The name of the talk was taken from the title of the 1999 article by A. A. M. Duncan: ‘Yes, The Earliest Scottish Charters’, Scottish Historical Review 78 (1), 1–38. One of the shoutiest titles of any piece of academic writing, it was in fact part of an academic conversation that had been going on since 1958.[1]

The dispute was about the authenticity of a set of charters issued in the last decade of the 11th century by Kings Duncan/Donnchad II and Edgar of Scotland, and King William II (Rufus) of England. The talk concentrated on the Edgar and William ones, for lands in what is now Berwickshire, as Donnchad II’s charter relates only to what is now East Lothian. The reason for concentrating on Berwickshire is the 3-year Leverhulme-funded project ‘Recovering the Earliest English Language in Scotland: the Evidence of Place-Names’ (for more details, see Carole Hough’s piece, this issue).

The dispute about the genuineness or otherwise of these charters has no real bearing on the analysis of the names as names, since both Donnelly and Duncan agree that at least some of the charters in question are original documents from the 1090s. namely those issued by William II in 1095, so the forms reflect usage of the late 11th century, whatever the agendas of those compiling the documents were. The forms in King Edgar’s charters are from later medieval copies of originals issued at the same time, and in the same context. This context is Edgar’s attempt to take the Scottish throne from his uncle, King Domnall III (Bán), with the support of the English king. In fact the attempt failed, and Edgar did not become king of Scots till 1097.

William II (Rufus)’s confirmation, which exists in Durham Cathedral’s archive as Durham MC558a, is of King Edgar’s grant to God and the church of Durham and to St Cuthbert the confessor and to Bishop William and to the monks there. It contains two central places, Berwick and Coldingham, each referred to as a manor (Latin mansio),[2] the former with 20 manors (mansiones), the latter with 10 manors, attached, and all defined as being in Lothian (in lodoneio). The names of those attached to Berwick are (in the order as listed in the charter), along with their modern and medieval parish: Graden (Coldstream, formerly Lennel), Lennel (Coldstream etc), Dylsterhale # (probably Coldstream etc), Birgham (Eccles), Edrom (parish), Chirnside (parish), Hilton (medieval parish, now in Whitsome), Blackadder (Edrom), Kimmerghame (Edrom), Hutton (parish), Renton (Coldingham), Paxton (Hutton), Foulden (parish), Mordington (parish), Lamberton (medieval parish, now in Mordington) and the other Lamberton, Edrington (Mordington), Fishwick (medieval parish, now in Hutton), Horndean (medieval parish, now in Ladykirk) and (Wester) Upsettlington (medieval parish, now part of Hutton). It is notable that half of these manors are recorded as parishes in the later Middle Ages.


Before looking at these names more closely it is necessary to think about language. The big question in southern Scotland in general, and in Berwickshire in particular, in dealing with a name of Germanic origin is: which language was it coined in? The two main contenders are Old English (up to about 1100) and Scots (from about 1100). As there is to a large extent a continuum between Old English and Older Scots in this part of Scotland, it is often difficult, sometimes impossible, to know which of these languages a name was coined in. It is all to do with timing![3]

In answering this question, the following considerations are useful to bear in mind:

status of named place: OE names tend to survive in high-status settlements, especially those which have given their names to medieval parishes, e.g. Greenlaw and Fogo.

date of earliest record: OE names are found attached to settlements or estates recorded already in the earliest charters (late 11th century), as well as, of course, in Bede and some earlier records relating to Lindisfarne, e.g. Coldingham, Hutton, Foulden.

place-name elements: certain elements seem to have been used to form place-names only or predominantly in the Old English period; e.g. hām ‘settlement’ (literally ‘home’), as in Edrom; wīc ‘(specialised) farm, settlement’, as in Berwick, Fishwick.

A problematic element is OE tūn and its direct descendant, Scots toun ‘a farm, estate’. It seems to have been productive in both periods. Often found combined with personal names, these can provide the vital clue as to date of coining. For example Penston by Tranent ELO, although first recorded in 1381 (as Paynistona RMS i no. 638) contains the personal name Pain (Latin Paganus), specifically Pain of Hedleia (Hanle?), who was granted land in Tranent in c.1170 by Robert de Quincy (Barrow 1980, 23 fn. 104). We cannot say exactly when it became known as *Pain’s toun, but it cannot have been earlier than c.1170. We can therefore confidently assign this name to Scots.


Let us now home in on the names in our Berwick group, which pose some extremely knotty toponymic puzzles, to not all of which solutions can be claimed, certainly not at this stage of the Project. There are 21 names in all. 8 of these contain the OE tūn; 3 contain OE hām; 3 OE denu ‘valley’; 2 OE wīc; 2 OE halh ‘haugh’; 1 OE sīde ‘(hill-)side’. + 1 is a river-name used for a settlement, Blackadder. This comes to 20 including Berwick, as Lamberton contains 2 mansiones.

The three hām-names are Edrom, Birgham and Kimmerghame (noting that in the two last the g is pronounced as j in English jam). Taking the generic first: the most recent in-depth discussion of this element is to be found in Alan James’s article in JSNS (2010). It starts by giving a useful overview of previous scholarship on this much-debated element. He writes (with footnotes omitted):

Barrie Cox has demonstrated that [OE hām] was the most favoured habitative term among the earliest Germanic-speaking settlers and that it remained in use up to around the time of Bede [died in 735], being superseded from around 750 by a range of other nouns, most commonly -tūn. This observation is vividly illustrated by the only concentration of place-names in ‑ham to be found in (or just outwith) Scotland, which is in the Tweed basin below Melrose, down to the Merse. Norham NTB and (probably) Yetholm ROX were heads of sċīras; both of these became mediaeval parishes, as did Edrom BWK, Ednam ROX, Oxnam ROX and Smailholm ROX. Midlem (Bowden ROX), Birgham and Leitholm (both Eccles BWK) and Kimmerghame (Edrom BWK) …’

So we may well be seeing what James terms ‘vigorous and determined English-speaking colonisation by the early eighth century’ (2010, 104). However, James reminds us of one meaning of hām already pointed out by A. H. Smith in his important English Place-Name Elements of 1956: that it can also refer to a monastic household or monastery, a religious house, a minster, as expounded also by Victor Watts in his 1994 article on Hexham. And as such it was certainly in use later than Barrie Cox’s mid-8th-century date, perhaps up until about 900 or even later. We have to bear in mind that this ecclesiastical usage is only one of several applications of –ham and –ingham , but it should always be carefully considered in cases in which there is supporting evidence. Such supporting evidence can be written or material (such as archaeological or sculptural), as for example for Norham, just over the Border from Upsettlington, as well as for Hexham and Coldingham, and possibly for Edrom. In names such as Birgham and Kimmerghame, however, there is no comparable religious context.

Moving on to the specifics of these names: in Birgham (brygh<am> 1095 Durham MC559; bricgha<m> 1095 Durham MC558a; Brigham 1189 x 1195 RRS ii no. 317) this is OE brycg f. ‘a bridge; a causeway’ (VEPN s.v.). Given its proximity to the Tweed we might assume it referred to a bridge across the Tweed, possibly to Carham, which lay on the opposite (English) side. But surely there was not such an ambitious bridge so early. It is more likely to have referred to a small bridge or causeway across a boggy or frequently flooded area. A local place-name, and an old map, provide just such a context.

Fig. 1. Armstrong’s map of Berwickshire, 1771, showing Lochton (Loughtoun) by Birgham, with its eponymous loch and a burn draining out of it into the Tweed. It is probably the bridge or causeway which carries the old road across this burn that is referred to in the name Birgham. Image courtesy of http://maps.nls.uk/

Fig. 1. Armstrong’s map of Berwickshire, 1771, showing Lochton (Loughtoun) by Birgham, with its eponymous loch and a burn draining out of it into the Tweed. It is probably the bridge or causeway which carries the old road across this burn that is referred to in the name Birgham. Image courtesy of http://maps.nls.uk/

The metathesis of y/i/r starts to appear in the late medieval period. The modern pronunciation with palatal g has still to be investigated, but suggests that the final consonant in brycg was palatalised early i.e. /g/ >/dʒ/, unlike Scots brig.

There are several places in England containing the same elements, such as Brigham, Allerdale parish, Cumberland, which first appears as Briggham c.1175 (PNCumberland 2, 355), and Brigham Bank, Bolton parish, Westmorland (PNWest-morland 2, 140). These are both pronounced with a hard g.[4]

In Edrom (Edrem 1095 Durham MC559; ederha<m> 1095 Durham MC558a; Ederham x 1138 ESC no. CXVII), whose importance as a church is signalled by the fine Romanesque arch surviving there, the specific is the river-name Adder, as in the Whiteadder and Blackadder. It is on the Whiteadder, suggesting that the distinguishing white and black, which are already in place by the late 11th century, developed after the coining of this settlement-name. It is one of several OE settlement-names in south-east Scotland with a river-specific + a habitative element, e.g. Tyningham, Pefferham # (both ELO), Edrington (also the Whiteadder), Ednam (Edenham) ROX, Leitholm and Ayton. Then there is also the settlement-name Blackadder, without any habitative element, another of the mansiones of Berwick.[5]

The third of the hām-names is Kimmerghame (Edrom parish) (Chynbrygh<am> 1095 Durham MC559; cynebrihtha<m> 1095 Durham MC558a; Robert de Kynbriggeham 1296 Ragman Roll, 146). May Williamson analyses the first element as OE, probably ‘cows’ bridge’, referring to a bridge over the Blackadder Water, though she does admit the possibility that it is the male personal name Cynebriht (Williamson 1942, 15). In fact, the earliest form, which is from our William Rufus charter (Durham MC558a), reads not Cynebritham, as she thought, misled by an error in ESC no. XVI, but cynebrihtha<m>, which looks even more like the personal name.

The next generic in this group, and found in almost half of all the names, is OE tūn, the dating problems of which have already been discussed. These are Hilton, Hutton, Renton, Paxton, Mordington, Lamberton and the other Lamberton, Edrington and Upsettlington. This is not surprising, given that tūn was the most productive habitative element in the Old English period in England, as well as in Older Scots, in the form of toun. From the 12th century onwards, the unit referred to in Latin charters as villa almost certainly represented this word, the equivalent in Gaelic Scotland being baile. It could be argued therefore that it was also the vernacular term which in these very early charters was represented by Latin mansio.

As to the specifics, they can be divided up into personal names and topographical features, with one possible reference to animals (Lamberton), and one to a murder or murders (Mordington). Those which definitely contain personal names are Renton and Paxton. For Renton (Regninton 1095 Durham MC559; reinintun 1095 Durham MC558a; Regnintun c.1100 Durham MC555), May Williamson suggests that the first element is OE Regna, a short form of the male personal name Regenwald. This would become Regnan in the genitive. Alternatively the second element, -in-, may be a reduced form of the particle -ing, meaning something like ‘associated with’, so ‘farm associated with Regna’.

Paxton on the Tweed, appearing as Paxtun or Paxton consistently from the earliest record seems to contain a ‘strong form, *Pæc(c) of the OE [male] personal name Pac(c)a’ (Williamson 1942). There are several places in England which seem to contain this name, including Great Paxton (Huntingdonshire), which Ian Cowan confused with this Paxton, hence his listing it as a parish in his book on medieval Scottish parishes (1967, s.n.).

Lamberton, a medieval parish now part of Mordington, was already divided into two parts by the 1090s, suggesting a relatively large and important estate (lamberton … aliam lamberton 1095 Durham MC559; la<m>bertun … aliam la<m>bertun 1095 Durham MC558a; Johanne de Lamertoun 1201 x 1233 Midl. Chrs. (Soutra) no. 16; (ecclesia) de Lambirtun c.1250 St A. Lib., 31). It may contain the Flemish personal name Lambert, although this would be remarkably early for Flemish settlement in Scotland. On present evidence a derivation from OE lambra tūn ‘farm of lambs’, containing the genitive plural of OE lamb ‘a lamb’, seems most likely. A full discussion of this name will appear in volume one of The Place-Names of Berwickshire.

Mordington (Morthyngton 1095 Durham MC559; morðintun 1095 Durham MC558a) seems to contain OE morð ‘murder’, again with the associative particle –ing, so ‘a farm or settlement associated with a murder or murders’, which may be compared with the Northumbrian place-name Morpeth, ‘murder path’ (Morthpath c.1200).

The remaining tūn-names are combined with topographical features: Hilton, formerly a parish, now part of Whitsome, is on a low hill.

Edrington (Hadryngton 1095 Durham MC559; hædrintun 1095 Durham MC558a; Edrington’ 1321 RRS v no. 187), as with Edrom, it contains the river-name *Adder (now Whiteadder Water), with the common –ing particle, signifying ‘associated with’ etc.

Fig. 2 Hutton Castle on its hōh ‘spur, projecting heel of land’ (whence Scots heugh). Huton 1095 Durham MC559; hotun 1095 Durham MC558a. Photo: Wikipedia.

Fig. 2 Hutton Castle on its hōh ‘spur, projecting heel of land’ (whence Scots heugh). Huton 1095 Durham MC559; hotun 1095 Durham MC558a. Photo: Wikipedia.

Upsettlington, which supplied the names of two medieval parishes, namely East and Wester Upsettlington, appears as Vpsetinton 1095 Durham MC559; upsetintun 1095 Durham MC558a; Hupsetligtun 1153 x 1160 Kelso Liber i no. 273; vpsetling’tune 1177 x 1204 Kelso Liber i no. 253; Hupsetinton’ 1194 SEA i no. 232). Smith 1956 gives under OE setl, with variants seld, sedl, *seðl ‘a seat, an abode, a dwelling’. ‘In a few cases the meaning is “seat”, probably used to indicate a lofty situation, as in Warshill, etc’. While OE up(p) he gives ‘up, higher, upon’. A lot depends on how we interpret the ubiquitous but quite troublesome particle –ing. While it may simply have its associative force, it may in this case refer to a group of people, so perhaps ‘the farm or settlement of the people living high(er) up’, or even ‘of a group called the *Upsettlings’. One problem that has to be addressed is the lack of l in the earliest forms. Perhaps *set or *seð was regarded as a variant form of setl, sedl etc. Williamson 1942 interprets this as ‘the [upper] farm or village on or by the ledge’, etc.

The talk also discussed the lost Dylsterhale (possibly in Coldstream) and Lennel, both of which contain OE halh ‘a haugh, water-meadow, etc’, with Lennel probably meaning ‘lean or meagre haugh’. There was no time to cover Berwick, Fishwick, Foulden, Graden, Horndean and Chirnside – perhaps matter for another talk, or a further article. However, there was enough to show that the early-recorded place-names of Berwickshire are, unlike Lennel, anything but lean, and will throw light not only on early settlement and language of this corner of Scotland, and on the very earliest roots of Scots, but also on the place-nomenclature of England.

Simon Taylor (from his talk at Galashiels)

Bibliography and References


Barrow, G. W. S., 1980, The Anglo-Norman Era in Scottish History (Oxford).

Cowan, I. B., 1967, The Parishes of Medieval Scotland, Scottish Record Society vol. 93.

Clancy, Thomas Owen, 2013, ‘Many strata: English and Scots place-names in Scotland’, in Perceptions of Place: Twenty-First-Century Interpretations of English Place-Name Studies, ed. Jayne Carroll and David N. Parsons (English Place-Name Society, Nottingham), 283-319.

James, Alan G., 2010, ‘Scotland’s –ham and –ingham names: a reconsideration’, Journal of Scottish Name Studies 4, 103-30.

Johnston, James B., 1940, The Place-Names of Berwickshire (The Place-Names of Scotland Series, No. 1, published by the RSGS, Edinburgh).

Watts, Victor, 1994, ‘The Place-Name Hexham: a Mainly Philological Approach’, Nomina 17, 119-36.

Williamson, May G., 1942, ‘The Non-Celtic Place-Names of the Scottish Border Counties’, unpublished PhD, Edinburgh University [on-line on <www.spns.org.uk>].

[1] Duncan, A. A. M., 1958, ‘The Earliest Scottish Charters’, Scottish Historical Review 37, 103–35.
Donnelly, Joseph, 1989, ‘The Earliest Scottish Charters?’, Scottish Historical Review 68, 1-22.
Duncan, A. A. M., 1999, ‘Yes, The Earliest Scottish Charters’, Scottish Historical Review 78 (1), 1–38.

[2] One of the Domesday Book terms for a manor, and one that was replaced by manerium.  Mansio has been found in a Scottish context only in this group of charters. Coincidentally, in the Lindisfarne Gospel hamas glosses mansiones in the NT passage ‘in my father’s house there are many mansions (mansiones)’ (Smith 1956 i, 226).

[3] For a good discussion of some of the problems involved, see Clancy 2013, especially pp. 286-9.

[4] Thanks to Alan James for confirming these pronunciations after the conference.

[5] For a BBC feature article on the REELS project, which mentions these river-named settlements, see <http://www.bbc.co.uk/news/uk-scotland-south-scotland-38090671>, an article that appeared on 5 January 2017.