decisions over what we teach. My response to the last question would be that for most people with tenure and on permanent contracts, we make our own syllabi (I acknowledge, as shown by the two respondents who were not in control of their own courses, that the situation is somewhat different for more precarious teachers). For those of us in that privileged position, there is rarely someone prescribing the content of our courses to us. We each decide on what is essential, so to say there is no time in a course or that it is outside the syllabus (which you created) is to say that you do not consider it a priority. This is, actually, not that far away from the stronger views being expressed by e.g. #120. A responsibility is tied to course creation; this should not be hidden behind a ‘no time’ excuse.

On a positive note, however, it is worth noting that 17 respondents to Q4d took the opportunity to respond with positive reasons for including these topics in their teaching. One respondent explicitly noted that ‘it is important to… inform students about the shaping of ideas and the role of bias in many conventional notions about the Roman world’ [#1; US; Assistant Prof.]. Another observed that rather than devoting particular courses to these topics, they ‘consistently play a role in my teaching’ [#97; US; Associate Prof.]; this interweaving of these topics in to all that we teach is a useful direction for pedagogical thinking about Roman archaeology (see also Atkinson et al. 2018: 82–83). Furthermore, of the 35 respondents to Q4c (comment on reasons for making changes to topics being taught), 6 answered that it was due to changes in current scholarship and 5 that it reflected the interests of their students (so providing a counter-balance to the respondent to Q4d who felt that these topics were not in their students’ interests). Comments included that these topics ‘make students excited’ [#32; US; Assistant Prof.], that ‘students engage most with the lectures that incorporate these aspects’ [#85; US; Assistant Prof.] and that these ‘issues come up more in student questions’ [#26; US; Assistant Prof.]. These perspectives are very encouraging and suggest that we may be on the cusp of change, driven by students and what I hope will become the future generation of tenured/permanent academics.

Reading lists and the meritocracy issue

Questions 6a–6d around reading lists were, perhaps, the most contentious parts of the RATS survey. Questions 6a and 6b asked whether respondents had made a deliberate effort to include BAME people, women, and early career researchers on their reading lists and to offer an explanation, if they had not. While some progress seems to have been made regarding the inclusion of women on reading lists with 50.8% respondents having made a deliberate effort to include women, only 27% of respondents have done the same for BAME scholars and 34.9% have made no effort at all to include any of these groups on their reading lists (Figure 14; Appendix 3.5). In addition, it is worth noting that people teaching in the US have been more willing to make change than in other areas. People teaching in the UK and mainland western Europe are significantly behind the US with 40% and 60.9% of UK and mainland western Europe respondents, respectively, answering none in comparison to 8.3% of US respondents.

Figure 14

Respondents who have made a deliberate effort to create inclusive reading lists or not (%; respondents = 126; some respondents have deliberately included more than one group). Count = Appendix 3.5.

Respondent #120 is again instructive (negatively) here. Regarding their strong views, #120 clarified that:

‘The intentional questions regarding gender etc I think are not useful for promoting and improving teaching in Roman Archaeology and reflect much more political than scientific views. In Germany, we only judge researchers according to ability, achievements and aptitude.’ [email correspondence – quoted with permission]

There is a lot to unpack here. Firstly, the assertion that the survey is ‘more political than scientific’ demonstrates that this respondent is of the view that our study of the past can be objective; this is a position that has been thoroughly debunked in the past 20–30 years. Most researchers across a broad sweep of disciplines would now agree that objectivity is a veneer and that all research takes place from a position of bias, be that positive or negative. Secondly, the accusation that the survey is political is in itself to take a political stance, one that rejects an inclusive and decolonial agenda. Finally, the claim that this respondent only judges other scholars ‘according to ability, achievements and aptitude’ betrays a lack of awareness of, or possibly even a lack of willingness to engage with, the idea of implicit bias. Again, this comes from a refusal to acknowledge that all the selections, decisions, and choices we make in our research and teaching are not objective and rather are subject to bias.

What worries me most about the comments from #120, who expressed some of the most extreme views in the survey twice calling it ‘bullshit political correctness’, is that there are distinct overlaps with #120’s opinions and those of many people who had not made deliberate efforts to create inclusive reading lists. Of the 85 synthesised reasons given for not making a deliberate effort to create inclusive reading lists, 62 (73%) claimed that: their reading lists did not require attention, they chose on quality and relevance, and, emphatically, did not discriminate (Table 10).

Looking in detail at some of the specific reasons given in Q6b nuances the choices being made. Respondent #9 (Lecturer; Netherlands), for example, states that they ‘start from the canon’, which again raises the issue of what, precisely, constitutes the canon, who created it and what biases are inherent within such a canon. Respondent #14 (Lecturer; UK) specifies that they use relevant material that is ‘published using appropriate current academic practice’. This is an interesting stipulation as it brings up issues around publishing and who is welcomed into the world of academic publishing; work looking at publishing biases also suggests that similar sets of problems are at play (Wenneras and Wold 1997; Budden et al. 2008; Mitchell et al. 2013; Kaatz et al. 2014; Kelly et al 2019; ), so a bigger question becomes how we make sure that published scholarship is more diverse in order to create inclusive reading lists. Finally, are those who are suspicious of attempts to create deliberately inclusive reading lists:

• #22 Senior Lecturer; UK: ‘I will choose them because they are worth including, not for some box-ticking reason’

• #114 Associate Prof: ‘reading lists are focused on (e)quality, not discrimination’

• #124 Prof; UK: ‘I will choose them based on what is written, not who wrote it and I am concerned by attempts to do otherwise’

These responses – and their focus on quality, relevance, and non-discrimination – seem to show a disturbing lack of awareness of implicit bias within the profession. There is now an ever-growing number of studies that demonstrate that implicit bias is alive and well across all areas of academic life, with negative bias being routinely shown due to gender, race and class (Wenneras and Wold 1997; Bagilhole and Goode 2001; Budden et al. 2008; Knobloch-Westerwick et al. 2013; Mitchell et al. 2013; Kaatz et al. 2014; Handley et al. 2015; Storage et al. 2016; Criado Perez 2019: 92–111; Friedman and Laurisen 2019). The claims by respondents to this survey not to be discriminating are at best naïve. The idea that by making a deliberate effort to include e.g. women and BAME scholars, you would be discriminating (presumably against privileged white men) is laughable. We can no longer take refuge in the myth of meritocracy; deliberate efforts towards inclusivity are the only way to combat this persistent bias.

Another related issue that came out in response to Q8 (any aspects that a respondent felt had not been covered in this survey) was publication language. Respondent #124 [Prof.; UK], for example, wrote:

‘Most pernicious is the view that scholars can discuss Roman Archaeology without engaging with languages other than English. This is in fact a form of prejudice and it contributes in its way [to] enhance the dangerous insularity we are now encountering. Students, at both UG and PGT level should be exposed to non-English texts… Roman archaeology could and should lead a fightback against this trend.’

In addition, this issue cuts both ways. For colleagues in non-English speaking countries the problem becomes what to put on reading lists that is not in English. Colleagues from e.g. Jordan and Russia in the survey pointed out that the bias towards Anglophone publication made it difficult to find material to give students whose English was not of a high standard. We must work hard to make sure that we are not potentially alienating talented students and future scholars by imposing language barriers. Actions to counter this are discussed below.

Taking action

As previously, there are numerous actions on a range of levels that we can each take to shake us out of this disciplinary inertia and to continue the good efforts that are already underway.

Students, would you like to see changes in your syllabus? Then, lobby your lecturers, explaining why particular subjects are interesting and relevant to you. You have more power than you think in a university setting.

Academic staff at all levels should consider taking an implicit bias test, for example Harvard Project Implicit (2011), and act on those results: you could start by looking at the suggestions given by Umachandran (2017) for countering white fragility. Learning a new region or topic to teach that is outside your comfort zone will enrich your teaching, help to disrupt the teaching>research>teaching cycle and may even open up new research avenues. This need not be an onerous task; introducing a single new case study would be a good starting point.

Furthermore, this is not simply about what we teach, but also about how: are you teaching in ways that create an inclusive classroom? Some of the topics I am advocating for a decolonial Roman Archaeology – e.g. slavery, migration, multiculturalism, ethics – have the potential to raise high feelings and so need to be handled with delicacy; for useful and adaptable suggestions on how to manage such topics, see Mairs 2019 and MRECC n.d. Furthermore, it is worth finding out who is in your classroom, what preconceptions they might bring with them and as a consequence how they respond to more innovative material. You can do this by creating your own survey (as a supplement to those provided by university managers); for inspiration, see Blouin 2019 and my homage to Katherine Blouin in Kamash 2021. You may also find that your students speak a wider range of languages than you assume.

If you believe an essential canon exists for Roman Archaeology, take some time to consider what that means and whether disrupting those previously-held beliefs might be beneficial both to your teaching and to your students. Similarly, take some time to check the balance of your reading lists. Are you really as inclusive as you might be? One simple way to expose the balance of your reading lists, at least for gender and potentially also for race, is to include first names; this amendment takes a very short time and will make both you and your students think. This is an action that journal and book editors can also take: change your house style to include first names. I asked TRAJ for permission to do so in this article and am pleased that it was granted; this is encouraging for the future. Journal and book editors also have a responsibility to check the balance of their publications: ‘manels’ and ‘wanels’/’whinels’ should also not exist in our publications. We need to make sure that diverse voices are present in our publications, so that we can create inclusive reading lists for our students. Again, research and teaching are interlinked here. Across all of these actions, I would make a special plea to senior colleagues to embrace these actions, set precedents and support upcoming scholars to make these changes in their own practices: no-one should be threatened with not getting a job for being innovative and forward-thinking.

The language issue is more of a challenge. In the UK there is a widely-acknowledged crisis in modern language teaching (APPG for Modern Languages 2019); this will take some time to reverse, but there are efforts to do so as discussed in the APPG report. We can contribute to these efforts by encouraging our students to develop their language skills at university. If a student wishes to undertake a PhD in a Middle Eastern country, for example, then that should also include learning Arabic, as well as the language of archaeology publications, which often reflect colonial histories (e.g. the predominance of French-language excavation reports from Syria). With the advent of translation software, it should also be possible to begin thinking about automating translations of publications; the results may not be perfect, but this would ameliorate the situation. Again, journal and book editors have a role to play here in e.g. providing multi-lingual abstracts of articles and chapters.