Deep thoughts on the Devon, and a fresh look at the Nith[1]

Devon < *Domnona?

A fine river cutting eastward down from deep Glendevon in the high Ochils, the River Devon PER KNR CLA veers southward through the Yetts o’ Muckhart, turning sharply back at the Crook o’ Devon to drain westward into the winding Forth downstream of Stirling and two miles upstream of Alloa CLA.[2] The shorter, smaller Black Devon KNR CLA runs westward, roughly parallel about three miles to the south, from the Cleish Hills in Kinross to the Forth a mile downstream of Alloa. The Devon (aquam de Douane c.1173 (Glen)dovan 1210 (Glen)dofona 1271) has always been derived from British (B) *Dubona ‘black river’, ‘Black One’, or even ‘Black Goddess’[3], a meaning which Duibhe, the Gaelic form of the name, appears to confirm. Yet this may only be an assimilation to a perceived dubh ‘black’.

In 1860 Glendevon village was Downhill[4], while in the fourteenth and fifteenth centuries the highly visible hillfort of Down Hill [NO 001036], which guards the Yetts o’ Muckhart beside the now-dammed river, seems to have been Dundovane[5], evidently an exact equivalent of the Monklands LAN Dundyvan (Dundyvene 1545, Dundyvane 1582, Dundovane 1587 Dundovan 1587[6]), though there is no fort here.[7]

Pardovan WLO hails from the Cumbric reflex of Welsh (W) pa(w)r + dwf(y)n ‘deep pasture’, ie ?‘lush grazing’[8]: Pardufin in 1124 (an almost ‘Welsh’ form), Purduuyn 1282, later Pardovin and Pardovan in 1541.[9] Namesakes exist in Pardivan by Haddington ELO, Parduvine by Gorebridge MLO and Perdovingishill RNF, this last in fifteenth century orthography.[10]

Devon, the English shire, is itself derived from Defnas ‘men of Devon’ < B Dumnonii (‘the name of the Celtic aborigines, …transferred to their Saxon conquerors’ as Ekwall perceived it[11]), an Iron Age kindred probably related to the so-called *Damnonii (after Ptolemy’s Damnonioi[12]) of Strathclyde, recte the Dumnonii[13], perhaps ‘worshippers of (the god) *Dumnonos’, ?‘the mysterious one’, a figurative use of B *dubno- *dumno- ‘deep’.[14] A proto-Pictish or North British form *domno- or *dobno– has been deduced and proposed[15], which perhaps confused Ptolemy’s sources, and gave us forms in dov-.

There is another Glendevon in Winchburgh WLO (Glendaven 1754), probably transferred from the Ochils (if in fact this represents the same toponym), as ‘the topographical features do not in the least correspond with the meaning of the name (Celt. gleann, a narrow valley; Dubona, the black river); indeed, no river is near.’[16] Local antiquary Hardy Bertram McCall tells us that

on 1st October 1484, there is a charter by Gawin of Levinstoun [Livingston] of the lands of Howatstoun lyand in ye barone of Caldor [(Mid)calder MLO, now WLO] in favour of Mergrete Hay, spouse of John of Glendony. Six years later an action was raised in the Court of Session by John of Glendovyn and Mergrete Hay his spouse against Gawin Levinstoun of yt ilke for wrangwis vexacioun and distrubling thame in ye peaceable possessioun of ye landis of Howatstoun… On 5th July 1492… [four tenants raised] an action against Johne of Glendony and Mergrete Hay his wife, [and Gawin’s son and apparent heir Henry], for taking dowble malez of the said landis, and that Glendony had wrangwis awaytuke and withhalden fra the forsaid tenents five horses and four hed of nolt…[17]

While there is no absolute assurance of this, it seems likely that Glendony and Glendovyn are alternative late fifteenth century spellings of a Glendevon, not noted elsewhere.

Despite some confusion over vowels—note again though the above-mentioned alternative W dyfn ‘deep, dense’[18] (?as preserved in Dundyvan)—it now seems highly likely that the river-name too springs (as its current form suggests) < B *dumno-, *dubno- ‘deep’, or more likely from its North British reflex *domno-, in a form such as *Domnona ‘Deep One, Mysterious One’, whether divine (and there is considerable evidence of Pictish river-worship[19]) or not.

In truth the oft-proposed derivation seems further challenged by the more southerly Black Devon which rises by the suggestively named Aberdona House (Aberdonie 1652), itself likely < *domn-. Note further Devon, Kettle parish FIF, earlier Dovan < Gaelic (G) domhain ‘deep, low-lying’, also Baldovan near Dundee and a Ball Domin in the Gaelic Notes to the Book of Deer.[20] Devonburn by Lesmahagow LAN may prove to have a similar source, while Blendewing by Kilbucho PEB seems to offer us the element compounded with the Cumbric equivalent of W blaen ‘end; source or upper reaches of river or stream; uplands’.[21] And here the English river Devon LEI NTT (Dyvene 1252 Deven 1342, whose perhaps more southerly phonetics Ekwall equated with the Scottish Devon as ‘black, dark’[22]) naturally shifts its semantics to flow into our pool of ‘deep’ names.

To the northernmore *Domnonii Ptolemy attributed Colania (Camelon? STL), Vindogara (by Irvine AYR), Coria (Barochan Hill? RNF), Alauna (Ardoch PER), Lindum (Drumquhassle? STL) and Victoria (Inchtuthil? PER), ‘so that their territory should have extended from Ayrshire and Renfrewshire across the Forth-Clyde isthmus into Dunbartonshire, Stirlingshire and southern Perthshire’.[23] Could it be that our river Devon once marked a boundary of their land?

Though possible, it is not very likely that the river commemorates the Iron Age kindred-name: ‘The Deep One’ is sufficient. However, Cardowan Wishaw LAN and Dowanhill Milngavie DNB may preserve the ethnicon of these first recorded inhabitants of Strathclyde, and hint at former strongholds,[24] as may apparently unfortified Dundyvan itself. Devonshaw (Hill) LAN (beside the Clyde), and Devonside LAN (by the Douglas Water), where archaeology has found apparently Iron Age features, also merit further study.[25]

*NOVIUS > Nith

The Nith DMF KCB is a longer, more direct, yet equally contrary river: it rises about twenty miles east of Ayr, yet chooses to flow south-eastward the thirty-odd miles through fertile Nithsdale to Dumfries, thence out into the Solway Firth. Its watershed defines part of Ayrshire’s boundary, its estuary part of Kirkcudbrightshire’s.

Ptolemy refers to it (or another river in this area) as the *Novius (Noouiou potamou[26]), < B *nowiio- ‘new’ (Modern W newydd ‘new, fresh’ etc.[27]), an element found also in Gaul.[28] Likely named from the river (a contentious matter[29]), the Novantae, an Iron Age kindred of what is now Galloway, have recently been credited with a guest appearance in The Gododdin, as Nouant (or Ënouant).[30]

While there have been (as ever) some dissenting voices, W J Watson asserts the likelihood of Novius > Nith, though ‘it is difficult to say with certainty what precisely is impossible in the case of a name… trnsmitted from Old British through Welsh into Gaelic and thence into English’.[31] He is followed here explicitly by Rivet and Smith.[32] Watson, though, pursued the matter further, adducing Newburn FIF to his argument: formerly Nithbren, its second element may well be W pren ‘tree’, the first a form of newydd, used like the G úr-chrann ‘a green tree (lit. a new tree)’, and doubly cognate núa-chrann, to mean ‘Green-tree’. He concluded that ‘it would be rash, therefore, to deny that Nith may represent Novios.’[33]

The river Nith is (Stra)nit c.1124 (Strad)nitt 1124 x 1140 Nud 1181 Nyth c.1240, the earlier names showing P-Celtic *strad ‘dale’ (cf W ystrad ‘vale’ < Latin strata[34]), as is to be expected in the south-west and other parts of Scotland (cf 12th c. Stradeern ‘Strathearn’).

To strengthen Watson’s argument even more, there is a further number of toponymic analogues which will have percolated through the same strata of tongues (where his ‘Welsh’ is to be interpreted as ‘Cumbric’, his ‘English’ as ‘Scots’):

  • Niddry Kirkliston parish WLO (Nudreff 1370, Nudry 1392, Nudre 1410, Nidre c.1542, Nudery 1571, Nuddrie 1614[35])
  • West Niddry WLO (West Nwdry 1521, Westnudry 1534[36], West-nethrie post1545[37])
  • Niddrie MLO (Nudreth 1140, Nodrif 1160 x 1214, Noderyf 1264 x 6, Nudreff 1296[38], Nudref 1290, Nodref 1335[39])
  • Longniddry ELO (Nodref, Langnodryf 1315 x 1321, Loungnudrethe 1380–1[40], Langnudre 1424[41])

The elements’ order hints at their antiquity[42]: these must all derive directly from the Cumbric development of B *nowiio- ‘new’ + *treb- ‘steading etc.’, Modern W newydd + tref = newydd dre(f), as in (Y) Drenewydd (Newtown MTG). Ultimate -dd tends to drop in Welsh (eg i fyny ‘upwards’ formerly i fynydd ‘to the mountain’; *Castell newi = Castellnewydd (Emlyn) CRD); the following voiced dental here would have assisted this process, and the forms may also have been influenced by G nodha (fem. of nuadh), Irish nuadh < nue[43], or indeed merely by Scots new: cf Kirknewton MLO (now WLO), locally *The Nitton or *Kirknitton.[44] Whatever the process, all these place-names have descended into today as Ni-.

And whatever be proposed as the Cumbric form fossilised in these toponyms (?*nouíd > *nówið), there can be little doubt that the Nith too comes down < B *nowiios ‘new, fresh’, etc., and is indeed Ptolemy’s *NOVIUS.

[1] Published in Nomina (Journal of the Society for Name Studies in Britain and Ireland) Vol. 25 (2002), pp. 139-45.

[2] An earlier condensed version of the first section of this article was published in Scottish Place-Name News (The Newsletter of the Scottish Place-Name Society / Comann Ainmean-Aite na h-Alba), 4 (Spring 1998), Notes and Queries, 9–10. County abbreviations (pre-reorganisation) follow Scottish toponymists’ usage and are as listed in W. F. H. Nicolaisen, Scottish Place-Names. Their Study and Significance, new edn (Edinburgh, 2001), pp. xxi–xxii.

[3] W. J. Watson, The History of the Celtic Place-Names of Scotland, (Edinburgh, 1926, repr. 1993), p. 438; J. B. Johnston, Place-Names of Scotland, 3rd edn (London, 1934), p. 155; A. Macdonald, The Place-Names of West Lothian, (Edinburgh, 1941), p. 46; E. Ekwall, The Concise Oxford Dictionary of English Place-names, 4th edn (Oxford, 1960, repr. 1991), p. 143; Nicolaisen, Scottish Place-Names, p. 228, A. Watson, The Ochils: Placenames, History, Tradition, (Perth, 1995), p. 56.

[4] Watson, The Ochils, p. 56.

[5] Ibid., p. 56. Note Cornish down, Breton doun ‘deep’.Geiriadur Prifysgol Cymru / A Dictionary of the Welsh Language, edited by R. Thomas et al. (Cardiff, 1950–2002), s. dwfn; (henceforth GPC).

[6] This last spelling is from P. Drummond, Placenames of the Monklands, (Monklands Library Services Department 1987), p. 11, where it is derived from Gaelic dun dubh-abhainn ‘the fort by the dark water’. My gratitude to him for providing me with the first three forms (pers.comm. March 2001).

[7] To judge from its absence in (R[oyal] C[ommission] on the A[ncient and] H[istorical] M[onuments of] S[cotland], Lanarkshire: An Inventory of the Prehistoric and Roman Monuments (Edinburgh 1978).

[8] dwfn, dyfn ‘deep, dense’; pawr, pl. porion, ‘pasture, grass, a grazing;’ < pori ‘to graze’ etc., (GPC, s.vv.). It is important to note that W *par does not feature in GPC, but see parlas ‘green patch of ground’ s.v.; Adpar, Llandyfriog CRD (locally *Atpar, W ad- < B *ate- ‘second; again’ etc., GPC) has been explained to me by a native Welsh-speaking farmer as ‘pasture with good regrowth’: thanks to Gareth Ford for this. On par- in Scotland, see Watson, The History of the Celtic Place-Names of Scotland, pp. 372–3; W. F. H. Nicolaisen, The Picts and their Place Names, (Rosemarkie, 1996), p. 27.

[9] Macdonald, The Place-Names of West Lothian, p. 62; J. G. Wilkinson, West Lothian Place-Names, (Torphin, 1992), p. 29.

[10] Watson, The History of the Celtic Place-Names of Scotland, pp. 372–3. Note early Welsh spellings: duuin (12th c.), Duvin (12-13th c.), dwfyn, dwuyn (13th c.), dyfwn, dofyn (14th c.) (GPC).

[11] Ekwall, The Concise Oxford Dictionary of English Place-Names, p. 143.

[12] Ptolemy, Geography, II, 3, 7. A. L. F. Rivet and C. Smith, The Place-Names of Roman Britain (Cambridge, 1979), pp.138–40 (Chapter Three, ‘Ptolemy’s Geography’, where the edition of C. Müller & C. T. Fischer (Paris 1883–1901) is referred to).

[13] Rivet and Smith, The Place-Names of Roman Britain, pp. 342–4.

[14] Ibid. It is often proposed that the Dumnonii were miners; they could just as well have been mariners.

[15] J. T. Koch, ‘The Stone of the Weni-kones’, Bulletin of the Board of Celtic Studies, 29 (1980–82), 87–9.

[16] Macdonald, The Place-Names of West Lothian, p. 46; Wilkinson, West Lothian Place-Names, p. 25 (where the traditional derivation < *Dubona was also followed).

[17] H. B. McCall, The History and Antiquities of the Parish of Midcalder, (Edinburgh, 1894), p. 131 (italics mine). No specific authority is cited, but see his Introductory Observations, pp. 1–6 (especially pp. 2–3) for his many potential sources, probably Acts of the Lords of Council in Civil Causes (1478–1495) here. These alternate forms suggest that the Glendevon’s Innerdounie or Innerdownie (Hill) also refers to the river and glen. Note too the Scots surname Cardownie.

[18] GPC, s.v.

[19] Nicolaisen, The Picts and their Place-Names, pp. 21–23.

[20] I am grateful to Dr. Simon Taylor for these three names (personal communication, June 1997).

[21] GPC, s.v.

[22] Ekwall, The Concise Oxford Dictionary of English Place-Names, p. 143.

[23] Rivet and Smith, The Place-Names of Roman Britain, pp. 342–4.

[24] These references were taken in good faith pre-2002 from local informants but are at best secondary localisations of the names. Read: ‘However, Cardowan Glasgow LAN (Car<d>euien 1144 x 1124, Curdowane 1562 x 1592) and Dowanhill Govan LAN (no early forms found) may preserve the ethnicon of these first recorded inhabitants of Strathclyde…’ Thanks to Dr Simon Taylor for pointing this out and providing the correct locations, January 2014.

[25] Lanarkshire, (RCAHMS): Devonshaw Hill [NS 962286], p. 50 (cairns), pp. 100–1 and p. 150 (enclosures & fort); Devonside [NS 916394], p. 87 (settlement). If *Domnonii represents a derivative of a divine name [see Rivet and Smith, PN Roman Britain, loc. cit.], then *Domnowalos > Dyfngual/ Domhnaill > Donald, by analogy with *Luguwalos ‘strong in Lugus’, the eponym of LUGUVALIUM now Carlisle, is not necessarily ‘world-strong’ but may be ‘strong in *Domn(on)os/Domnona’ or the like.

[26] Ptolemy, Geography, II, 3,2; Rivet and Smith, The Place-Names of Roman Britain, pp. 133–35.

[27] GPC, s.v. Note Middle Cornish nowyth, newyth, Old Breton nouuid (ibid.).

[28] Rivet and Smith, The Place-Names of Roman Britain, pp. 425–8.

[29] Ibid.

[30] J. T. Koch, The Gododdin of Aneirin: Text and Context from Dark-Age North Britain, (Cardiff 1997), pp. lxxxii-lxxxiii.

[31] Watson, The History of the Celtic Place-Names of Scotland, pp. 54. Colin Smith’s ‘quotation’ from Watson is not verbatim and should be emended as above. I am grateful to Dr. Carole Hough for pointing this out to me.

[32] Rivet and Smith, The Place-Names of Roman Britain, pp. 428.

[33] Watson, The History of the Celtic Place-Names of Scotland, pp. 54-5. Note, therefore, that some names in New- may be translations through two Celtic tongues.

[34] W ystrad ‘(floor of) valley, vale, plain’ is actually < Clt *strato- < IE *sterə- ‘widen, spread’ with influence from probably cognate L strāta (via) in its secondary and later meaning ‘street’ (GPC, sv., [which hadn’t reached S-Z at that time!]). Thanks again to Dr Simon Taylor for first querying this derivation.

[35] Macdonald, The Place-Names of West Lothian, pp. 43–4.

[36] Ibid.

[37] Cited in RCAHM[ & Constructions of]S, Tenth Report with Inventory of Monuments and Constructions in the Counties of Midlothian and West Lothian, (Edinburgh 1929), p. 211, from Sir Richard Maitland, History of the House of Seytoun (undated). This form was missed in both Macdonald, The Place-Names of West Lothian (p. 43-4) and Wilkinson, West Lothian Place-Names, (p.29).

[38] Nicolaisen, Scottish Place-Names, p. 169.

[39] Watson, The History of the Celtic Place-Names of Scotland, p. 363.

[40] Nicolaisen, Scottish Place-Names, p. 169.

[41] Watson, The History of the Celtic Place-Names of Scotland, p. 363.

[42] K. H. Jackson, Language and History in Early Britain, (Edinburgh, 1953, repr. Dublin 1994), pp.225–6: perhaps pre-seventh century.

[43] Nicolaisen, Scottish Place-Names, p. 169; Watson, The History of the Celtic Place-Names of Scotland, p. 363.

[44] Where orthographic -tt- (locally uttered as a glottal stop) serves to keep the preceding vowel short; -i- is as in fish or him in West Lothian parlance: a terser equivalent of Welsh -y- (eg as in Dyfed).

[Space did not permit this addition to the published article: ‘I should like to thank Dr. Simon Taylor again for his encouragement and typically enthusiastic reading of the initial versions of these two pieces; belated but lasting gratitude likewise goes to the late Professor Colin Smith, who appreciated the evidence of the second.’]

Comments are closed.