On 17 July 1929, time stood still on Hadrian’s Wall. The moment is enshrined in a photograph taken by Ian (later Sir Ian) Richmond of the excavation at Birdoswald undertaken by the Cumberland Excavation Committee (Richmond and Birley 1930). Eight men stand around a mock Roman altar perched on stone flags, pouring a libation (Figure 1). Three of the men, F. Gerald Simpson, Robin Collingwood and Eric Birley, were to dominate Hadrian’s Wall studies for nigh on 70 years (Breeze 2014a; Stewart 2024). When the flags were lifted on 18 and 19 July, after Collingwood’s departure, it was discovered that two were building inscriptions, one dated to about AD 200 and the other a century later (RIB 1909, 1912). Their discovery led to the formulation of a new hypothesis (theory really) for the history of Hadrian’s Wall, which became known as the ‘Wall periods’ hypothesis (Birley 1930: 164–174). The history of the Wall was divided into four main periods: AD 122–197 (with a break for the occupation of the Antonine Wall); c. AD 200–296; c. AD 300–367; AD 367–385 (this final date was later modified to about AD 410), with each of the first three periods ending in violent destruction (Breeze 2014a: 91–100). Thereafter, for the next 40 years or more, the results of all excavations were shoehorned into this framework and all guidebooks reworked to reflect the new interpretation.

It is, however, instructive to note just how narrow the foundations of this hypothesis were. The excavations of John Pattison Gibson and F. Gerald Simpson at milecastle 48 (Poltross Burn) in 1909 had led to the creation of a chronology for that site (Gibson and Simpson 1911) and it was this that was tweaked in 1929 as a result of the discovery of the two inscriptions; the new canon, now applied to the whole frontier, was based on two excavations and two inscriptions (cf. Hodgson 2008).

The author of this article imbibed this orthodoxy as an undergraduate at Durham University from 1962 to 1965, when he was taught by Eric Birley. Subsequently, he became a (junior) member of the team reviewing the evidence for the history of Hadrian’s Wall, publishing his first paper on the frontier in 1968 and then, with Brian Dobson, Hadrian’s Wall, the first full history of Hadrian’s Wall rather than a guidebook with some history attached (Breeze and Dobson 1976). Essentially, therefore, over the years from 1962 to 1976, this author’s interpretations of the northern frontiers of Roman Britain were formed and hardened. Through the ensuing decades, new evidence and alternative interpretations were judged against these early views. In 2025, a series of lectures by colleagues on Roman frontiers at the Roman Army School in Durham led to a realization by the author that his views required review and indeed that he had been overzealous in rejecting views which did not coincide with his own. The tipping point was the argument by Matthew Symonds that the additional protection provided to the fortlets in south-west Scotland in relation to those in Wales reflected a more volatile situation on the northern frontier than in Wales (Symonds 2018: 89–91). The fact that the northern example cited was the fortlet at Barburgh Mill excavated by the author in 1971 drove home the message (Breeze 1974).

My first task is to describe the early challenges to the Wall periods hypothesis leading on to a consideration of the publication of Breeze and Dobson’s Hadrian’s Wall in 1976 and then how my own views became solidified resulting in strong reactions to other interpretations. Analysis of my Damascene moment leads onto a review of John Gillam’s re-dating of the history of the northern frontier in the second century and reactions to it before concluding with a note on the value of regularly reviewing our evidence and the interpretations drawn from it.

The first challenge to the orthodoxy, which had been established as a result of the excavation at Birdoswald in 1929, arose from another excavation. As a result of his work at Halton Chesters in the 1950s, Michael Jarrett, a pupil of Eric Birley in Durham, realised that his discoveries did not agree with the approved interpretation of the history of Hadrian’s Wall which he had been taught. However, so strong was the framework of the ‘Wall periods’ hypothesis that Jarrett did not publish his views for another 10 years, and then in Germany in German, out of the shadow of Hadrian’s Wall — and the Durham School (Jarrett 1967). As part of his reaction to the Wall periods hypothesis, Jarrett argued that the history of each site should follow its own evidence rather than be forced into the straitjacket of an overarching theory.

To Michael Jarrett falls the honour of opening the first hole in the dyke. His lead was followed by additional pupils of Birley. The Durham and Newcastle Universities training excavations at Corbridge in the late 1960s gave the staff the opportunities to discuss the chronology of the northern frontier (Breeze 2021). In 1970, Charles Daniels organized the first Roman Northern Frontier Seminar, bringing together archaeologists on Hadrian’s Wall and the Antonine Wall, where different chronologies for the two frontiers ran concurrently: this is hard to believe today, but it was the case. The next stage was that articles started to appear reviewing the evidence (e.g. Gillam and Mann 1970; Jarrett and Mann 1970). They led to John Gillam’s seminal Horsley Memorial Lecture at the Society of Antiquaries of Newcastle in January 1974, where he re-wrote the history of Hadrian’s Wall in the second century (Gillam 1974; 2021 [published posthumously]).

As a Durham graduate, I was brought up to believe in the ‘Wall periods’ and noted that to challenge the hypothesis was to court oblivion. As a supervisor at Corbridge from 1966 to 1969, I more observed than participated in the discussions, but then from 1967 to 1969 I excavated Carrawburgh fort on the Wall and found no indication of any destruction contrary to the ‘school solution’ (Breeze 1972). This led me to a general examination of reports on Wall sites for records of destruction, and how any such evidence might be interpreted (there was a lot of discussion about the existence of bent nails, indicating purposeful demolition, or not, at the time). Eric Birley, the propagator of the 1930 interpretation, himself had noted that in some cases burning might relate to debris from hearths rather than enemy action (Birley 1930: 171). I concluded that the evidence for the destruction of structures along the Wall, be it from enemy action or accident, was very slight; after a struggle to find a journal willing to publish my heresy, my conclusions appeared in print (Breeze in Breeze and Dobson 1972: 200–206). John Gillam, directing excavations at Corbridge, Halton Chesters and Rudchester in the 1950s and ‘60s came to the conclusion that the evidence for destructions at these forts related to an invasion, and that this was the event in the early 180s recorded by Cassius Dio, with the invading army using the Roman road, Dere Street, to penetrate as far south as Hadrian’s Wall (Gillam 1974: 8–10). In writing the report on my excavation at Carrawburgh, I was fortified by Jarrett’s argument that the evidence from each site should be treated on its own merits (Breeze 1972).

The end of this long period of revision came with the publication of Breeze and Dobson’s Hadrian’s Wall in 1976 (the text was completed in 1974, but we had difficulties in finding a publisher: it was rejected by Thames and Hudson and then published by Allen Lane) (Figure 2). In this book, the Wall periods hypothesis was simply ignored (Breeze and Dobson 1976). The writing of the text was divided between the two authors, with myself taking on the chapters on the history of the Wall, two, three and six, though the final text was approved by both authors.

The ability to produce the book together followed from writing booklets and articles on Roman frontiers and the Roman army together over the previous six years, but also through the joint running of an annual week-long study course on Hadrian’s Wall and Hadrian’s Army from 1968. These were fundamentally significant events for both of us. Here, we were faced with producing lectures about Hadrian’s Wall and through that reviewing our own positions on literally every aspect. We were also challenged by our audiences, several of whom went on to publish on Hadrian’s Wall themselves (e.g. Hunneysett 1980; Hill 1981; 2006; Waldock 2002).

It was in these years that I sharpened my own thoughts on the purpose and operation of Hadrian’s Wall. I was heavily influenced by R. G. Collingwood’s 1921 article in which he argued that soldiers did not fight from the top of the Wall because they did not have the appropriate weapons, there were too few soldiers to man the whole line of the Wall, too few access points, and in any case, the Roman army preferred to march out from their forts and fight in the open where they usually won; as a result he envisaged the Wall as an elevated sentry walk (Figure 3). To this was added John Mann’s 1969 observation that, as the palisade serving as the frontier in Germany patently did not have a wall-walk, we should not expect there to be one on Hadrian’s Wall (quoted by Dobson in Breeze and Dobson 1972: 187) (Figure 4). From this position, that the Wall was not built to be fought from, I have not budged, though I accept that the width of the original stone wall, 10 Roman feet at base, may point to an intention to have a wall-walk, abandoned when forts were added to the linear barrier (Breeze 2019: 79).

By 1976, my interpretations were set as firmly as those of Birley, Richmond and Simpson nearly 50 years before, though in mitigation, I would argue that mine had been forged in public debate at adult education courses and argued in detail in a series of papers in archaeological journals as opposed to the ex cathedra statements of my predecessors. My position was determined, set in stone. It was based on a belief in the superiority of the Roman army in relation to its foes in northern Britain, the lack of evidence along the Wall, with the above noted exceptions cited by Gillam, for any destruction by enemy action, and the strong view that the linear barrier — the curtain wall itself — was not defensive, i.e. that it was not manned by soldiers who fought from or patrolled along its top. All other evidence was bent to this view — my interpretation, my prejudice, my bias.

It was through this lens that I interpreted all other evidence over the next five decades, pushing aside evidence and interpretations that did not accord with my views, and defending them in print as the citations in the bibliography demonstrate. The following examples will suffice.

Yes, there are references to trouble in Britain at the beginning of the reigns of Hadrian, Antoninus Pius, Marcus, Commodus and Severus, but most could be explained away as being clichés relating to the unhappy inheritance of each emperor in a manner observable today with political parties (Breeze and Dobson 1976: 109; Breeze 1982: 125).

Yes, there are epigraphic references to military activities beyond the Wall by the Roman army (Breeze 1982: 126), but there is evidence for raiding on every Roman frontier, so the situation in Britain is not special (Breeze 2019b: 188–190).

Yes, the western sector of Hadrian’s Wall was constructed in turf, but that did not imply a need to build at speed because of a threat from the north-west (e.g. Graafstal 2012: 136–138; Hodgson 2017: 53, 67) but rather a practical choice of building material (Breeze 2009: 92; 2019a: 83), a position latterly supported by the argument that the very large stones used in the construction of some milecastle and fort gates indicates building at leisure (Breeze 2019a: 83) (Figure 5).

Yes, there were milefortlets and towers down the Cumbrian coast as part of the Hadrianic scheme, which have been taken to imply a threat to the province from beyond the Solway, but most were not reoccupied on the return to Hadrian’s Wall from the Antonine Wall, suggesting that there may have been a different reason for their construction; perhaps the original plan had been overkill (Breeze and Dobson 1976: 126).

Yes, there was a concentration of fortlets beyond the Wall in south-west Scotland in the Antonine period. I have interpreted this as indicating a concern for the ‘security of communications and protection and regulation of traffic’ (Breeze 1974: 152) or, to put it another way, the presence of a restless local population (Hodgson 2009), but other reasons could be offered. These include the possibility that the fortlets reflected an economic use of manpower (Miller 1952: 219–221) or overkill to ensure the smooth success of the invasion (Breeze 1975: 76). In addition, the existence of fortlets elsewhere in southern Scotland at the same time were not interpreted in the same way, that is, as implying an insecure situation (Breeze 2020: 294). More significantly, Symonds (2018) saw the fortlets in an even wider context (Figure 6). While acknowledging their relationship to the landscape, for him they represent ‘the apogee of early imperial experimentation into highway security in the north-west provinces’ (Symonds 2018: 77).

Yes, pits that were emplacements for above-ground obstacles or entanglements were discovered on the berm in the eastern sector of Hadrian’s Wall, but that did not prove that troops fought from or patrolled its top (Bidwell 2008b). Indeed, the reverse is possible, that there was no wall-walk and so it was necessary to provide protection for the sectors between the towers and milecastles to help prevent illicit crossings (Lindsay Allason-Jones pers. com.) (Figure 7).

Yes, some milecastles on the stone sector of Hadrian’s Wall appear to have been given priority in the building programme, but, in spite of having invented the term ‘priority milecastle’ (Hooley and Breeze 1968: 109), I remained unconvinced that the order of construction related to anything more than the happenstance of the building programme (but see Symonds 2005: 77–78).

Yes, Hadrian’s Wall and the Antonine Wall were very strong in the materials of construction and the existence of a ditch in relation to the German frontier but we argued that the former may have been created as a memorial to Hadrian (Breeze and Dobson 1976: 33), with the latter following suit in its size (Gillam 1975). Hadrian’s interest in architecture may have played a part (Breeze 2009), while the plan suggests that it was formulated far from northern Britain (Breeze and Dobson 1976: 38), possibly by Hadrian (Graafstal 2012: 133; 2018: 87). In considering the differences between the British frontiers and those elsewhere, it was possible to invoke Mann’s (1974: 519) suggestion that provincial governors were allowed leeway in the construction of military works and were not encouraged to work together.

Yes, some hinterland forts had defended annexes (Bidwell and Hodgson 2009), but rather than this being a general phenomenon, most seem to relate to high-status cavalry units, precisely those who could afford the cost of their construction (Breeze 2011a: 129–130).

Yes, the Antonine fort at Birrens was destroyed by fire, like its predecessor, to be rebuilt in AD 158, but the destruction did not have to be the result of enemy action; even the excavator acknowledged that there was no definite evidence (Robertson 1975: 283. cf. Breeze 1977: 459). There is evidence for Roman units demolishing and burning their forts when the regiment in residence was changed or on abandonment (e.g. Gillam 1977: 55–56; Pitts and St. Joseph 1985: 279–280; Hanson 2007: 649; Breeze 2016: 375), just as the Boudiccan revolt provides evidence for the sacking of Roman places (Figure 8). Further, it seemed to me that it would be a remarkable coincidence for a fort to be destroyed by enemy action at the same time that the regiment based there was about to be changed, as was the case argued for Birrens (Robertson 1975: 283).

Yes, there was fierce burning in the South Shields fort about AD 300, when it is believed that the unit in residence was replaced, though here the failure to recover high-value items of gold, military equipment etc. from several of the barracks rather counts against any kind of deliberate fire, even one that got out of hand (Hodgson 2005).

Yes, the excavation of Birdoswald by Tony Wilmott (1997: 403–404) failed to produce evidence for burning, but it did demonstrate a period of abandonment during its construction. Rather than interpreting this pause as resulting from warfare on the northern frontier, I suggested that it could be explained by a shifting of resources elsewhere for a time during the long building programme for the Wall (Breeze 2003: 14).

Yes, there was a large army based in Britain; that might not indicate that the enemies to the north were particularly fierce, but that they had not yet been conquered and so the army remained waiting for the order to advance (Breeze 1988: 8).

In short, I felt that I was able to counter any argument advanced by colleagues for the Roman army operating on the back foot, under pressure from its northern neighbours, or internal unrest, or adopting a defensive posture. My confidence in my interpretations was, in retrospect, strengthened by the success of Hadrian’s Wall (Breeze and Dobson 1976). Not only was this the first history of Hadrian’s Wall, as opposed to a guidebook with some history added, but it attracted — and has continued to attract — good reviews (e.g. Caruana 1976; Daniels 1979; described as ‘revolutionary’ by Hodgson (2009: 1); and ‘fundamental’ by Wheeler (2017: 692)). It rapidly established itself as the new paradigm, to use Kuhn’s (1962) term. In the following years, I sat through many lectures in which the starting point was a reference to the book. However, this is not to say that I was unaware that not all agreed with the theses set out in the book. As a result, when I published the fourteenth edition of the Handbook to the Roman Wall in 2006, I organized a conference at South Shields with Paul Bidwell in the same year with the intention of giving colleagues whose views did not agree with mine the opportunity to challenge them (Bidwell 2008a). Many, though not all, colleagues rose to the challenge. Voices were mostly raised in the margins of the debates not in relation to what I saw as the core issues, not least the purpose of Hadrian’s Wall and how it was intended to operate and, most importantly, how we determine our answer(s) to these questions. This is, of course, a general aspect of life; most of us are more comfortable dealing with the nitty-gritty issues rather than the wider strategic or philosophical questions. It can be exemplified by the development of the Congress of Roman Frontier Studies. Established in 1949, and in its early days dealing with wide-ranging subjects, 30 years later, the great majority of the papers were reports on excavations (Breeze et al. 2024: 157–159). Starting in 1986, there was a move to change from geographical to thematic sessions, resulting in greater in-depth treatment of frontier issues and greater international cooperation.

Also, I can see, with hindsight, that I sat within the cocoon of the Durham approach to the purpose and the operation of Roman frontiers as enunciated by Eric Birley, John Mann and Brian Dobson in particular (Breeze 1982: 12–13). Hadrian’s Wall challenged Eric Birley’s chronology for the Wall, his ‘Wall periods’, but not his general view of its function and operation. Here, the attitude of the Durham people has remained consistently stable and uniform.

There was a different approach taken by colleagues in Newcastle to which I could — and did — react. My Newcastle colleagues saw Hadrian’s Wall as being more defensive, with a wall-walk and soldiers lined up on the top protected not just by a substantial ditch but with obstacles on the berm (Daniels 1979: 360–361; Bidwell 2008c). My reactions to this position strengthened, rather than challenged, my own views and did not lead them to change.

I remain puzzled as to why there is this strong divide. Is it the result of teaching? Perhaps, but Paul Bidwell, who entered the study of Hadrian’s Wall from Exeter and settled on Tyneside, followed the same approach. One answer may be ‘group thinking’ or the ‘corporate’ syndrome. People within a group talk to each other and reinforce each other’s views (see the discussion in King and Crewe 2014 about the decision to impose the poll tax in Britain in the 1980s). The same could be applied to my own situation. Each rejection of an alternative interpretation supported and strengthened my own biases.

My views remained secure until the meeting of the Roman Army School in March 2025 on the subject of Roman frontiers. The lecturers on this occasion were Nick Hodgson on Hadrian’s Wall, Bill Hanson on the Antonine Wall and the Valu lui Traian in Romania, and Matt Symonds on the military landscapes of northern Britain (my contribution was on the frontier in Germany). Of course, I was conversant with their views. Hodgson had contributed a discussion to my Festschrift detailing the evidence for unsettled conditions north of Hadrian’s Wall in the second century, an extension of his general position in emphasizing the defensive nature of Hadrian’s Wall (Hodgson 2009; for a wider discussion of the defensive nature of the military occupation of north Britain see Bidwell and Hodgson 2009: 31–33, 42–45). Hanson focused on the evidence for the building of the Antonine Wall. Here, the divide is on whether all the forts were planned from the beginning or whether some were added later. I had previously written in support of the first argument in Hanson’s Festschrift (Graafstal et al. 2015), but here was Hanson’s counterblast. It carried with it a reason to add forts to the Antonine Wall during its construction, namely, a reaction to that construction from the tribes living beyond the Wall, in short, another ‘defensive’ argument (Hanson 2020).

I also knew the views of Symonds (2015; 2018) through his publications, but I had not thoroughly taken to heart his arguments. Now, here, in his lecture, Symonds presented his evidence for the importance of fortlets in ensuring local security. He examined the differences between the fortlets in Wales and those in southern Scotland, pointing out the special defences of the gates of the fortlets in Scotland which did not appear in Wales, implying a greater need for security in Scotland. This was the tipping point. It is, however, difficult to be sure why. The reason may have been that the example that Symonds chose to illustrate his argument was the fortlet at Barburgh Mill in south-west Scotland which I had excavated in 1971 (Breeze 1974). An ‘abstract’ argument suddenly became personal and my brain automatically adjusted.

The combined presentations of these three colleagues, coupled with knowledge of their previous publications, broke through the dam. It suddenly became impossible to ignore, for example, the clear evidence of Symonds for the stronger defences of the Antonine fortlets in south-west Scotland when compared to those in Wales. The greater density and strength of the fortlets in south-west Scotland had to be acknowledged and therefore the implication of a greater concern for security in that region (Figure 9). The argument of Hanson that the additional forts on the Antonine Wall related to pressure on the frontier was plausible. Hodgson’s discussion of the building of Hadrian’s Wall emphasized that although problems remained about the date and nature of warfare in Britain under Hadrian, it was recorded by several ancient authorities in both literature and epigraphy and it was a possible explanation for the hiatus in the building of Hadrian’s Wall noted at several sites and hence its extended building programme, probably not completed by Hadrian’s death (Breeze 2019a: 90; Graafstal and Breeze 2023).

It is difficult to know whether my period of adjustment will be painful, but after 50 years of seeing the northern frontier in one way and persuading myself that I was right, I must now review my biases, throw them out and start the process of re-thinking. My view will be more nuanced. It is not the major invasions and unrest attested on the northern frontier that need to be reassessed, nor the epigraphic evidence for Roman forays north of the Wall. The main element of reconsideration must be the Romans’ appreciation of the threat level. It was higher than I had wanted to believe. Perhaps the Roman army had not been able to control the peoples of northern Britain at will, to go freely where they wished, as Hodgson correctly characterized my view (Hodgson 2009; cf. Hodgson 1995: 42). Tighter control than I had wanted to accept was necessary. Communication routes needed protection, as can be seen in the fortlets and towers along the roads in the hinterland of the Antonine Wall as I had pointed out fifty years before (Breeze 1974: 151–152). Symonds’ (2018) emphasis on the additional defences of these fortlets suggests that the security situation could have been more serious than elsewhere. The reason for the continuing strength of the army of Britain in the second century, and on into the third, was not just the Roman army leaving troops in Britain because it could not be bothered to move them elsewhere or the need to maintain a strong army in the province because of difficulties in providing urgent support (which, of course, was provided in the fourth century when required) but quite simply because a strong army was necessary to protect the northern frontier (Breeze 1988: 8).

There will be sticking points: In the debate about whether the camps on each side of Burnswark Hill were erected in an actual siege or were a training area for the Roman army, I remain firm in my belief that the latter interpretation is to be preferred (Breeze 2011b; Wilkins 2024) (Figure 10).

The issues surrounding the interpretation of burning in Roman forts is long standing. The interpretations have included the debris of hearths, accidental fires, the demolition of forts by its occupants when they moved out and enemy action (Breeze 2005). At the annual South Shields conference in 2003, a fire officer approached the issue of interpretation by considering the processes and results of fires, including that in the fort at South Shields about 300 (Morley 2005). The reasons for fires in Roman forts could be considered in an even wider context: a recent survey revealed that an average of about seven heritage buildings in the UK were lost to fire each month in the four and a half years from January 2002 to June 2006 (Maxwell 2008: Chapter 3: 6). The chances of accidental fires in Roman forts appear to be higher than we normally assume.

It is important in undertaking this reassessment not to allow the pendulum to swing too far in the opposite direction. My views, originally laid down in print in 1972, were a reaction to the prevailing ‘school solution’ (cf. Daniels 1979: 357). But they were evidence-based and, while I still accept most of the views expressed in Hadrian’s Wall, there are some which have been overturned, which is as it ought to be. In many areas, the debate continues. We are all governed by our underlying views. To take just one example: were the two linear barriers in northern Britain surmounted by a wall-walk? There is no direct evidence. How the question is answered can reflect the underlying beliefs/biases/interpretations of the responder (Breeze 2014b). To put it crudely, to reject the wall-walk is to see the Roman army pre-eminent in the field; to argue in favour of a wall-walk may imply belief in a more defensive posture, or simply an acceptance of a need for greater security in being able to scrutinize the north from any point along the Wall. But what is the dividing line between ‘defence’ and sensible security measures?

On Hadrian’s Wall, there was a move from the construction of a relatively simple, though massive, linear barrier with relatively few troops along its line — the lowest estimate could be less than a thousand — to the addition of a regular series of forts, mostly for whole units, placed, in many cases, astride the frontier (Figure 11). As I see it, this was a move from a ‘passive’ frontier to a situation in which the army appeared to be more concerned about what was happening to the north. But was this precautionary or a reaction to an actual threat, and how do we decide? On the Antonine Wall, was the addition of closely spaced forts a reaction to a threat (Hanson 2020: 218–220) or was it part of a different assessment of the operation of the frontier, abandoning towers in favour of more closely spaced forts (Graafstal 2020: 179; 2024: 15–17)? Interpreting such evidence correctly may not help with understanding the Grand Strategy of the Roman Empire, but it will provide an insight into the measures which the Roman army undertook to protect its empire and the thought processes of the generals.

One might ask why we do not review our interpretations more often. There are probably several answers to this. There is first the intellectual, physical and emotional effort involved in the task. Further, the older we become, the harder it is to challenge our own long-held views. Then there is the anticipated opposition, the counter challenge, and the ridicule, yes, the ridicule.

I have already mentioned that in the 1960s there were different chronologies for the northern frontier in play, one held by the archaeologists based in Scotland and the other by the denizens of Durham and Newcastle. Those colleagues working in Scotland placed the end of the Antonine Wall in the early AD 180s, that date relating to the invasion recorded by Cassius Dio, with a break in its occupation in the years from about AD 155 to 158 and with a brief third period of occupation (Macdonald 1934: 364; Robertson 1960: 37; Steer 1964: 11). Archaeologists working on Hadrian’s Wall, however, saw the Antonine Wall continuing in occupation, after a break, until the end of the second century, that is, to the destruction deposit at Corbridge dated to AD 197 (Gillam 1974: 5–6).

The differences in interpretation came to a head with the publication of Kenneth Steer’s report on his excavations at Mumrills on the Antonine Wall in 1963. In this, the Samian pottery in the destruction deposit was dated by Brian Hartley (1963) to within the decade after AD 150. Kay Hartley (1963) gave the mortaria a terminus post quem of AD 150–155 and possibly 160. John Gillam, however, preferred a date of AD 175–185, basing this on his date for the Corbridge (and Hadrian’s Wall) destruction of AD 197 (Gillam 1963).

The ‘protagonists’ knew each other well and the problem was discussed on several occasion, but it was not until 1971 that they all sat in the same room and thrashed out the issues. Gillam realised that he had to review his position and he re-examined all the pottery from Corbridge — the crucial site because of its position in relation to Hadrian’s Wall and to the occupation of Scotland and the most prolific by way of quantities of pottery — and revised his date for the destruction deposit from AD 197 to c. 180. He thereby brought his date for the coarse pottery at Mumrills into line with those for the Samian and mortaria and accepted Brian Hartley’s preferred date of c. 165 for the abandonment of the Antonine Wall (Gillam 1974: 10; 2021: 100–113). The new dates for the occupations of the two Walls were reinforced by Brian Hartley’s (1972) review of the samian from the northern frontier which indicated that the two Walls were not occupied at the same time.

In January 1974, Charles Daniels organized a conference at Newcastle where the results of the rethinking of Hadrian’s Wall over the last decade were presented. Here, Gillam set down his new dating for the coarse pottery from the two Walls. The ramifications extended well beyond the dating of Mumrills and Corbridge as they were a major challenge to the concept of the Wall periods. I chaired Gillam’s lecture and I can still remember my horror at the intensity of the attacks from the audience. Gillam stood his ground, and in an attempt to persuade the doubters wrote a paper setting out his position (alas, this was not published, posthumously, until 2021). In this, he was not wholly successful. Ten years after the conference, Frere introduced a footnote to his new edition of Britannia in which he stated that ‘the dating was unacceptable because it would leave a typological gap [in the pottery sequence] for the period 180–200’ (Frere 1987: 177, n. 4). Ridicule this may not have been, but it was certainly a slap down by the leading authority of Roman Britain of the man who had done more than any other at that time to create a baseline for the dating of Roman pottery in northern Britain and also, it is important to note, explain his methods (Gillam 1950; 1956). Equally terse was Frere’s reaction to John Mann’s observation that, as there was no wall-walk along the top of the German palisade, there need not have been one on Hadrian’s Wall; in the margin of the paper submitted in 1971 by Brian Dobson and myself to Frere as editor of Britannia, he wrote the single word ‘rubbish’. In such circumstances, to challenge even one’s own previous interpretation was to excite hostile comment.

We are apt to forget that each of our interpretations is but a step along a road towards a new interpretation formed as new evidence accumulates and also as we are influenced by the changing zeitgeist. It would be tedious to list the number of put-downs I received in earlier years, starting with the classic of being described as ‘a young man in a hurry’. And the triumphant cries when new evidence contradicted, or rather was deemed to contradict, my own position. Humility in the face of a correct challenge based on new evidence is not the easiest virtue for humans to espouse but it is necessary in our archaeological world.

The publication of Hadrian’s Wall in 1976 came at a particular point in the re-interpretation of the northern frontier of Roman Britain, the moment when the dust settled (even so, in advance of publication, I was informed by one senior colleague that this was not the right time to write a history of Hadrian’s Wall: I omitted to ask, when would be the right time?). The authors were fortunate in being the first in the field — and, as it happened, the only ones to produce such a synthesis. The book therefore served two purposes, a statement of the current position and a measure against which new ideas could be tested. Three subsequent editions allowed new information and interpretations to be absorbed (a fifth edition is in preparation).

There is one major advantage in reviewing one’s position; this is that forgotten publications, interpretations and arguments are rediscovered and those missed before unearthed for the first time (in re-reading Breeze and Dobson 1976: 55 I discovered that we had cited C. E. Stevens’ suggestion that the Wall was started before Hadrian’s arrival and he modified it by placing the forts on the line of the Wall, only to reject it: now I accept the suggestion: Breeze 2019a: 86; Stevens 1966: 39; cf. Graafstal 2012: 133). Such a review is therefore a rewarding, fulfilling exercise. And, at the end of the day, it is all theoretical. Little can be proved; it is all a balance of probabilities. Even the literary statements have to be treated with care bearing in mind that the ancient writers would have had their own biases, as do we today. The best that we can do is retain open minds, give our colleagues the benefit of the doubt and challenge our own biases, acknowledging that these biases shape the way we handle the evidence. This is one lesson I took from those discussions about the dating of the occupation of the northern frontiers in the 1960s and ‘70s.

There may be a ‘rule’ relating to interpretations, one could call it the ’40-year rule’. John Collingwood Bruce published his first book on Hadrian’s Wall in 1851 and rapidly became the ‘king’ of the Wall, dying 40 years later in 1892. The Wall period theory was promulgated in 1930; Breeze and Dobson’s Hadrian’s Wall appeared in 1976 (completed in 1974). It would appear that we are due a new fundamental treatment of the frontier. But, how ‘fundamental’ would be — could be — such a review? The present view of Hadrian’s Wall is the result of observations, ideas, interpretations and new discoveries over centuries (Breeze 2014a). The body of evidence grows and is regularly challenged, but at the core of the discussion is the purpose and operation of Hadrian’s Wall — and indeed all Roman frontiers — and we await the appearance of the decisive piece of evidence or argument which will convince the participants in the discussion one way or the other: would that be a good or a bad thing?