1 A book on Romanesque architecture may well have over 95 per cent of its text devoted to ecclesiastical building, as for example the volume of that name by H. E. Kubach (London, 1979), and the same point could be made about Focillon’s, Henri much older The Art of the West in the Middle Ages: I Romanesque Art (Oxford, 1969 Google Scholar, first published in French in 1938), a work largely devoted to buildings and their decoration. Similar discrimination afflicts many books on Gothic architecture such as Bony’s, Jean recent French Gothic Architecture of the 12th and 13th centuries (Berkeley, Los Angeles and London, 1983)Google Scholar. These are just sample instances of a widespread reluctance to admit non-ecclesiastical building within the confines of ‘style conscious’ architecture at this period. By contrast, it would be hard to conceive of books on Renaissance architecture omitting palaces. To find any sustained discussion of a castle or domestic building it is generally necessary to turn to the articles, books, even whole journals devoted to these subjects. There are exceptions to this segregationist approach: among writers imbued with the ‘Antiquarian’ tradition (a good example would be Boase, T. S. R., English Art 1100-1216 (Oxford, 1953))Google Scholar, and among those concerned with the socio-economic side of medieval building, particularly trade organization and the supply of materials. But writers such as L. F. Salzman, and Knoop and Jones have little call to concern themselves with individual architectural designs.

2 In particular the late R. A. Brown in various work over the last four decades. For a recent review of the current standing of various aspects of the subject, archaeological and historical, see the article by Eales, R., ‘Royal Power and Castles in Norman England’, in The Ideals and Practice of Medieval Knighthood, III, ed. Harper-Bill, C. (Woodbridge, 1990), pp. 49–78 Google Scholar. Brown’s, comment in ‘Framlingham Castle and Bigod 1154-1216’, Proceedings of the Suffolk Institute of Archaeology, 25 (1952), pp. 127–48 Google Scholar, that ‘English medieval castles have been largely neglected by historians from every point of view, save the architectural’, may have seemed true from his perspective then but is impossible to justify now.

3 In Benson, R. L. and Constable, G., eds, Renaissance and Renewal in the Twelfth Century (Oxford, 1982), p. 255 Google Scholar. He also offers more muted expressions of the same attitude, e.g. on p. 257, ‘This culture was largely a reflection of the knowledge and forms of expression which were being “reborn” in the creative centres of the church’.

4 The documentation on Orford is set out in Colvin, H. M., ed., The History of the King’s Works II (London, 1963), pp. 769-70Google Scholar. Of the total expenditure of £1,413 95. 2d., £986 95. 8d. was spent in the financial years 1165-66 and 1166-67; presumably this is when nearly all the building materials were purchased and building begun. In the next three years, to 1170, outgoings averaged just over £100, probably spent on wages. See also Redstone, V. B. ‘Orford and its Castle’, Proceedings of the Suffolk Institute of Archaeology, 10 (1900), pp. 205–30 Google Scholar, and Hartshorne, C. H., ‘Observations on Orford Castle’, Archaeologia, 29 (read 1840), pp. 60–69 CrossRefGoogle Scholar.

5 Perhaps the most extreme expression of the links between the Normans and castles was made through the intermediary of yet another problematic term, feudalism, by Brown, R. A. in his Origins of English Feudalism (London and New York, 1973)Google Scholar, passim.

6 Eales, ‘Royal Power and Castles . . .’, esp. pp. 50-54.

7 Bibliography on the English castles mentioned in this paper may be found in D. Renn, Norman Castles in Britain (2nd edn, 1973) and King, D.J. C., Castellarium Anglicanum, 2 vols (New York, 1983)Google Scholar. More recent bibliography or that which relates particularly to my arguments will be cited independently.

8 On Castle Acre see Coad, J. G. and Streeten, A. D. F., ‘Excavations at Castle Acre Castle, Norfolk, 1972-7: Country House and Castle of the Norman Earls of Surrey’, Archaeological Journal, 139 (1982), pp. 138–301 CrossRefGoogle Scholar. Harold’s first-floor hall at Bosham is shown in Wilson, D. M., The Bayeux Tapestry (London, 1985), pls 3–4 Google Scholar; there is an earlier literary reference to an upper hall in the ‘miracle’ of St Dunstan at Calne. The building type, complete with exterior staircase, was quite probably of continental origin, as for example seen in the reconstruction of the Emperor Henry III’s palace as Goslar and its antecedents, see Labbé, A., L’Architecture des Palais et des Jardins dans les Chansons de Geste (Paris and Geneva, 1987), pp. 475-79 and pl. 33Google Scholar. Ibid. pp. 500-01 suggests that Bosham might be a Torhalle. Archaeological exploration of late Anglo-Saxon aristocratic sites is still in its infancy. The most helpful for those who support the idea that substantial enclosures were at least an occasional option is Goltho: see Beresford, G., ‘Goltho Manor, Lincolnshire: the buildings and their surrounding defences c. 850-1150’, Anglo-Norman Studies IV, ed. Brown, R. A. (Woodbridge, 1981), pp. 13–36 Google Scholar. In the post-Conquest period both Richmond and Castle Acre belonged to major Norman barons.

9 See Brown, R. A., ‘Some observations on the Tower of London’, Archaeological Journal 136 (1979), pp. 99–108 CrossRefGoogle Scholar, for a reconstruction of the Tower’s original state.

10 A substantial report on the keep at Norwich has been prepared by Paul Drury. It is to be hoped that it will shortly be published. I would like to thank Sue Margeson and Barbara Green at the Castle Museum, Norwich for showing me the typescript.

11 It is not clear that contemporary siege engines were very effective. Jordan Fantosme’s Chronique de la Guerre entre les Angloi et les Ecossois, in which Orford first ‘saw action’, includes a single attempt to use such machinery, at the siege of Wark (lines 1191-284), ‘Hear, lords, of the stonebow how it went on/The first stone which it ever cast at them/ The stone was scarcely parted from the sling/When it knocked one of their [own] knights to the ground/ . . . / Much must he hate the engineer who contrived that for them’. See Chronicles of the Reigns of Stephen, Henry II, and Richard I, ed. Howlett, R., Rolls series, 4 vols (London, 1884-89), III, 202–377 Google Scholar.

12 R. A. Brown, ‘Framlingham Castle and Bigod’, see note 3 above.

13 The earliest evidence of commercial activity at Orford comes from 1102. For this and other references see Beresford, M., New Towns of the Middle Ages: town plantation in England, Wales and Gascony (London, 1967), p. 489 Google Scholar. Stephenson, C., Borough and Town, a study of Urban origins in England, Medieval Academy of America, monograph 7 (Cambridge Mass. 1933), p. 165 Google Scholar discusses the growth of new towns and the east coast ports and gives the comparative taxations in his Appendix iv. Orford first appears in 19 Henry II paying an assise of 20 marks — directly comparable with Yarmouth’s. However Yarmouth had already paid an ‘auxilium’ in 2 Henry II and it is not clear whether the absence of Orford in this year was because it was part of the feudum of Eye or because it was too insignificant, or both. By 1187 its tallage at 31 + marks, was more than Ipswich (surprisingly low at 24 marks) but less than Yarmouth’s (38 + marks).

14 The use of openings specially designed as archery loops is still very restricted during Henry II’s reign, though clear examples are to be found at his castle at Chinon on the Loire where they appear in the wall flanking the main gate of the Chateau de Milieu (the Tour de l’Horloge) and in the mural tower at the opposite end of the site (the Tour du Moulin). The earliest topographical view of Orford Castle, from around 1600, does not indicate the presence of loops in the mural towers. Reproductions of this early watercolour can be found in Colvin, King’s Works, 11, p. 47A, and (on a larger scale and in colour) in the recent English Heritage guidebook by Renn, Derek, Framlingham and Orford Castles (1988), p. 33 Google Scholar.

15 Brown, R. A., English Castles (London, revised edn 1962), p. 52 Google Scholar.

16 O’Neill, B. H. St J., Castles (London, 1954), p. 14 Google Scholar.

17 Curnow, P., ‘Some Developments in Military Architecture c. 1200: Le Coudray-Salbart’, Proceedings of the Battle Conference, II (1979), ed. Brown, R. A., pp. 42–62 Google Scholar.

18 The type of fireplace is distinct from that in the kitchen on the first floor (which is like the double fireplaces in some contemporary monastic kitchens) and it seems that it had some other function. However it is the position of the sink which most clearly suggests that the room was intended to have a particular function, one which required low level drainage.

19 At Henry II’s most expensive castle project, Dover, the keep was provided with piped water: MacPherson, E. and Amos, E. G. J.., ‘The Norman waterworks in the keep at Dover’, Archaeological Journal, 86 (1929), pp. 253–55 Google Scholar. Dover was probably not begun until the 1180s but it is quite possible that the provision of running water in a keep was known by the 1160s.

20 There are relatively few references in contemporary literature to a notion of privacy which involves withdrawing to a separate room, and many of these stem from the segregation of the sexes, for example when a woman changes her clothes or goes to bed.

21 The measurements here are derived from the large-scale cross-section drawings based on a photogrammetic survey conducted for English Heritage. On the use of root 2 in ecclesiastical architecture see Kidson, , Systems of Measurement and Proportion in Early Medieval Architecture 2 vols (unpublished Ph.D. thesis, London, 1956)Google Scholar and various works by Fernie, Eric, for example ‘The Ground Plan of Norwich Cathedral and the Square Root of Two’, Journal of the British Archaeological Association, 129 (1976), pp. 78–86 Google Scholar and ‘Reconstructing Edward’s Abbey at Westminister’, in Romanesque and Gothic, essays for George Zarnecki (Woodbridge, 1987), pp. 63-67 at 64-65.

22 It has not been possible to measure directly the wall thickness at the top of basement level, which I have calculated trigonometrically. This can be done at two points and produces figures of 10 ft 8 in. and 10 ft 9 in. When added to the 27 ft 10 in. of the interior the total is 49 ft 3 in., but there is some margin for error in the calculation.

23 Shelby, L. R., Gothic Design Techniques: the fifteenth-century design booklets of Mathes Roticzer and Hans Schmutter-mayer (Carbondale, 1977), p. 119 Google Scholar. I am grateful to Peter Kidson for pointing me in this direction.

24 It may well be significant that the 11 ft deep tower is also the widest, as though each tower’s depth was derived from its own baseline.

25 Fernie, E., ‘Anglo-Saxon Lengths: the “Northern” System, the Perch and the Foot’, Archaeological Journal, 142 (1985), pp. 246–54 CrossRefGoogle Scholar for divisions of the perch. See also the support given for the use of foot units by Bony, Jean, ‘The Stonework Planning of the First Durham Master’, in Medieval Architecture and its Intellectual Context, studies in honour of Peter Kidson, eds Fernie, E. and Crossley, P. (London and Ronceverte, 1990), pp. 19–34, at 31-32 and note 15Google Scholar.

26 Knowledge of the formula 22 over 7 for the relationship of the circumference of a circle to its diameter can be found in a few sources earlier than Orford. Three of these are printed by Bubnov, N., ed. Gerberti Opera Mathematica (Berlin 1899, reprinted Hildesheim 1963), p. 356 (also pp. 346, 354) from Geometria incerti auctoris probably of the early eleventh centuryGoogle Scholar; p. 546 from the treatise of Epaphroditus and Vitruvius Rufus architectus, perhaps of the sixth century; and p. 304 from the Epistola Adelboldi ad Silvestrum II papam of 999-1003. The first and last of these texts were certainly circulating in England during the twelfth century. See also Practical Geometry in the High Middle Ages: Artis cuiuslibet Conummatio and the Pratike de Geometrie , ed. with trans, and commentary by Victor, S. K. (Philadelphia, 1979), p. 173, n. 64Google Scholar.

27 E.g. Newark Castle, built by Alexander, bishop of Lincoln. The feature occurs in ‘castles’ of all dates, (e.g. Langeais, on the Loire, c. 1000) because it was common in Roman city walls and gates.

28 In Colvin, , King’s Works 1 (1963), pp. 369–71 Google Scholar.

29 Mango, C., The Art of the Byzantine Empire 312-1452 (Englewood Cliffs, 1972), pp. 228-29Google Scholar.

30 L’Orange, H. P., Studies on the iconography of Cosmic Kingship (Oslo, 1953), pp. 19–27 Google Scholar and passim.

31 Grabar, O. ‘From Dome of Heaven to Pleasure Dome’, Journal of the Society of Architectural Historians, 49 (1990). pp. 15–21 CrossRefGoogle Scholar.

32 Lavin, I., ‘The House of the Lord’, Art Bulletin, 44 (1962), pp. 1–28 Google Scholar.

33 Le Voyage de Charlemagne à Jérusalem et à Constantinople, ed. Aebischer, P. (Geneva, 1965), pp. 52–53 Google Scholar. On the whole subject of palace architecture in twelfth-century romance literature see most recently the work of Alain Labbé, cited at note 8 above.

34 Heinermann, T., ‘Zeit und Sinn der Karlsreise’, Zeitschrifi für romanische Philologie, 56 (1936), p. 55offGoogle Scholar.

35 Frankl, P., The Gothic, literary sources and interpretations through eight centuries (Princeton, 1960), p. 167 Google Scholar. The most recent edition of the poem is Gautier d’Arras, Eracle, ed. Lage, Guy Raynaud de (Paris, 1976)Google Scholar.

36 Their fame is due in large part to their inclusion by Viollet-le-Duc, E. in his Military Architecture, trans. Macdermott, M. (London 1879)Google Scholar. More recently see Salch, Ch.-L., Les Plus Beaux Chateaux Forts en France (Strasbourg, 1978), pp. 98–101 Google Scholar (Etampes) and 213-16 (the so-called Tour de César at Provins).

37 The connections between Prester John, the Magi and Barbarossa have been convincingly outlined by Hamilton, B., ‘Prester John and the Three Kings of Cologne’, in Studies in Medieval History presented to R. H. C. Davis, eds Harting, H. Mayr and Moore, R. I. (London and Ronceverte, 1985), pp. 177-91Google Scholar. The land of Prester John was often seen as the place where the apostle Thomas had preached and where he had built a palace for the king. Descriptions of the functions of the rooms in this mythic palace circulated throughout the middle ages (Lavin ‘The House of the Lord’) in the legend of the saint and would have provided further stimulus for a high degree of comfort and functional specificity. On the palace of Prester John, see L’Orange, Studies on the iconography of Cosmic Kingship, pp. 18-19.

38 Ibid., pp. 10, 18-19 and passim; and Grabar, ‘From Dome of Heaven . . .’, especially p. 18, for the poem of c. 1200-20 on an Azerbaijani Palace built by Shida, ‘A master in the work of drawing too, and in surveying famed geometer ... I’ll make a seven domed house ... for the seven days of every week and the seven planets . . .’.

39 It is probable that there were three shorter rafters in addition to the thirteen for which corbels are provided (Figs 3,7). If wooden arches were sprung from the corbels either side of the large window niches, a large rafter might then have risen from each of the three apices to the point of the roof. The arches could have been planked on the underside towards the windows, and the remainder of the ‘domed’ plank ceiling could then have been brought down to corbel level.

40 Cited by Paul Frankl, The Gothic . . ., p. 172. See Veldeke, Heinrich von, Eneasroman, ed. and trans. Kartschoke, D. (Stuttgart, 1986), at p. 528 Google Scholar. Kartschoke, pp. 852-56, considers that the text was begun in the early 1180s.

41 Two of the three royal servants responsible for overseeing expenditure on Orford, Bartholomew de Glanville and Wimar the Chaplain, subsequently became sheriffs, and therefore would have resided at the castle. The third member of the team, Robert de Valeines, was closely related to the Glanvilles by marriage. The remarkable success of this extended family in gaining important royal posts has been explored by Mortimer, R., ‘The family of Ranulf de Glanville’, Bulletin of the Institute of Historical Research, 54 (1981), pp. 1–16 CrossRefGoogle Scholar.

42 On Trim see now McNeill, T. E., ‘The Great Towers of Early Irish Castles’, Anglo-Norman Studies XII, ed. Chibnall, M. (Woodbridge, 1990), pp. 99–117 Google Scholar at 104-09.