Print this pageHistorical Text Archive © 1990 - 2013
This etext is copyright © 1989, 1991, 2001 C.R.J. Currie. It may be freely circulated in full in machine-readable form for the purpose of research or private study, provided that:
PREFACE TO THE ILLUMINATED VERSION, 2001
When the text was made accessible thtough the Internet in 1991 by being mounted in Don Mabry's FTP Archive, there was no means of illustrating it. Readers, of course, wanted more, and thanks once more to Don Mabry, who has scanned my original slides and mounted them in an HTML version, most of the illustrations to the original lecture are now included as clickable thumbnails (click on the thumbnail for an enlarged version). In some cases I have substituted a new scan of a drawing. Many of the illustrations were from books which cannt be reproduced here, but the references to the originals have been retained in the footnotes.
I have decided not to revise the text for reasons stated in the original Preface.
Dendrochronology has, however, continued to make progress, and where a date has been
proved wrong, or where a very precise felling date has been obtained within the
approximate range that I quoted, I have added the tree-ring date in square brackets in the
text and a reference to the source in the notes. The main effect of the revisions is to
make some of the examples of Decorated carpentry in the Midlands somewhat later: even in
1989 it was apparent that some aspects of the Decorated styple persisted in the West
Midlands into the 15th century, but, as stated, I wished to stress innovation rather than
conservatism. Most of the English dendro-dates, and references to their detailed
publication, can now be obtained online through the Vernacular Architecture Group's online
database: from http://ads.ahds.ac.uk/catalogue,
click to accept the terms, then on the sidebar select 'special collections' and then
If I were revising the text today, I don't think that I should change much of the argument. Nevertheless I would want to stress the importance of the 12th-century fashion for aisled halls in the development of the English bay system, and to take account of recent discoveries of more Romanesque aisled halls, with curved braces, and possible continuities with Gothic carpentry. I should also wish to allude to Lynn Courtenay's work on the evolution of the hammer-beam roof, and to Richard Harris's article on the grammar of carpentry (which appeared in Vernacular Architecture, vol. xx, 1989), a grammar which I believe took shape as a part of the aestheticizing movement that I discuss in the talk. I would also want to explore in more detail possible consequences of the split in upper-class housing arrangements which appears to have taken place between England and northern France in the 13th century (12th-century aisled halls of English type are now being found in Normandy, where later upper-class houses are always storeyed logis; some evidence has been found to suggest that such halls existed in Brittany too). Finally, I would wish to discuss the development of ornamental facades on timber buildings on the Continent from the 15th century. But these are details compared with the general thrust of the argument.
[Perhaps more significant would be a discussion of the very different development of roofs and timber-framing in Northern England, which I largely dodged in the talk. Ornate roofs of the late Middle Ages are much rarer in the north than the south, and are restricted to a much narrower social spectrum. Until about 1300 or a little later Northern carpentry appears to have followed national trends, with the early use of common-rafter roofs and probably crucks, and appearance of crown-posts and base-crucks towards the end of the period. I have seen wall-framing of c. 1300 outside Durham which is virtually identical with forms of the same date in North Berkshire. Thereafter there is a hiatus until the middle of the 14th century. The high Decorated principal-rafter roof with purlins and windbraces appears to have made no impact outside Cheshire and South Lancashire (and a very limited impact even there). From the middle of the century new forms of roof were developed or adopted within the region, particularly at York. They include the newly identified 'truncated-principal' roof, whose variants have recently been identified in a model study by Mr. Robin Thornes (as yet unpublished), and simple roofs with raking queen posts supporting purlins. King-post roofs are found particularly in the Pennines; they may have originated in the 14th century, but, as noted in the talk, do not become common until after 1450. In addition, Northern wall framing remained in certain respects more conservative than Southern.
These differences, and the hiatus which precedes them, may perhaps be related to a cultural crisis which could be attributable to the effects of war with Scotland and climatic deterioration on the economy, more marked than further south. One consequence of the war, in the border counties rather than in Yorkshire, was that everyone who could afford to do so built defensibly (or at least non-inflammably) in stone rather than in timber, so that the divide between upper and lower-class housing was probably as great as on the Continent.]
It should be noted that several buildings said to be in Berkshire, as they were when erected, have since 1974 been in Oxfordshire. Further changes in English local government organization in the 1990s might well render a correction obsolete, so I have not made it. The selection of many slides of Oxfordshire and Berkshire buildings is for two reasons: first, there is an unusually large number of buildings in that area of c. 1270 to 1350: second, because I have slides of them. The same points could have been made with slides of buildings from many other parts of England.
THE AGE OF CARPENTRY: THE NEW ART AND SOCIETY IN PLANTAGENET ENGLAND
This talk is about the curious incident of the dog in the 1 night-time. The title is of course a parody of that of last year's Age of Chivalry exhibition. The idea was provoked by an extraordinary omission from that otherwise excellent exhibition, an omission which (I subsequently discovered) had been pointed out to the organizers before they made it. The exhibition was intended to illustrate all aspects of Gothic art in England between 1200 and 1400. Of course almost all parts of Europe had their Gothic architecture in stone or brick, their Gothic choir stalls, their Gothic font covers, their Gothic sculpture, their Gothic stained glass, their Gothic metalwork, their Gothic manuscript illumination. The exhibition showed the English forms of those arts. But it gave practically no idea of the art which was almost exclusively English in that period, in which England certainly excelled all other parts of Europe, and which was demonstrably a creation of the period. Moreover it was, within England, the most widespread public and private art, capable of making the greatest visual impact on the largest number of people in their everyday life. As such, it should tell us more than other fine or applied Gothic art about the peculiarities of English society which have been alleged by, for example, Alan Macfarlane.
I refer to Gothic architecture in wood--Gothic structural carpentry. I don't mean merely carpentry of the Gothic period - which of course existed everywhere - but the deliberate adaptation of the design of buildings and roofs to the requirements of Gothic style and Gothic ornament. By 1400 English carpenters routinely applied Gothic arcuation and Gothic ornament to the roofs and walls of quite humble buildings such as barns and peasant farmhouses, and to the facades of urban terraces, whose Continental counterparts were by comparison quite plain and functional. This contrast with the Continent was apparent to English writers such as William Harrison in the 16th century. It has also been noticed by continental architectural historians such as Viollet-le-Duc and Friedrich Ostendorf. Ostendorf observes in his monumental Geschichte des Dachwerks, published in 1908, that 'in no country has visible or open roof-carpentry played such a significant role as in England'.
For the purposes of this talk I'll restrict myself as far as possible to the period covered by the exhibition - i.e. down to 1400. This is justifiable in that practically all the main types of roof design and wall framing had been invented by then. The main trend between 1400 and 1550 was a consolidation and standardization of regional differences, although of course some highly elaborate roofs were built over churches and royal palaces. Most of the examples I'll show are of what might be termed upper middle-class buildings: ordinary manor houses and the roofs of parish churches, although there will be some grander structures and a few peasant houses. I'll begin by talking about roofs, since there is much more evidence about them than about other parts of the building. If there's time, I'll say a little at the end about wall framing and the design of timber-framed facades.
I'll try to keep technical jargon to a minimum, but some is unavoidable. The handouts explain most of it. By convention most drawings show cross-sections and it's often difficult to visualize them in three dimensions. So I thought it helpful to show a couple of three-dimensional views to emphasize a key distinction in roof-types.
The upper drawing shows a double-framed roof, in which the weight of the roof is transmitted to the walls at intervals by trusses with principal rafters. The common rafters are supported, and lengthwise rigidity is provided, by purlins and sometimes a ridgepiece, supported by the principal rafters. The lower drawing shows two kinds of single-framed roof in which there are no principal trusses or purlins and each pair of rafters is supported directly by the walls.
The other handout shows types of joint used in the 12th and 13th centuries. The main distinction is that between the various types of halved joint (nos. 1 to 4), with which one timber is lapped onto another, and the mortice and tenon. The latter gives a tidier appearance but logically requires a different method of assembly, since the end of one timber must be inserted into another. Halved joints were widely used in the Romanesque period, but by 1300 in England and northern France they were usually restricted to a few positions in the structure; frames were predominantly mortised. In parts of Germany that change happened much later.
The first two slides show the contrast we've already mentioned. Anyone with a good knowledge of European Gothic styles, but totally ignorant of timber-framing, could identify the country and approximate date of the first example. The Tudor arches across the building, repeated along the roof-plane, give it away. It's clearly English, in the late Perpendicular style. It is in fact the Long Gallery at Abingdon, built in the early 16th century. 2
It's worth noting some of the techniques which help to create these effects. Naturally curved timbers are used to form the arches, and are cut through and through with a saw to make planks and boards, giving depth to the patterns without using a large number of trees. Moreover, because two or more braces are cut from the same timber, they can be arranged to achieve a fairly regular symmetry. The regularity is enhanced by applying chamfers and mouldings to the edges of the timbers to give them a tidy appearance, and by using mortised joints.
The next slide shows a much grander building, but one would need to know a good deal about timberwork to place and date it. It's in 3 fact a French market hall built in 1511, perhaps a few years earlier than the last one. There's nothing particularly Gothic about it. The timbers are largely straight, there's no use of natural curvature to achieve aesthetic effect, there's no arcading in the plane of the roof, and the timbers are square in section, each probably made from a whole tree-trunk, and not tidied up with chamfers. They are probably sawn, but they could equally well be dressed with an axe or adze. If I had shown a Bavarian building of the same period, the contrast with the English one would be even more marked. Yet if we had looked at an early 13th-century English timber building, we would find it had the characteristics we noted in the 16th-century French market hall--or even that it was more like the Bavarian building. How and when did the change come about?
Early roofs, not only in England, but all over Europe, fell into two categories. Some were not meant to be seen. The others were not meant to be looked at. So if the next few roofs seem boring, please remember that they were not meant to be anything else. We'll begin by going up to mount Sinai and taking our laws from Justinian. The roof that he paid for at St. Catherine's monastery there(slide) is a typical Classical king-post roof, with tie-beams, principal rafters, and purlins, like the one in the handout. Such roofs continued to be put up in Southern Europe, particularly in Italy, with little or no change in design, right through the Middle Ages.
In Northern Europe, on the other hand, the earliest roofs which survive or have left clear evidence, those over Romanesque churches, were single-framed common-rafter roofs. Each pair of rafters was jointed to a heavy tie-beam which helped to support a flat ceiling, as 4 at Odda's Chapel, Deerhurst, Glos. (slide). The roofs above such ceilings were not visible, served only to keep out the weather and hold up the ceiling, and were made of crudely squared timbers put together with halved joints. There was a system of struts between tie-beams and rafters, and generally a collar about two-thirds of the way 5 up. (slide). The earliest such roofs had a pitch of 30 degrees or less, so the structure was bottom-heavy and quite stable. These designs were revived for English houses in the 1960s.
In the 12th century, several external pressures forced carpenters to modify this system. First, masons started to build vaults whose crowns rose above the eaves; the roofs above had therefore to dispense with tie-beams. Secondly, from the late 10th century onwards but particularly from the mid 12th, masons wanted steeper-pitched facades, so the roof-pitch had to be increased too.
Thirdly, the new thin Gothic walls required a light roof, as Villard de Honnecourt points out in his notebook. Finally, the steady increase in population and consequent pressure on timber supplies also encouraged a reduction in the number of tie-beams, which were made from larger and thus rarer trees than the other members.
The succession of roofs at the collegiate church of Soignies6 in Hainaut (slide) is a well known example of the introduction of intermediate trusses without tiebeams. At the nave of Kempley in 7 Gloucestershire (slide), not later than about 1140, the carpenter seems to have got rid of the tie-beams altogether, producing the first surviving evidence of an open roof in England and perhaps in Northern Europe. After a period of experiment the new designs were rationalized. During the 12th century English carpenters introduced the soulace, a brace between rafter and collar, to produce a polygonal 8 arch, as at Canon Pyon (Herefs.) (slide). An alternative, derived 9 from timber towers like Brookland in Kent (slide), was scissor-bracing, as in the mid-13th century roofs of the Blackfriars at Gloucester10 (slide). Such roofs could be left open or ceiled and boarded, as at Ely cathedral. Nevertheless, the open roof had appeared before any attempt to make it look nice.
Getting rid of the tie-beams and increasing the pitch made the roof more top-heavy and created a new problem: the rafters started to rack, that is, to lean in a lengthwise direction. Various devices were introduced to stop this. One of the earliest was rafter-bracing, which the sharp-eyed will have spotted in the Canon Pyon roof and which had a long later history in Germany. That too was revived in the 1960s.
The other methods all led to the revival of some form of double-framing. In parts of northern France, purlin roofs seem to have been in use by 1300, at least in barns and houses. In England they appear first in lean-to roofs. In the triforium roofs at Wells Cathedral, perhaps as early as the 1190s, the carpenter introduced a clasped purlin between the rafters and collars, shown here in Hewett's drawing 11(slide). This technique had a long later history, in England. Purlins also occur in a mid 13th-century lean-to roof at Salisbury, trapped between two rafters and a blocking-piece 12in a manner found in early principal-rafter roofs in France. (slide) An alternative was a roof-plate set square, as at the Wheat Barn, Cressing Temple, Essex,13 of the later 13th century (slide). All these devices allowed the carpenter, if he wished, to omit the collars at most of the rafter-couples and thus to save timber or labour.
If the collars were retained, another alternative to purlins, and also of French origin, was the "sous-faitage", the crown-plate or collar-purlin, which runs under the middle of the collars and stops them, with luck, from slipping. Its is stabilized at intervals by a post or strut rising from a tie-beam; the king-strut roof of the nave at Grosmont (Gwent) (slide) is one of the earliest British examples, perhaps of 1240. In England it became more usual to end the post at the collar, when it is called a crown-post in modern jargon.14 (slide)
Accounting documents show that by the early 13th century English carpenters had a more radical alternative to the common-rafter roof in15 the form of the cruck truss (slide), at first apparently restricted to utilitarian structures, particularly farm buildings. Pairs of curved timbers, called crucks or furcae, support both the roof and wall frame. For reasons which are not clear, but I suspect were economic, crucks are not found in East Anglia and the south-east. 16(map). It's worth noting that crucks are made from forked trees which could not be used in the roofs we've looked at so far; on the other hand, cruck buildings need fewer standard woodland trees. A modified form of cruck, perhaps also subsumed under the furcae of the documents, is the base-cruck, which was used to replace the wall-posts and aisle-posts of aisled structures.17 (slide) Base-crucks are found in the South-east, and in the late 13th century were taken up in Flanders and Holland, but they're not found in East Anglia.
We can sum up the story so far by saying that until about 1270 technical change was driven, directly or indirectly, by economic factors or by aesthetic change in masonry work, not by the aesthetics of timber-framing itself. All the roofs we've looked at have been purely functional, and betray no Romanesque or Gothic design: from an artistic point of view, they could date from any period.
To understand the earliest form of ornamental or architectural roof design, we need to go back a century or so and look at the palaces of the upper aristocracy in 12th-century England. In the second half of the century what is now thought to be a new fashion for aisled and clerestoried halls, imitating Romanesque churches, appeared. The first known was at Leicester Castle in the mid-12th century, but the best surviving example, built about 1175 , is at the Bishop's Palace at Hereford.18 (slide) It has timber arcades with scalloped capitals and semicircular arches with mouldings and nailhead ornament. There were large transverse arches as well. We don't know how it was roofed. Trees don't grow in quadrants, and such huge quadrant-shaped timbers must have been very hard to obtain and such buildings therefore rare.
There is now some evidence of similar structures in northern France, and the techniques used in them may have inspired French carpenters to invent the wagon or cradle roof, a modification of the earlier English soulaced roof. The first known example is again domestic, in the great hall of Blois castle about 1200. The slide (slide) shows I think the dorter at the abbey of Noirlac19 on the Cher, about 1250. By the 1230s the technique had spread as far south-east as Lake Constance, and as far north as Lincoln, as Hewett's drawing of the Greyfriars shows.20 (slide) The curved braces were repeated at every rafter couple, and had to be carefully selected to provide a regular lower surface, which was often boarded or plastered. The design was therefore costly in timber.
It was also labour-intensive, particularly after the switch to mortised joints in the later 13th century, since many long mortices had to be cut. Nevertheless the wagon roof forms the basis of practically all ornamental roofs on the Continent. The wagon vault could conceal much structural variation above it. In France it was often combined with king-post trusses, and a simplified form also occurs with curved braces only at the trusses. Since wagon roofs were expensive they were not suitable for peasant houses or even small manor-houses, and were largely restricted to churches, cloisters, and upper-class buildings. They were nevertheless much cheaper than a stone vault, and perhaps that's why they are quite common in English churches. In21 the 14th century English carpenters modified the form, (slide) thickening up some of the rafters to form principals, and adding at 22 first one purlin below, and later several (slide). The effect is that of a vault with moulded ribs and bosses at the intersections, which were often painted. After 1400 this design became very common in churches in the South-West. It was re-exported to Brittany, whence a modified form with decorative king-posts and tiebeams spread along the Channel to Belgium. It's worth mentioning at this point that from the late 13th century onwards some English churches had true vaults in timber, the earliest being apparently at St. Albans and the best known at the Octagon at Ely and at York Minster.
From the point of view of vernacular carpentry the wagon roof was a dead end: it was too expensive. The wholly English alternatives have their origins in domestic architecture and reflect a fundamental feature of English houses, throughout the Middle Ages and at all social levels. By the 14th century the Continental upper classes, including the urban bourgeoisie, lived in houses of two or more storeys with chimneys, often, indeed, towers. There was thus a sharp division between them and the single-storeyed peasant house with its open hearth or, in parts of central Europe, stube and storage loft.
By contrast the English house was essentially bungaloid. It focused round a single-storeyed hall, open to the roof and heated by an unenclosed hearth, at first in the centre and later at one end. The other parts of the house might be two-storeyed, but even the great first-floor chambers of the rich were open to the roof. As a result English carpenters developed systems which could be modified not only for houses of different classes, but also for other single-storeyed buildings, such as churches and barns.
The Romanesque style of the great 12th-century timber halls seems to have been a dead end and had little effect on smaller houses. Although 13th-century manor-houses and even peasant houses, like this 23 one of c. 1250 (slide) were often aisled, their carpentry was not ornamental. There might be elaborate carved capitals below the springing of the braces, but everything above that was built with 24 straight, undecorated, roughly squared timbers, not at all Gothic.(slide).
In the last third of the century there are signs of a change. The manor-house at Bourn in Cambridgeshire, rebuilt after a fire in 1266, has arched braces to the tie-beam, as has a late 13th-century one at Sullington in Sussex25 (slide). At first the roof was unaffected. Even in the 1290s the upper roof of the hall at St. Mary's Hospital, 26 Chichester, has completely plain crown-posts (slide). In the west, early base-cruck halls, like the Old Deanery, Salisbury, built between 1258 and 1274 27(slide), though structurally quite impressive, at first lack any decorative arcuation. At York Farm, West Hagbourne, Berks., a freeholder's hall, perhaps as early as 1280,[1284/5] 28(slide) there is a splendid double-chamfered arch to the main truss but the crown-post roof above is quite rough. There is no ornamentation in the longitudinal plane either. At West Bromwich Manor-House 29(slide), about 1275, the upper roof is still plain, but the arcade plate is chamfered and the braces to it are curved. These arch-braced trusses were soon adopted even in peasant houses, like these two in 30 Berkshire (slide), one apparently of the 1290s, the other of 1305.
A more decorative treatment of the upper roof began about 1280. It seems to start in first-floor chambers, where the roof was closer to the floor than in an aisled hall, and was coming to be lit by 31 larger windows than before. The rectory house at Sutton Courtenay (slide), where this roof was built probably for Solomon of Rochester between 1284 and 1290, though the present windows are later, illustrates the change. As Eric Mercer has pointed out, the crown posts of these roofs could be treated as columns, with capitals and bases, and curved four-way braces mimicking vault-ribs. There may also be a deeper, if probably unconscious, symbolism (slide of crucifix). Crown-posts of this type were certainly soon used in 32 churches; that in the chancel at Harwell, Berks. (slide) can be closely dated to the early 14th century by the arms of Piers Gaveston in the east window.
The great hall at Sutton Courtenay, also probably 1284-1290, 33 (slide) and originally timber-framed, brings all these features together for the first time in a surviving building. We have the great moulded arch, treated as a chancel arch with an applied hood-mould; the octagonal crown-post above; the polygonal arch formed by the soulaces; the moulded cornice and curved braces of the arcade; and at the rear the arch of the spere truss, formed by two planks cut from the same tree, an innovatory device.
Again, these new ideas spread very quickly to the houses of those 34 peasants who could afford them. (slide) Lime Tree House at Harwell, which we looked at briefly just now, was built about 1250 on a 21-acre holding and reconstructed about 1300 in the new Gothic style, with chamfered base-cruck arch, curved arcade braces, and ornamental crown-post above. There is a striking contrast between the finish of the old and the new timberwork.
By the 1290s broadly similar treatment was applied to aisled buildings in Eastern England, where the base-cruck arch was not used. 35 The (slide) shows Hewett's drawing of Place House at Ware. Among early variants of the crown-post technique is the 'raised-aisled' roof, of which the earliest surviving example, of winter 1299-1300, is in the old Warden's Hall at Merton College, Oxford (slide). It's now restored and more visible. Crown-posts could also be combined with a wagon-roof to give the effect of a two-light window: an example at South Moreton, Berks., probably built in the late 1320s 36 (slide) imitates one of 1318 at Chiselhampton, Oxon., and anticipates those built for the earl of Huntingdon at Maxstoke Castle, 37 Warwickshire, about 1340 (slide). The crown-post, often ornamented, over an arch-braced tiebeam, remained typical of the interiors of hundreds of ordinary farmhouses in South East England until the 16th century : that of c. 1400 at Bayleaf, Sussex, reconstructed in the Weald and Downland Museum, shows what they originally looked like38 (slide).
By then, however, in Western England the crown-post was obsolescent. Carpenters there had long since moved on to purlin roofs which allowed ornamental arcading, appropriate to the high Decorated and Perpendicular styles, to be carried on to the plane of the roof itself and sometimes right up to the apex. At Stokesay Castle, Shropshire, (slide), the cruck trusses of the hall roof, finished by about 1305,39 have three tiers of arches right up to the apex. By the middle of the 14th century that was unexceptional, and is even found in barns, as at Bradford on Avon. The manor house at Sutton 40 Courtenay, (slide) probably of about 1340, has a base-cruck with two tiers of cusped arches, as well as curved windbraces in the roof plane. A more spectacular raised-aisled building at Lewknor, Oxon., 41(slide) of about the same date, has curved and cusped timbers in three planes, producing astonishing spatial effects. These ideas too soon affected peasant houses (slide); Tudor House at Steventon, Berkshire, with its ogee base-cruck arch and cinquefoil above, is a customary 42 tenant's house made from timbers felled apparently in 1355.
These base-cruck and aisled designs were obviously not suited to church roofs. Yet by the middle of the 14th century, and probably in the 1320s, West Midland carpenters had developed a new technique, the arch-braced collar-beam roof, which combined French principal rafter construction with English windbracing.43 (slide) Perhaps the earliest one, and certainly one of the most exciting, is that formerly over the Guesten Hall at Worcester, traditionally dated to the 1320s but perhaps of the 1350s. This form persisted until the 16th century; a well known one traditionally ascribed to the 14th century is at44 Colwall church, Herefs. (slide), but it could be early Tudor. Such trusses could alternate with ornamental tie-beam trusses, as at the former guesten hall of about 1340, still with typically Decorated45 reticulated tracery in the windows, at Great Malvern Priory (slide).
Decorated tie-beam roofs were not restricted to the West Midlands: the 46 service wing at Fyfield Manor in Berkshire, (slide) now Blackwell's antiquarian bookshop, is of much the same date. The roofs we've just been looking at have a good deal of cusped timberwork, reminiscent of masonry canopies, tomb recesses, and so on of the Decorated period. Masons had of course introduced cusping in the traceried windows of the late 13th century, and some carpenters were quick to follow. Among the earliest manifestations of cusped timberwork are bargeboards, like those of the late 13th century [felling date 1323]on the South wing of Middle Farm at Harwell, 47 (slide)on the left of the slide; the North wing 48(slide), usually attributed to the 1350s but perhaps of 1316,[felling dates 1367, 1371] still has Geometric barge-boards, but by the mid 14th-century ogee heads were introduced like those recently restored 49 (slide) on the barn at Leigh Court, Worcs. The French equivalent was the trefoiled flying rafter.
Structural cusping was also attempted on the base-cruck arch at 50 Eastington Hall (Worcs.), (slide if poss.) perhaps c. 1280, and the central arch at Southchurch Hall in Essex, allegedly early 14th-century.[Felling date 1321 x 1363] 51 (slide). A more dramatic idea, apparently invented at the same time, was the hammerbeam truss. This produced a massive trefoiled arch;52 (slide) at the Bishop's kitchen in Chichester, late 13th-century, a pattern of four trefoiled arches framed a square. The same idea was used apparently in the 1290s in the so-called Pilgrims' Hall, 53 Winchester. (slide). This was decorative, not structural, in intent, since base-crucks were used further along the building. Carved portrait heads form the cusps of the trefoil. On the Continent, Dutch and Flemish carpenters achieved a trefoiled barrel-vault by combining the wagon roof with base-crucks, as at the Bijloke in Gent about 1300 [c. 1250] 54(slide).
In churches, the late Decorated and still more the Perpendicular style, with buildings of taller proportions, clerestoreys, and parapets, required a low-pitched roof,which emphasized the contrast of horizontal and vertical lines. Hewett claims that the low-pitched king-post roof 55 at Bristol Cathedral (slide) is original, early 14th-century, and much earlier than any other known example. However that may be, around the middle of the century carpenters developed king-post roofs, at first with elaborate cusping, like the Porch at 56 Wellington (Herefs.) (slide), or the nave roofs at Charlton on Otmoor and Adderbury (Oxon) (slide). The pitch was gradually reduced 57 by shortening the king-post, (slide) or even getting rid of it 58 altogether (slide - Weobley). The rafters framed boarded panels and were often enriched with chamfers and mouldings. In the North of England a simpler form of king-post roof was used in houses and barns in the 15th and early 16th centuries.
The Age of Chivalry exhibition did show a picture of Hugh Herland's roof of the 1390s at Westminster Hall, simply as something to gawp at. (slide). I hope that it's now clear that it didn't suddenly emerge like Athena from the head of Zeus, but was the culmination and synthesis of the developments we've been looking at. Combining massive arch-braces with hammerbeam construction provided an elegant trefoiled arch, framing the windows and enriched with mouldings. The angels form the cusps. The head of the arch is already four-centred, as are the arcade braces. There is a central king-post, and the plane of the upper roof is a windbraced arcade. The tracery panels in the spandrels of the arches and above the tiebeams seem to be a new idea. It was widely imitated in the late 15th and earlier 16th century.
Now a brief word about wall-framing. Wall-framing rarely survives from before the 15th century, and where it does it's usually been much altered. The best corpus of earlier wall-framing in a small area is in North Berkshire, from which some of my examples are drawn; the forms can be paralleled elsewhere. Aisled buildings of the 13th and early 14th centuries gave little scope for fancy framing on the sides, since most of the elevation was 59 taken up by the roof. (slide) There were, however, two-storeyed buildings, for example chambers and gatehouses, which presented more opportunity for decoration. At first wall panels seem to have been large plain rectangles with straight braces concealed behind them 60(slide). By the early 14th century the braces were exposed (slide). 61
From the late 13th century carpenters began to use curved braces 62 to achieve a Gothic effect, (slide) as in the south wing of Middle Farm, Harwell, [D 1323] where in the west bay they form an arch over the former staircase entry. Later the braces became more plank-like and shorter; buildings at Weobley in Herefordshire, probably late 14th or early 15th century, show the effects that could be achieved, though spoilt 63 by later subdivision (slides). In the early or mid 14th century[felling date 1391--2] a64 reticulated effect, analogous with contemporary window tracery, (slide) appears at the gatehouse at Mavesyn Ridware in Staffordshire.
From the 1330s- -the first instances seem to be at York--convex braces were
sometimes used; combined with arch-braces they could form 65
a scallop pattern as in an example at Coventry (slide), which again 66
echoes window tracery (slide). Particularly in the North-West there 67
was a fashion for large planks framing cusped panels; (slide) it already appears at
Baguley Hall (Ches.) in the 2nd quarter of the 14th century.
It remained popular in that area in the 16th, for example at Gawsworth (Ches.)
(slide),68 where it's combined with a herringbone pattern.
The hall at Great Malvern Priory69
(slide), however, of c. 1340, like Zacharias's shop in Ship Street, Oxford, built
as an inn in the 1380s (slide)70,
has the downward bracing; it's not clear how we're supposed to read this pattern, though
it may be influenced by Perpendicular traceries in which verticals intersect or surmount 71 arches. In the Wealden house in the south-east (slide) this type
of bracing in the wings alternates with arch-bracing under the eaves in front of the
recessed hall, giving a striking plasticity to the facade. With the triumph of the
Perpendicular style close-studding 72 came
into fashion; (slide) this may be an import from Normandy, but the close studding at
Mancetter Manor in Warwickshire is allegedly mid 14th century.
During this talk I've deliberately stressed innovation. I haven't emphasized- -as most writers on the subject have done- -the persistence of old forms alongside new ones, or the marked regional contrasts in types of roofs and framing, which are very obvious from the 15th century onwards but certainly began before that. I think this is justified. The emphasis on regional differences arises partly because far more houses and barns survive from the late Middle Ages than from before 1350, and this may be misleading. In all parts of the country change in the 15th and early 16th centuries was much less marked than in the previous century and a half. Carpenters went on, with minor alterations, building what they had been putting up before 1400, if not before 1350. This is probably an aspect of the conservatism which affected all forms of architecture in England, and indeed in most of Northern Europe, from some time after the Black Death. In England the Perpendicular style, already appearing by the 1330s, was still dominant at the Reformation, and has never been altogether abandoned since. Conversely, most of the artistic innovations in carpentry we've been discussing seem to have occurred between 1270 and 1350, precisely the period when English masons, too, were the most innovative in Europe.
For timber buildings, late-medieval conservatism has a corollary: buildings are difficult to date. This difficulty is often extrapolated back to the earlier period, so that a building with Decorated mouldings is attributed to the 15th century simply because it is thought to be vernacular. Recent work on tree-ring dating has tended to show that on the one hand, archaic-looking buildings, especially barns, at one time attributed to the late 12th or mid 13th century are often several decades later, while on the other hand, more advanced types attributed to the late 14th or the 15th century are several decades earlier. The traditional approach underestimated the rate of change in the peak period.
We saw that the open roof developed in England in the 12th century, long before it became Gothic, and that several of the technical devices which were used to achieve aesthetic effects between 1270 and 1350 had been introduced somewhat earlier, for different reasons. So the contrast between English and Continental carpentry cannot be explained in technical terms.
And here one must return to the social implications of carpentry. The early appearance of the open roof in churches may simply reflect the relative poverty of a country which couldn't afford much stone vaulting. Yet by the 15th century the choice of a timber roof rather than a vault was clearly a conscious preference. Some of the roof-types we've looked at - the wagon roof and the hammerbeam - were intrinsically expensive or sophisticated and therefore largely restricted to churches and upper-class buildings. But most of the designs could be, and were, adapted to churches and great houses on the one hand, or peasant houses on the other, simply by reducing or increasing the degree of complexity and ornamentation - which was also varied, and this is a point we haven't had time to consider, depending on the importance of the position of the roof or roof-truss in the building. Moreover, domestic, rather than ecclesiastical, architecture seems to have been the motor of change. And new forms appeared in peasant houses very soon after they appeared in great houses. That is now clearly established by dendrochronology.
On the Continent, on the other hand, there was one dominant form of Gothic roof- -the wagon roof- -which was restricted to upper-class buildings; peasant houses, where they survive, don't have ornamental roofs. This contrast surely reflects the absence in England of a sharp distinction between noblesse and peasantry. It suggests a relatively fine stratification of the social hierarchy, allowing quite subtle ways of displaying wealth and status in building. It also indicates an extensive market for professional carpentry, in which peasant clients were active participants. It must also reflect the ubiquity of the open hall at all social levels. Possibly, too the prestige of English carpentry was an important factor in preserving the open hall in the houses of the nobility and gentry when their continental counterparts had often abandoned it. When the open hall went out of fashion in the 16th century, it was the middle classes, not the aristocracy, who led the way.
NOTES AND SOURCES FOR SLIDES
1 i.e. 1988.
2 Slide of Abingdon Abbey, Long Gallery (early 16th century).
3 Mereville, Seine et Oise: Scientific Methods in Medieval Archaeology, ed. R. Berger (170), p. 34.
4 C.R.J. Currie, 'A Romanesque Roof at Odda's Chapel, Deerhurst, Glos?' Antiq. Jnl. lxiii (1,1983), plate xiv a, facing p. 61.
5 Ibid. fig. on p. 59.
6 J.M. Fletcher and P.S. Spokes, ' The Origin and Development of Crown-Post Roofs', Medieval Archaeology, vi (1964), p. 151, fig. 43.
7 Antiq. Jnl. lxv (1985), p. 102, fig. 1.
8 Canon Pyon church, North Aisle; cf. Vernacular Architecture, xxi (1990), 23--5.
9 Archaeologia Cantiana, lxxxix (1974), p. 44.
10 Med. Archaeol. xxii (1978), p. 108, fig. 2. [For tree-ring date (1230 x 1269), Vernacular Architecture, xxiv. 47.]
11 C.A. Hewett, Eng. Cathedral and Monastic Carpentry, p. 85, fig. 81.
12 Hewett, ibid. p. 86, fig. 82.
13 Hewett, Eng. Historic Carp. p. 103, fig. 89.
14 Many examples could be shown.
15 A late-medieval example was illustrated; few, if any, surviving true crucks date from the 13th century.
16 Maps in A Catalogue of Cruck Buildings, ed. N.W. Alcock (2nd edn., Council for British Archaeology, 1981).
17 E.g. Siddington barn, Glos. (Hewett, Eng. Historic Carp. p. 88, fig. 75).
18 Medieval Archaeol. xxxi (1987), 66. [For felling date of 1179, Vernacular Architecture, xx. 46.]
19 Medieval Archaeol. vi (1964), p. 167, fig. 53.
20 Hewett, Eng. Cath. & Mon. Carp. p. 28, fig. 25.
21 Slide of Steventon church, Berks., nave (wagon roof perhaps of the mid 14th century, with mouldings in the Decorated style).
22 The slide illustrated a late-medieval example in Herefordshire.
23 Slide of Lime Tree House, Harwell (wrongly published as a manor house in Archaeological Journal (1979)).
24 The slide illustrated the 12th or 13th-century aisled hall at Temple Balsall (Warwickshire).
25 Sullington Manor (unpublished; from author's drawing.)
26 For date, see guide by J.T. Munby on site.
27 N. Drinkwater, 'The Old Deanery, Salisbury', Antiq. Jnl. xliv (1964), p. 42, fig. 1.
28 Author's drawing (unpublished). [See now Oxoniensia, lvii (1992), 126--32; for tree-ring date, Vernacular Architecture, xxiv. 55]
29 Victoria History of the County of Stafford, xvii (1976), plate facing p. 17. [Tree-ring date 1269 x 1293, Vernacular Architecture, xx. 41.]
30 J.M. Fletcher and C.R.J. Currie, 'Two Early Cruck Houses Identified by Radiocarbon', Medieval Archaeol. (1972), p. 139, fig. 141. The date of the second example is now  thought to be c. 1365 (unpublished dendrochronological work). [Dates not confirmed because of weak statistical correlations].
31 Slide from J.H. Parker, Some Account of Domestic Architecture in England, ii (1853), facing p. 87. Recent research indicates that the roof is earlier than the 14th-century windows.
32 Medieval Archaeol. vi (1964), p. 170, fig. 54.
33 Slide from J.H. Parker, Some Acct. of Dom. Archit. in Eng. ii, facing p. 32.
34 Slide of second phase of Lime Tree House, Harwell (author's photograph).
35 Hewett, Eng. Hist. Carp. p. 122, fig. 109.
36 Slide from unpublished drawing by J.M. Fletcher and author.
37 N.W. Alcock et al., Archaeol. Jnl. cxxxv (1978), figs. on pp. 206, 208.
38 Slide of Bayleaf Farmhouse by Richard Harris.
39 This statement derives from the published drawings, which are misleading. In fact there is no arch at the apex, and the upper collar is soulaced rather than arch-braced; Stokesay thus appears to belong to the earlier stylistic phase in which only the lower roof was treated with arcuation. [Tree-ring date 1287--90: Vernacular Architecture, xxviii. 160.]
40 Author's unpublished drawing. [See now Oxoniensia, lvii. 214--22.]
41 Oxoniensia, xxxviii (1973), 339--45. For date, Vernacular Architecture, xxi (1990), 47.
42 Slide from author's drawing. For date, Vernacular Architecture, xx (1989), 44. Evidence for tenure from research by the former owner, Dr. E.B. Evans.
43 Slide of Guesten Hall roof, Worcester, supplied by Avoncroft Museum of Buildings, where the roof is being re-erected.
44 Author's slide. For a photograph, see N. Pevsner, The Buildings of England: Herefordshire.
45 J.H. Parker, Some Acct. of Dom. Archit. ii, facing p. 35.
46 Author's unpublished section drawing.
47 Author's slide of Middle Farm, Harwell. For a photograph, Berks. Archaeol. Jnl. lxii (1965), plate X. [Tree-ring dates, Vernacular Architecture, xxiii. 48.]
48 Author's slide. [Tree-ring dates, Vernacular Architecture, xxiii. 48.]
49 Author's slide of porch at Leigh Court, 1989.
50 F.W.B. Charles and Mary Charles, Conservation of Timber Buildings (1984), fig. 269. The cusping is visible on site.
51 Hewett, Eng. Hist. Carp. p. 134, fig. 118. [Tree-ring date, Vernacular Architecture, xxiv. 52.
52 Chichester: Scientific Methods in Medieval Archaeol. p. 254, fig. 9.
53 Winchester Pilgrim's Hall: ibid. p. 255, fig. 10. For date, Vernacular Archit. xv. 69.
54 Slide by J.T. Munby. For photograph, H. Janse and L. Devliegher, 'Middeleeuwse bekappingen in het vroegere Graafschap Vlaanderen', Bull. Commission Royale des Monuments et des Sites, xiii (1962), p. 353, fig. 70. [Tree-ring dated to c. 1250 by P. Hofsummer.]
55 Hewett, Eng. Cath. and Mon. Carp. fig. 48.
56 Author's slide, 1988.
57 Author's slide of church porch at Dilwyn (Herefs.).
58 Author's slide of camber-beam roof of church porch at Weobley, Herefs.
59 Author's slide of St. Mary's Hospital, Chichester, showing south elevation.
60 Author's slide of the Cottage, Aston Tirrold, Berks. (now Oxon.), provisionally dated by dendrochronology to c. 1270 (unpublished) [See now also Oxoniensia, lvii. 103--7; for felling dates 1282, 1284, Vernacular Architecture, xxiii. 58.]
61 Author's slide of east elevation of 42-42A High Stree, Milton, Berks. Some idea of the framing can be obtained from published photograph, taken from NW., in Victoria History of Berkshire, iv (1924), facing p. 361.
62 Berks. Archaeol. Jnl. lxii (1965), p. 48, fig. 2. [Tree-ring dates, Vernacular Architecture, xxiii. 48.]
63 Author's slides of Red Lion Inn and adjoining buildings, Weobley.
64 Author's slide of Mavesyn Ridware gatehouse. [Tree-ring dates Vernacular Architecture xxvii. 79]
65 Illustrated in Charles and Charles, Conservation of Timber Buildings.
66 Author's slide of east window of aisle of North Moreton church, Berks.
67 Author's slide of Little Moreton Hall, Cheshire.
68 Author's slide of Gawsworth Hall, Cheshire.
69 Parker, Some Acct. of Dom. Archit. ii, facing p. 258.
70 Author's slide of Zacharias. For dates, Vernacular Architecture xix. 43, [xxiii. 43. See now Oxoniensia lvii. 244-309.]
71 Slide of Bayleaf, Weald and Downland Museum, Singleton, Sussex.
72 Slide of Wales Farm, Barton under Needwood, Staffs.