The Castles of Kent No. 2: Thurnham Castle

Never in the history of Kentish archaeology has a single ancient monument had so many theories put forward as to its origin, on such little evidence, as has Thurnham Castle.

Overlooking the wide vale in which flows the River Len, Thumham Castle (TQ 8080 5817) stands near the top of the North Downs some 250m. to the north of, and 90m. above, the village of Thumham. The castle is reached by following a narrow country lane which, as it winds its way up the scarp slope, turns into a classic hollow-way (fig. 1).

As with most settlements there is no reference to the manor of Thurnham before Domesday Book (1086) which tells us the land was held by Ralph de Courbepine a tenant of Odo, Bishop of Bayeux and Earl of Kent. A church and mill are mentioned, but no castle. However, as Domesday Book was primarily an economic document the failure to mention a castle at Thurnham is of no significance. The present writer suspects that there was not a castle present in 1086, but is unable to offer reasonable proof.

What can be done is to dismiss many of the previous theories as to its origin. The most impressive statement about the history of the site comes from Jerrold's work Highways and Byways of Kent (1907, p.295) where he states, ' ..... no authentic history of the castle has come down to us ..... '. True then and true now. Unfortunately other writers have teased us with information, stated possibilities and probabilities without producing any evidence, been economical with the truth or told downright lies. To explain in detail most of the theories put forward at various times in the last four hundred years would require a work significantly longer than the Newsletter would allow, only a brief summary can appear here.

In his manuscript De Casleflius Ca11lii, written in the late sixteenth century, William Darell 'conjectured' that the castle was built by a Saxon called Godardus (Philipot 1659, p.341-2). This manuscript is stated as being in the archive of the Royal College of Arms (Stephens 1888, p.53-4), but is now lost (Yorke, per. comm.1984; 1993). Even assuming an individual called Godard us existed it seems unlikely he was a Saxon. Whilst the personal name can be either Old French or Old German (Reaney and Wilson 1991, p.191) no Old English form is known. The statement about Godardus seems to be an attempt to explain the two alternative names for the site, Godards or Godwarde Castle (Leland c.1540 Vol.ii, p.30; Lambarde 1576, p.56). The derivation of this name will be discussed in a longer article (Ward, in prep. a).

In his Villare Cantianum, Philipot states that the monument was a Roman watch tower (1659, p.341-2). A statement followed blindly by all and sundry down to the present day. This idea falls apart if looked at in a systematic manner. It is necessary to understand how the Roman army worked. A watch tower would be part of a chain, connected to forts in which would be stationed larger garrisons who would take offensive action against an enemy. No such chain of watch towers along the North Downs has ever been found or suggested. The remains at Thurnham became a watch tower of Roman date purely because the masonry and earthworks were visible, not because they could be dated in any way. The Roman urns (i.e. complete pottery vessels rather than sherds) supposedly found nearby (Harris 1719, p.317) are more likely to come from a cemetery connected with the villa in the vale than any other structure.

In 1659 Richard Kilbume tells us (p.259) that the castle was founded by a Sir Leonard Goddard in the time of King Stephen. Our problems are over, we have a definitive statement. Unfortunately Kilburne is totally unreliable and. in addition we have a relatively good run of names for the late eleventh and twelfth century holders of the manor, unfortunately Sir Leonard does not appear in the family genealogies of de Magimot, de Say or de Thurnham.

The suggestion that the monument was a British hill fort is not supported by any evidence (Payne 1893, p.129).

The new medieval documents which directly mention the castle at least gives us a terminus ante quern for its foundation. The castle must have been in existence by c.1220, for the walls are mentioned in a charter (Anon. 1862, p.215). At a slightly earlier date (1174-84) another document tells us that Stephen de Thurnham had a private chapel within his 'court' (ibid. p.201). This court would almost certainly be the castle. By deduction we can go further. This type of fortification whether ringwork or motte and bailey would, for southern England, be old fashioned by the second half of the twelfth century. A date of construction in the 'Anarchy' of King Stephen's reign (1135-54) is possible, but his successor, Henry II (1154-89), by tradition, pulled down these o called 'adulterine' castles (i.e. fortifications built without royal permission), in which case a date prior to 1135 becomes Likely. The use of masonry for a private fortification, at least on this scale, was unknown in the Anglo-Saxon period, therefore a date of 1067 can be regarded as a safe terminus post quem. We can go fmther, even the Tower of London was not begun until the 1080's so a stone rural castle would be very unlikely before a date of c.1085.

The reader is directed to Hasted for the first detailed description of the castle (1798, Vol.v, p.520-32). Despite statements to the contrary (Page 1974, Vol.l, p.424) it appears not to have changed in overall appearance since his day (Ward 1985). However, there was far more flint facing in place twenty years ago than there is now (Spall, per. comm.1992;1994). Much of the north wall of the castle still stands with small areas of coursed flintwork face surviving. Adjacent to the gatehouse the wall survives to a height of c.3 m. above modern ground level. The truncated west wall can be seen in places above, and following the ,--Jine of, the hollow-way. That this wall was constructed along the edge of the lane rather than on the scarp slope tells us that the hollow-way was there before the castle. The south wall has long since been destroyed, its tumbled down ruins could be seen in Hasted's day at the base of the cliff formed by chalk quarrying. That the ruinous castle was not totally destroyed is probably due to the presence of a barn, stables, buildings, gardens and orchards within and adjacent to the courtyard (K.A.O.D. U348 T3 and T6). All of these structures had been there in the seventeenth century (K.A.O.D. Ul258 Pl, Ul258 P2 and U588 Pl) which perhaps implies that the castle site had not been abandoned.

Thurnham is always referred to as a matte (mound) and bailey (courtyard) castle. If it is - then it is a very strange one. A matte should be shaped as is shown in figure 2a, Tonbridge being the classic example in Kent, whereas at Thurnham the profile is as shown in figure 2b. There are two possible solutions.

1. That a shell keep (i.e. a wall situated on top of a matte and encircling an open area, Tonbridge is again a good example), survives to a considerable height below the earth, trees and bushes that we see today. The actual matte in being considerably lower than first sight would suggest and certainly much lower that the 100ft. (30m.) quoted by Ditchfield and Clinch (1907, p.198). There is no purpose built matte in the country that high, only a very small number go above ten metres and none above fifteen.

2. That the profile we see is that of a ringwork. To the present writer this seems the correct solution. A rampart, rather than a mound, was constructed. When complete, if indeed it ever was, the rampart would have enclosed a small courtyard area which may or may not have had structures within. The rampart would have been constructed from the earth and chalk excavated from the ditch which can be seen separating it from the bailey. Just to the north of the chalk quarry a short stretch of flint wall, surviving up to a metre high can be seen for a length of c.Sm. Only clearance of the scrub would show whether or not the wall continued around the rampart, and only excavation might tell us if they are contemporary with each other and with the bailey.

Within the interior formed by this rampart a c.4m. diameter depression, first noted by Hasted can still be seen. First impression is that this feature is too wide for a well, although this could be accounted for by collapse of the sides. A more serious criticism of this feature being a wellhead is that the water table is c.90m. vertically below the site of the castle. However, it can be shown that on the dip slope of the North Downs the water table has dropped (Ward 1997; in prep. b and c), and on the South Downs an overall lowering of c.30m. since the Roman period has been suggested Oacobi 1978, p.77). It is possible therefore that the scarp slope water table may also have been higher in the medieval period and consequently the depth of the shaft would not have been so great. However, this seems unlikely, for the simple reason that it is the junction of the chalk and impervious gault clay which forces the water out. This junction has, obviously, not altered. It seems therefore that the spring level on the scarp side of the North Downs will have remained fairly constant. That deep shafts could be excavated in the medieval period of that there is no doubt, one only has to look at the well at Dover Castle. However, Dover is an exceptional case, it is a large royal castle and played a key element in defending the south coast, water may have been a decisive factor in any siege. Thurnham serves a different function. It would not have been a permanently manned defensive work, the lord of the manor would only be in residence occasionally and only then would there be relatively large numbers (tens of people, not hundreds) to cater for. If the depression was for water, a storage cistern lined with clay and perhaps covered by a structure is perhaps more Likely than a deep well. Fresh water could have been brought up from the village by some peasant as part of his servile duties.

As far as the ringwork is concerned it may be earlier, but due to its size and form a Norman date is to be preferred over any other. Indeed it is distinctly possible that masonry and earthwork are contemporary. If the earthwork were the earlier it seems logical that it should be constructed immeruately adjacent to the already existing hollow-way, which would then help emphasise the height of the ramparts. That the ringwork was constructed further to the east suggests a bailey was part of the initial construction. A timber palisade could have enclosed the bailey area prior to the masonry wall, but ii this is so it would appear to have been a very weak defence, for there is no sign of a rutch on the north. Such a feature may of course have been infilled, but a shallow depression in the ground surface, or in the side of the hollow-way would still be expected, especially as a ditch is visible between the ringwork and the bailey. Perhaps most importantly there is no sign that the masonry wall stands on, or has cut into, an earthen rampart (i.e. the upcast from a rutch). Without a ditch a timber palisade is inherently weak. A relatively high wall would form a better barrier and may therefore have been designed from the outset, although the absence of a rutch would still create a weak point.

Although great care is necessary when attempting to assign a date to flint walls, the way the facing stone is laid is similar to late eleventh or early/mid twelfth century examples elsewhere in the county; St. John's Hospital, Canterbury reredorter c.1085, St. George's Church, Canterbury, curved south wall c.1090-1150 (Ward. In prep. d), Eynsford Castle, c.1087-9 (Rigold 1971, p.112). Taking all the evidence together a date in the period 1085 to c.1125 for the masonry seems likely. Even ii the site were to be totally excavated the possibility of narrowing this suggested date range would be unlikely.

Successive writers emphasise the strength of the castle. Viewing (and climbing) from the south this is correct, but no enemy in his right mind is going to walk up a steep lane under fire in an attempt to assault the walls. The castle would be by-passed and attacked from the nothern side. This of course assumes that castles were primarily military establishments built as a protection against marauding armies and forming a strategic system of defence. Pure military scenarios for the construction of Thurnham Castle can easily be suggested, centred on the rebellion of Odo in 1088, or the anarchy of King Stephen's reign (Ward in prep., a), or the need of the Norman dynasty to defend London (Beeler 1956, p.581-601). However, most small rural castles, at least in southern England, were probably never threatened by national events. They allowed the lord of the manor to show off his wealth and power, and provided security for his produce and moveable property against boisterous neighbours, rather than being a viable defence against any invading force or baronial army.

If not primarily for defence why go to the effort of building Thurnham Castle on top of a waterless hill? Starting from the church down in the hamlet those who are brave enough to walk up the scarp slope will be rewarded by a magnilicient view to south, east and west, the expression 'Wow, look at that view' has been used. The following suggestion is not provable, but there is no reason to believe a Norman lord (Ralph de Coubepine, Gilbert de Magimot, a de Say or de Thurnham) would have thought rufferently. Whilst out hunting or picnjcking he may weU have srud the same and orderd his home to be constructed on top of the North Downs purely so that he could enjoy the view, just as later generations were to do elsewhere in the county and further afield. It seems to the present writer that the overaU aesthetic visual appeal of the site may have far more to do with its position than any military factor.

References: Charters of Cumbwell Priory. Part 1, Archaeologia Ca11tiana, Vol.v, p.194-222. Castles and Strategy in Norman and Angevin England; Speculum, Vol.xxxi, p.581-601. Memorials of Old Kent. Perambulation of Kent. Leland's Itinerary. Ed. L.T. Smith 1964. The History of Kent. History and Topography of Kent. Vol.v. Highways and Byways in Kent. Kent Archive Office Document. Topographie or Survey of the County of Kent. Population and Landscape in Mesolithic Lowland Britain; CBA Research Report No.21, The effect of man 011 the landscape: the Lowland Zone. Ed.S. Limbrey and J.C. E� ans, _ p.75-85. The VICtona County History: . Kent, Vol.1. Repnnt. . -- of 1908 edition. Collectanea Cantiana. Villare Cantianum or Kent Surveyed and Illustrated. Second Ed. 1776. A Dictionary of English Surnames 3rd. Edition. Eynsford Castle and its Excavation. Arcl,aeologia Ca11tia11a, Vol.86, p.109-172. Personal communication, letters dated Dec. 1992 and Feb. 1994. Dictionary of National Biography. The Three Castles of the Stockbury Valley (unpublished B.A. Hons. dissertation). The Castles of Kent No. 1: Stockbury Castle. K.A.S. Newslelter No.37. Ward. A. a. In preparation: Thumham Castle: Fact and Fiction. Ward. Ward. A. b. In preparation: St. Mary Magdalene's Chapel, Lidsing. A. c. In preparation: Field walking at Hartlip Roman Villa. Ward. Yorke. A. R.C. d. In preparation: 2.8.93. Personal Excavations at St. communication, George's letters Church, dated Canterbury. 9.7.84 and

My thanks are extended to Mr. H. Batchelor the then owner of the castle for readily allowing access to undertake surveying in 1983-5 and to all the many individuals who have patiently replied to my letters of enquiry or volunteered information over the last fifteen years.

NOTE: The castle is situated on private land and there is NO public access.

Alan Ward. c/o History Dept. Christ Church College, Canterbury.