Research Articles

Castles and the Children of Alfred

Castles in the air

Do not be sidetracked by this title, which is simply that of the dissertation behind this paper's proposal. My starting point was The Early Norman Castles of the British Isles (Armitage, 1912), a formidable book by a formidable woman. Seen by some as final settlement of the long, acrimonious debate over who built the first English castles, its continuing prominence - more than fifty years after publication it was still described as the only substantial book on Norman castles (Renn, 1968) - is less a monument to its achievement than a testimony to the dearth of research.

My research into the English Castle Debate went beyond Armitage's book, encompassing her sources and notebooks as well as other publications by her and her contemporaries. The dissertation (which deliberately avoided trying to settle the question) exposed frailties in her arguments and evidence and challenged the impartiality of her research.

I do not want to fan the flames of the English Castle Debate, encapsulated by Thompson (1991: vii) as "to an outsider like a dispute between the deaf and the blind". Nor do I intend to get embroiled in issues of 11th century ethnicity. One conclusion of the dissertation was that Armitage's Victorian mindset affected her impartiality. My contention is that her fudging of the evidence to support her conclusions is not unique.

"You find that people have copied from each other and never go back to the original stuff" (Kershaw, 1999)

In any field of research a few pieces of work attain 'tablets of stone' status. Archaeological dates and sequences are routinely reviewed in the light of improved radiocarbon and dendrochronological techniques, and the labels 'earliest', 'oldest' etc. regularly reassigned in the wake of subsequent discoveries. Meanwhile little thought is given to the actual evidence behind some works that are cited (or recited) like some sort of mantra. Such works do get re-examined as preludes to disagreements or pieces intended to take the original conclusions forward. Unfortunately, this consists of reviewing the finished work in isolation, which is analogous to a law court ruling on the basis of only the defence or prosecution case. Inevitably the author has not only selected which evidence is to be used but also chosen how it is to be presented. A comprehensive review requires all the evidence, including the background and mindset of the author.

The Case Study

Ella Who?

Born in 1841, Ella Sofia Armitage (née Bulley) was the first research student at Newnham College, Cambridge on its founding in 1871, at a time when men still argued that further education was harmful to women’s health. In addition to her historical research and lecturing she supported her clergyman husband in his work, and was a member of a number of education committees (Counihan, 1988).

In 1919 she was awarded an MA by Manchester University for her work in archaeology (Newnham College Archive). She was also an Honorary Fellow of the Society of Antiquaries of Scotland, and was one of the first six women offered Fellowship of the Royal Society of Antiquaries, which she declined (Evans, 1956: 389). Following the death of her husband in 1929 she went to live with her daughter. She continued to write articles and review manuscripts for The Antiquary, and correspond with colleagues, until her own death shortly after her 90th birthday (Counihan, 1988).

In the beginning...

In 1884 G.T. Clark published Mediaeval Military Architecture in England, a history of English fortifications. Armitage took great exception to Clark's equating English burhs with castles, and to his attributing English origin to many of the country’s motte and bailey mounds (Armitage, 1912).

She set out to disprove Clark's theories and papers presented to the Scottish Society of Antiquaries (1900), the English Historical Review (1904) and The Antiquary (1906) formed the basis of her book. She was supported by contemporaries such as George Neilson, Goddard Orpen and Horace Round, who also opposed Clark’s findings (Counihan, 1988).

The Truth, not quite the whole Truth and....

Arguments over the objectivity of research are nothing new. In the same year that Armitage’s book was published E.T. Leeds (1912) wrote that too many antiquarians were trying to make the archaeological evidence fit the history instead of doing impartial research.

A stated objective of Armitage was to establish whether the English built castles before the Normans came and if not, what sort of fortifications they did build (Armitage, 1912: 11). This was hindered by her narrow definition of the form and function of a castle. The Committee of Ancient Earthworks had already established categories for the wide range of features identified as castles or fortifications.

"A. Fortresses partly inaccessible, by reason of precipices, cliffs, or water, additionally defended by artificial works.

B. Fortresses on hilltops with artificial defences, following the natural line of the hill; Or, though usually on high ground, less dependent on natural slopes for protection.

C. Rectangular or other simple inclosures, including forts and towns of the Romano-British period.

D. Forts consisting only of a mount with encircling ditch or fosse.

E. Fortified mounts, either artificial or partly natural, with traces of an attached court or bailey, or of two or more such courts.

F. Homestead moats, such as abound in some lowland districts, consisting of simple inclosures formed into artificial islands by water moats.

G. Inclosures, mostly rectangular, partaking of the form of F, but protected by stronger defensive works, ramp arted and fossed, and in some instances provided with outworks.

H. Ancient village sites protected by walls, ramparts, or fosses.

X. Defensive works which fall under none of these headings." (Armitage & Montgomerie, 1912: 1)

Armitage specifically chose Classes D and E as features indicative of castles. The characteristic mounds were generally 20 to 40 feet high, usually artificial but occasionally incorporated natural features (Armitage, 1912). Pre-Norman artificial mounds were attributed to the Ancient Britons or Romans, originally constructed for burials or other rituals (Armitage, 1912).

Armitage saw the motte and bailey as almost a prerequisite of a Norman castle. She cited a 12th century treatise on castle-building, Neckham’s De Utensilibus, which stressed the advantages of mottes as part of castles, to support this view (Armitage, 1912: 86). She did however concede that the motte and bailey may have originated elsewhere and been adopted by the Normans, and that some castles had no mottes. These were seen as the exception, for example when incorporation of a large stone keep into the original design rendered a motte superfluous e.g. Colchester, Pevensey, Chepstow, or the site was already well protected by natural features e.g. Bamburgh, Peak and Tynemouth (Armitage, 1912).

"The question is," said Alice, "whether you can make words mean so many different things."

"The question is," said Humpty Dumpty, "which is to be Master - that’s all." (Carroll, 1871: 196)

Armitage was highly selective in her choice of evidence to support her arguments or refute Clark’s. Her reliance on the semantics of translation and terminology in Bede and Domesday to differentiate between types of fortification was not universally accepted; Chadwick (1905: 334) had already commented on Bede’s use of several different words to describe the same thing. Armitage herself (1912: 98) and later writers, such as Renn (1968), observed that for medieval writers the Latin castellum and Middle English castel seemed interchangeable.

Armitage rejected the Anglo-Saxon Chronicles’ earliest mention of English fortifications, AD547 at Bamburgh, on the grounds that this extract "is merely the interpolation of a 12th-century scribe, and is consequently of no consequence whatever, though there is nothing improbable in the statement, and it is supported by Nennius" (Armitage, 1912: 11). Although much of Nennius’ writing had been discredited (Ingram, 1912: 4), Armitage is arbitrarily disregarding evidence mentioned in more than one source.

However, she was prepared to use another 12th century writer, Florence of Worcester, to support her attack on Clark’s alleged misuse of the Old English word burh (Armitage, 1912: 19). She also used the Chronicles’ account of Cynewulfe and Cyneheard in support of her burh theory (Armitage, 1912: 13; Ingram 1912: 50-51), though the account was actually written at least 145 years after the events described (Mitchell & Robinson, 1968).

Armitage’s flexibility over translation extended to other documents. She decided an illustration in the Prudentius manuscript was not a castle because the English had translated the Latin urbem [sic] as burh (Armitage 1912: 19). She admitted that the use of the word castrum in Domesday "nearly always refers to a castle" (Armitage, 1912: 98) but would not accept that this applied to Arundel. Domesday’s entry for Arundel included features associated with towns (Arundel is mentioned in the Burghal Hidage) so Armitage decided there could not have been a castle and "we must therefore regard it as an instance of the fluctuating meaning which both castrum and castellum had in the 11th century" (Armitage, 1912: 98). Clark (1884: I 195) cited Arundel as "the only one therein [Domesday] specifically mentioned as in existence before the coming in of Duke William". Armitage (1912: 95) was adamant there were only three likely pre-conquest castles; Richard’s Castle, Hereford and Ewias, all built by Normans granted land by Edward the Confessor. Once again, whether to seek evidence of Saxon castles or refute Clark, Arundel should have been examined more thoroughly. Her reliance purely on the semantics of translation fell short of the standards of contemporaneous scientific research.

While she admitted that no evidence of English castles was found during her research, which she described as "negative evidence" for her assertion that the English had not built castles (Armitage, 1905: 11; 1912), she was less than exhaustive in her search for these structures. Her notes at Canterbury (Armitage, Y.A.S. MS521: N9) describe an entrance to the Middle Ward "either built on a Saxon foundation or entirely replaced some other defensive work of an earlier period". Canterbury’s ancient Northgate is said to "look like a Roman or Saxon work" (Armitage, Y.A.S. MS521: N55). The entry for Canterbury in her book mentioned, apart from the Norman castle, only Roman remains. The earthen bank of Northgate had also been reassigned to the post-Norman era (Armitage, 1912). The earthworks outside Canterbury, nothing to do with the Norman castle, were ignored.

She lamented the lack of archaeological excavations, which could have aided her research. However, her entry for Bramber castle (Armitage, 1912: 109) made no reference to earlier excavations there, in particular the 1909 work where her colleague Montgomerie had been involved (Barton & Holden, 1977).

Armitage’s entry for Bamborough [sic] (Armitage, 1912: 101) made only a brief mention of the Norman castle, in spite of a wealth of detail in her notes (Armitage, Y.A.S. MS521). Elsewhere in the book she referred to the site as "the only record which we have of any fortress-building by the invading Saxons" (Armitage, 1912: 12). This is clearly a site that deserved investigation in any search for English castles. Evidence of pre-Norman features was known in Armitage’s time. A pre-Norman well, uncovered in 1770 and visible today, is possibly the one mentioned in an 8th century manuscript and an 18th century history of the castle (Bamburgh Project, 1998). Armitage mentioned the well in her notes, but did not comment on its age (Armitage, Y.A.S. MS521: G13). She accepted there were pre-Norman fortifications but did not search for them.

Pontefract castle, "the key to Yorkshire" according to a letter from Ralph Neville to Henry III (Armitage, 1912: 188), is another example of her approach. She wrote "There is no proof that the hill at Kirkby was fortified before the Conquest", but added "Whether he [the builder of the Norman castle] found the place already defended by earthen banks we do not attempt to decide" (Armitage, 1912: 188). Pontefract was vital for anyone seeking control of this region. If the English had built castles, this location would have been ideal. Armitage’s decision not to look for earlier fortifications is also questionable because "its evidently pre-Norman earthworks" had been mentioned by Clark (1884: II 377).

Her account of Gloucester castle (Armitage, 1912: 156) is quite brief as it had, by her day, already been totally demolished. Her notebooks however give a little more detail (Armitage, Y.A.S. MS521: N74). She mentioned the Abbey Register’s reference to an "Old Castle" and Stukeley’s 1721 account of local stories about a huge, fortified mound. She may have been right to interpret the object of these references as "not a castle" but this interpretation (another example of negative evidence?) requires an explanation. A clue to her ideas about what is or is not actually a castle exists in her views towards the pre-Norman English, especially those of Alfred’s Wessex.

Ella loves Alfred

"It was given to Alfred to raise again the fallen standard of Christendom and civilisation, and to establish an English kingdom on so sound a basis that when, in later centuries, it successfully became the prey of the Dane and the Norman, the English polity survived both conquests. The wisdom, energy and steadfastness of King Alfred and his children and grandchildren were amongst the most important of the many factors which have helped to build up the great empire of Britain" (Armitage, 1912: 13)

As the above passage shows, Armitage saw the Norman Conquest as an aberration; a temporary setback to the inevitable fulfilment of English destiny. The eventual conquest of much of the Danelaw by Alfred’s successors diverted attention from the violence of the earlier absorption of Wessex’s English neighbours. Tenth century annals cited by Wainwright (1975) described not only Mercian resentment at being taken over after Æthelflæda’s (daughter of Alfred) death but also the force required to suppress unrest in Chester in AD924. Wainwright also attributed the alliance of the Welsh princes with Wessex to a choice between the lesser of two evils. Wessex had already taken over the Mercians, with whom the Welsh had long had violent territorial disagreements, and was seen as preferable to the Danes (Wainwright, 1975: 322).

Armitage’s rosy vision of Alfred and his successors, a common view in the 19th century with its romanticised attitude to pre-Conquest England, formed the main thrust of her argument against the existence of English castles. She used the Charter of Worcester, granted by Æthelflæda, to identify the burh as a place where the English could shelter when the Danes attacked: "it was by defending and thus developing the boroughs of England that Alfred and his descendants saved England from the Danes" (Armitage, 1912: 21).

Armitage saw much more than simple defence against the Danes behind this strategy. Alfred and his successors are credited with a strong concern for the long-term future of the kingdom. "By studding the great highways of England with fortified towns, Alfred and his children were not only saving the kernel of the British Empire, they were laying the sure foundations of its future progress in the arts and habits of civilised life" (Armitage, 1912: 30).

The networks of burhs along the border of the Danelaw created by Edward and Aethelflæda in the early 10th century suggest a strategy for both protection and counterattack against the Danes wider than individual fortified settlements.

This view of the farsighted purpose behind the construction of the burhs can equally be applied to the Norman castle-building programme. The primary purpose of the castles was to prevent insurrections by the (foreign) English against the Normans. The castle can therefore be seen as the means for protecting the community (the Normans) inside against the perceived threat (the English) outside. This interpretation (matching her analysis of the purpose behind the English burh) appeared in Armitage’s contribution to the Victoria County History of York (1912) in collaboration with Duncan Montgomerie. Most motte and baileys were noted as lying close to old Roman roads. This was done partly for ease of communication, but Armitage and Montgomerie also noted "they are often as near to one another as to suggest that mutual support was desired, or else that one baron was jealously watching another" (Armitage & Montgomerie, 1912: 19). This argument is missing from her book.

Conclusions

I believe this case study identifies serious flaws in Armitage's arguments, arising from her mindset and the selectivity of the evidence chosen to support her conclusions. She was looking for structures, which she believed the English were psychologically incapable of building. Apart from the motte and bailey issue, she failed to discredit Clark's views on the similarities between castles and burhs. Since disproving Clark was the raisson d’être for her book some of Clark's sites ought to have been investigated more thoroughly.

It is important to recognise that my conclusions could not have been reached if I had only used Armitage's book. It was her notebooks, sources and other writings, as well as the writings of Clark and other contemporaries that highlighted the problems with her work. I believe other works are due a similar level of scrutiny. I am not advocating that every book or article should be derived from first principles; there is no point in re-inventing the wheel. My plea is for an occasional second thought about pieces of work that are cited almost as a matter of routine.

A final point for modern archaeologists concerning these older works came out of my comparisons between Armitage's book and Clark's. In many cases Clark in particular describes the sites in great detail. Whether one agrees or disagrees with their interpretations is beside the point: they are describing these sites as they were between 90 and 150 years ago. These descriptions, rather than their interpretations, of features that have since decayed further or disappeared altogether make them a priceless record of the past for today’s academics.

References

Armitage, E.S. Various Dates 30 Notebooks, Y.A.S. MS521, Archive of Yorkshire Archaeological Society, Leeds.

Armitage, E.S. 1904 The Early Norman Castles of England. In English Historical Review XIX (R.L. Poole, ed.). London, Longmans, Green & Co.

Armitage, E.S. 1905 Alleged Norman Origin of Castles in England. In English Historical Review XX (R.L. Poole, ed.). London, Longmans, Green & Co.

Armitage, E.S. 1912 The Early Norman Castles of the British Isles. London, John Murray.

Armitage, E.S. & Montgomerie, D.H. 1912 Ancient Earthworks. In Victoria County History of York Vol 2:1-73, London, Eyre & Spottiswood.

Bamburgh Project 1998 Archaeological Research Project reported on Internet http://www.users.zetnet.co.uk/bamburghproject

Barton, K.J. & Holden, E.W. 1977 Excavations at Bramber Castle, Sussex, 1966-67. Five Castle Excavations: Reports on the Institute’s Research Project into The Origins of the Castle in England. Archaeological Journal 134:11-79.

Bede A History of the English Church and People. (Trans. L. Sherley-Price, 1955). Harmondsworth, Penguin.

Brown, R.A. 1969 An Historian’s Approach to the Origins of the Castle in England. Archaeological Journal 126:131 - 148

Brown, R.A. 1976 English Castles. London, Book Club Associates.

Carroll, L. 1871 Through the Looking Glass: and what Alice found there. In The Complete Works of Lewis Carroll (1939), London, Nonesuch Press.

Chadwick, H. Munro 1905 Studies on Anglo-Saxon Institutions. Cambridge, University Press.

Clark, G.T., 1884 Mediaeval Military Architecture in England (2 Vols.). London, Woman.

Counihan, J. 1988 The Growth of Castle Studies in England and on the Continent since 1950. Proceedings of the Battle Conference on Anglo-Norman Studies XI (ed. RA Brown). Woodbridge, Boydell Press.

Davies Pryce, T. 1905 Alleged Norman Origins of Castles in England. In English Historical Review XX (R.L. Poole, ed.). London, Longmans, Green & Co.

Evans, J. 1956 A History of the Society of Antiquaries. Oxford, University Press.

Ingram, J. 1912 The Anglo-Saxon Chronicle. London, J.M.Dent & Sons.

Kershaw, I. 1999 Interview in The Guardian, 16th October 1999.

Leeds, E.T. 1912 The Distribution of the Anglo-Saxon Saucer Brooch in relation to the Battle of Bedford AD571. In Archaeologia LXIII:159-202.

Mitchell, B. & Robinson. F.C. 1968 A Guide to Old English (5th edition). Oxford, Blackwell.

Newnham College Archive. Newnham College, Cambridge.

Renn, D.F. 1968 Norman Castles in Britain. London, John Baker.

Savage, A. ed. 1984 The Anglo-Saxon Chronicles. Godalming,Bramley Books.

Thompson, M.W. 1991 The Rise of the Castle. Cambridge, University Press.

Wainwright, F.T. 1975 Scandinavian England. Chichester, Phillimore.

Return to top.

Bob Hamilton graduated from the University of Sheffield in 1999 with a BA in Archaeology and Prehistory. He is currently studying for an MA in Landscape Archaeology, also at Sheffield. The idea for this article (using his undergraduate dissertation as a Case Study) arose from anecdotal reports (not only in archaeology) of academics who had taken liberties with evidence to support their theories. For the purposes of this article he would like to stress that it does not matter if Ella Armitage was right or wrong in her conclusions, but it is her approach that is being criticised. That said, please feel free to give him your views, advice, abuse etc. concerning either castles or distortion of evidence.

Research School of Archaeology & Archaeological Science
University of Sheffield
West Court
2, Mappin Street
SHEFFIELD S1 4DT
U.K.
prp99rh@shef.ac.uk