This was a joint paper by myself and Professor Richard Coates. The formatting of the special characters has not worked but a complete version can be obtained from the Hon. Secretary of SDNQ. Thanks to programs on popular television channels the interesting in Viking studies has increased, hence this blog.
A problem of particular linguistic and historical interest in and around the Severn Sea is the occurrence of place-names of Scandinavian (Old Danish or Old Norwegian) origin. Scandinavian names in the Western and Northern Isles and the west and north coasts of Scotland, Pembrokeshire, the Lake District, the Wirral and much of eastern and East Midland England are a well-known and well-understood phenomenon. Their distribution correlates with known waves of Scandinavian incursion and settlement from the 9th century onwards. The presence of such names for coastal features in parts of the south-west of England and south Wales is more mysterious, given the complete absence of Scandinavian settlement names in Somerset, Monmouthshire and Gloucestershire, and their almost complete absence in Glamorgan. Such maritime names which are beyond reasonable doubt of Scandinavian origin include especially those of islands such as the outlying Lundy and Caldey (including Old Scandinavian [OSc] ey ‘island’), but also, and more intriguingly, the up-Channel Sker Point and Tusker Rock in Glamorgan (including OSc sker ‘flat rock, skerry’), and the grange called Meles in Margam (Glamorgan; though the source word melr ‘sandbank, sand dune’1 may have been borrowed into the English language and used here as an English word. There is a very small cluster of inhabited places in south-east Wales with Scandinavian names: Lamby ‘long farm’ (Langby, 1401), Womanby ‘houndsmen’s farm’ (Houndemanneby, 1310), and Homri ‘Horni’s farm’ (Hornby, 14th century, Horneby 1540), all in the Cardiff area.2 In the inner Severn estuary, in the Aust narrows, we find the late-recorded but characteristically Scandinavian-looking Leary Rock (apparently from OSc leira ‘mudflat’, or leir(r) ‘mud’, + ey ‘island’), Gruggy (also apparently ‘mud island’; compare gruggóttr ‘muddy’), Guscar Rock (apparently from gás ‘goose’ + sker ‘rock’) and The Scars (an anglicized plural form of sker ‘rock, skerry’).
It is often claimed that Steepholm and Flatholm should be viewed in this light as Scandinavian, since hólmr is a Scandinavian word for a small island. But since Steep Holm contains an English adjective (Old English stēap ‘steep’), it is pretty clear that hólmr must have been borrowed into English and used as an English word in this name. Flat Holm may well contain an English noun, Old English flot ‘deep water’, since spellings with <o> predominate in early mentions. The sometimes-claimed evidence that it contains Old Scandinavian floti ‘fleet’, relying on an incident in the Anglo-Saxon Chronicle, can be refuted,3 though it could be argued on linguistic grounds that it has late Old Scandinavian flǫt ‘piece of flat ground’ + hólmr. But if that were the case, we would have to accept the odd conclusion that Steep Holm is English but Flat Holm Scandinavian. The upshot is that Steep Holm is definitely not a name given by Scandinavian speakers, and Flat Holm probably is not. Undoubtedly both names testify in a loose kind of way to Scandinavian influence in the region, in that hólmr has been borrowed into English, but the safest conclusion is that both these islands were named by English speakers.
The existence of the rest of the names mentioned above gives us confidence in interpreting other difficult names in the Severn region, including coastal Somerset, as Scandinavian. These include the various Dunballs and the like (DunballIsland and later Dunball Wharf in Huntspill, Dunball Island formerly in Easton in Gordano, Dunball alias Cures Island recorded in the 1770s in the River Parrett, and the lost Dunbal Point in Otterhampton), representing occasionally overflowed shoreline pasture (from OSc dunna ‘mallard’ + ból ‘lair, den’);4 Middle Hope Bay in Kewstoke (probably from hóp ‘small bay’); and Plotneys (for which see below). The two appearances of the previously unexplained word layer which appears on either side of Plotneys on Williams’s chart (1815; one instance shows on the extract in Illustration 2),5 suggested for the first time here as perpetuating OSc leir(r) ‘mud’, are also noteworthy;6 compare Leary above. The original name of Anchor Head at Weston-super-Mare (Ankygsetyn ‘super mare’ in 1480,7Ankershed in 1492)8 appears to be for OSc ankeris-sæti ‘anchorage’. The exact reason for the adoption and persistence of such alien maritime and coastal names in English- or Welsh-speaking territory is still not fully understood; the main problem is the sociolinguistic one of how an apparently non-resident population of Vikings can have had the effect of making their naming practices stick in particular cases in English- and Welsh-speaking territory. In our present state of knowledge, it appears too much to believe that the inhabitants of the three small Cardiff settlements with names in -by were home to a population capable of imposing their linguistic will in coastal Somerset and Gloucestershire without settling there.9
Plotneys: a case study
The name of the rock feature Plotneys on the Portishead/Portbury parish boundary has not been fully explained before. In the light of the preamble above, it could be explained as ‘flat promontory’, from OSc flatr ‘flat’ + nes ‘ness, land projecting into the sea’, but with a modern form in <P-> probably due to a misreading of an early manuscript <F->, though it should be noted that surviving spellings in <P-> are earlier than surviving forms in <F->. The misread manuscript in question must now be lost, or at least undiscovered. This suggestion may seem chronologically perverse, but it should be noted that trying to find an etymology if the first element is Scandinavian and begins with original <p-> proves to be a wild Arctic goose chase. Nes, a word ultimately related to nose, appears Scandinavian; but Scandinavian words beginning with <p-> are rare, and almost always borrowed from other languages.
If the name included, instead, the more or less corresponding (West Saxon) Old English [OE, c.450-1100] word næss, we would have to conclude either that it always appears in an inappropriate Mercian or Kentish dialect form with <-e->, or that it has been influenced by precisely the Old Scandinavian word we are considering, and the possibility of Viking influence remains. OE næss appears in Sharpness (Gloucestershire), in which spellings with <a> persist till the late 16th century before <e> takes over; and in Totnes (Devon), where <a> spellings are rarer, but appear till the 14th century. The English word is rare in the south-west of England, appearing in just these two major names. But OSc nes is quite common in the north and east, and appears in major names such as Furness, Holderness, Skegness and Orford Ness. It appears that by some combination of influence from the Scandinavian word, which must have been familiar in place-names to coastal seafarers, and vowel change in a final unstressed syllable, the ness type has pushed out the nass type. Since Plotneys appears late in the record, as we shall show, it cannot be ruled out that it originally contained næss. But if so, a problem with identifying the first element as English persists. The word flat is a medieval borrowing of Old Scandinavian flatr; there is no evidence for it before the Conquest. And if the first element is Middle English [ME, 1100-1500] or even Modern English plat or plot, the sense of this, ‘small patch of (cultivated) ground’, is hardly appropriate for a maritime feature.
The upshot of this difficult argumentation about Plotneys is as follows:
(1) OSc flatr + nes requires acceptance of a manuscript transmission error substituting <P-> for <F->.
(2) If it is a name coined with OE næss, no suitable first element can be identified. We might consider OE flot ‘deep water; ?fairway channel’, because of the fact that the channel runs close to Portishead Point, but the earliest spellings generally have <a>, not <o> (see below). Substitution of <e> for <a> must be assumed, whichever reason (borrowing or phonetic change) is preferred for this occurrence.
(3) If it is a name coined with ME flat + nass, it cannot be earlier than about 1100-1200, a manuscript error must also be accepted, and substitution of <e> for <a> must be assumed, whichever reason (borrowing or phonetic change) is preferred for this occurrence.
(4) If the earliest forms showing initial <P-> are authentic, no plausible etymology is available.
Clearly only (1) and (3) are viable possibilities. (3) cannot be ruled out, but (1) is the most straightforward, especially when backed with the other evidence presented for Scandinavian influence in the Severn. With due reserve, then, we propose that the name Plotneys is another Scandinavian name.
The evidence offered by the charting history of the place about its being named, and about the form of the name when mentioned, is quite complex. The first record of this feature on a printed chart was made by Captain John Williams, who appears to have executed a survey around 1706.10 He referred to a small, but nevertheless noteworthy, area of rocks to the east of Portishead Point as ye Platness, in which the first word is a scribal rendering of the definite article the which remained conventional until well into the 18th century. (It is not to be pronounced ye!) This feature was situated approximately halfway between Portishead Point and the entrance to the River Avon. As these rocks lay to the south of the important anchorage of King Road it is not surprising that this danger to navigation was shown, and more importantly named, by Williams.11 What is surprising is that only a few years later, in c.1712, it was not named in a chart by Captain Holliday, even though the group of rocks is clearly shown in his chart, published as ‘A draught of the Bristol Channel from the Holmes to King Road, including the River Avon’, and included in Nathanael Cutler’s A general coasting pilot.12
Despite Williams’ having marked and named the feature on his chart,13 it was not universally copied by other hydrographic surveyors, cartographers or publishers.14 Murdoch Mackenzie senior and his nephew Lieutenant Murdoch Mackenzie junior R.N. both show a feature on their published charts from the 1770s which appears to be a patch of drying ground, in the same area, but neither named it,15 and they were followed in this by Captain John Knight R.N. in his published chart of 1806.16 However, Mackenzie junior appears to have been the first person to make a detailed, and what must be classed as accurate, description of the ‘Flatness, or Platness Rock’ in 1772,17 the first use of a form with <F->, using wording which seems to suggest he thinks the form on ‘the chart’ may be misleading:
Flatness, or Platness Rock.
Flatness, or Platness Rock as it is called in the chart, is a rocky shoal that lies along the drying mud bank from the coast, between Posset [= Portishead] point, and Warth point; its outer or north side, is about ½ mile from the coast, and it lies ENE and WSW, and is about ¾ of a mile long, and about a cable’s length in breadth: its west end, bears ENE from Posset point, ¾ of a mile distant, and its east end bears west from Warth point, a little more than ½ mile. Its east end is dry with spring tide, and very nearly joins the mud bank from the coast; but on the middle, and on the west end of this rocky shoal there are from 1, to 6 feet water, with spring tides.
The fairway channel, from Fort point, to Kingroad; to the entrance of the River Avon; and to the entrance of the River Severn; is between the northside of Flatness or Platness Rock, the south end of New Bank, and the south side of Three Feet Shoal. New Bank, and Three Feet Shoal, are far nearer to the coast of Somerset, than they are to the coast of Monmouth; but as they lie on the north side of the fairway channel to Kingroad, and are always to be taken on the larboard hand in sailing from the Holms to the River Avon; we shall therefore differ our description of them, until we have described the other banks and shoals, that lie on the north or on the same side of the fairway channel as they do, and to the westward of New Bank.18
More evidence for the existence of Platness Rock was provided by Mackenzie when he described the ‘Three Feet Shoal’ and a channel between it and Platness leading to King Road. From this time forwards the Platness Rock appears not to have been included in published sailing directions until 1821 (where it appears as Plotness-rocks).19 In a manuscript description of 1815 it is described as ‘a rocky shoal called the Flatness, to be avoided, which extends more than a Cable’s length from the Shore, and is dry at low water’.20 This is where we begin to confront the most serious textual difficulty. We cannot tell for sure whether the author, probably Henry Brooker R.N. master, was continuing an authentic early tradition of a name beginning with <F-> which has been obscured by the (?mistaken) previous use of forms with <P->, or whether he was influenced by the existence of the ordinary word flatness, viewed as appropriate here. The rocks did appear on numerous subsequent printed charts as The Platness in 1794,21 The Platnese in 1800,22 a Rocky Bank in 1803,23 the Flatness opposite Chapel Pill in 1815,24 and as an unnamed area of rocks in charts produced by different publishers in 182025 and 1829.26 The form Flatness was also used in sailing directions in the 1870s.27
We have suggested above that there are cogent linguistic reasons for believing the form with <F-> must represent the earlier tradition, despite the appearance of the record.28 But if that is the case, we must explain why the description ‘flat promontory’ is appropriate for a rocky shoal. OSc nes normally names a solid land feature, but the evidence of Thomas’ chart of 1815 suggests that it may have named an ancestor of the projecting mud bank conveniently marked XL (above); the rocky shoal ‘very nearly joins the mud bank from the coast’, according to Mackenzie junior. Accordingly, the name of the Plotneys must have originally been given to designate the rocky shoal for being close to some ancestor of this feature. We might also consider the possibility that the flat promontory was the one east of the shoal, the warthland (coastal grazing land) of Easton in Gordano, including the long point which became Dumball Island and is now lost in Avonmouth Docks.29 That would certainly have been flat. The shoal is about equidistant from Portishead Point and Easton’s promontory, but there is no evidence that either feature was called ‘the flat ness’, and Portishead Point is not flat. There is no merit in a possible alternative suggestion that the last syllable represents, or includes, OSc or OE ey ‘island’ because there is no feature meriting the description nearby at the right time and Plotneys itself is not an island.
1. Old Scandinavian words are given in their literary Old Norse form, as is conventional.
2. B. G. Charles, Non-Celtic place-names in Wales (London, 1938), 240, 163, 158. Archaeological investigation in Womanby Street, Cardiff, has revealed nothing, according to G. O. Pierce, Place-names in Glamorgan (Cardiff, 2002), 183.
3. R. Coates, ‘The Severn Sea islands in the Anglo-Saxon Chronicle’, in Notes and Queries 61.1 [259 of the continuous series] (March 2014), 1-3.
4. R. Coates, ‘South-West English dumball, dumble, dunball ‘pasture subject to (occasional) tidal flooding’’, in Journal of the English Place-Name Society 39 (2007), 59-72.
5. U.K.H.O., L3895 R. Thomas, Survey of the River Severn … (London, 1815).
6. J. Wright, English dialect dictionary (Oxford, 1898-1905), has an entry for lair, sb.2, meaning ‘mud’, but only as a Scots and North Country word.
7. Bristol Archives (B.A.), AC/D/11/29 lease, 1480; reading uncertain.
8. Information kindly supplied by Brian Austin, Weston-super-Mare, 2016.
9. Some of these names are discussed in greater depth in A.J. Webb and R. Coates, ‘Somerset’s maritime place-names’ in A.J. Webb, ed., A maritime history of Somerset, volume three (2017), 27-82.
10. Williams appears to have executed a survey around 1706 with the title ‘A New and Exact Draught of the Channell of Bristoll from Hartland Poynt to the River Avon & from Caldy Isle to Red Cliff’, which was engraved and published (see A.J. Webb, Charts and surveys of the Somerset coast, c.1350-1824, Somerset Record Society 97 (2014), facsimile 6).
11. Copies of this are held at the B.L., Maps MAR III.79.1; Admiralty Library (A.L.), Vf2/33; National Library of Wales, 6013; Newberry Library, Ayer 135.E55 1766 between pages 42-43 (PrCt).
12. N. Cutler, A general coasting pilot; containing directions for sailing into, and out of, the principal ports and harbours thro’out the known World, with a set of sea-charts, some laid down after Mercator, . . . (London, 1728).
13. Webb, Charts, facsimile 6.
14. Webb, Charts, part two.
15. For details of the Mackenzies, see A.C.F. David, ‘Collins, Mackenzie and the hydrographic surveys of the Somerset coast’, in A.J. Webb, ed., A maritime history of Somerset, volume one: trade and commerce (Taunton, 2010), 93-106; Webb, Charts, 58-78, 144-51.
16. B.L., Maps1068(24); AL, Vu8/1; B.L., Maps1068(27).
17. David, ‘Collins, Mackenzie . . .’, 100.
18. Corporation of Trinity House, 1314 D5 f.59.
19. J. Barlow, The coasting pilot for the Bristol Channel and The Scilly Islands, from Scilly Islands and The Longships to Saint Ann’s Light, Milford, and King-Road (Ilfracombe, 1821), 84-5.
20. U.K.H.O., MP92 remark book of H.M.S. Myrtle, 1815.
21. B.L., Maps 1103(2) Laurie and Whittle’s chart of the Bristol Channel, 1794.
22. United Kingdom Hydrographic Office (U.K.H.O.), B265/3 Steel’s chart of the Irish Sea or St Georges Channel (London, 1800).
23. National Maritime Museum, Greenwich, D8824 W. Heather’s new chart of the Bristol Channel (London, 1803).
24. B.L., Maps 1244(1) Thomas junior’s chart of the lower reaches of the River Severn (London, 1815).
25. B.L., Maps 1103.11 Steel’s new chart of the Bristol Channel (London, 1820).
26. B.L., Maps 1068(29) Laurie’s Chart of the English Channel (London, 1829).
27. Captain E.J. Bedford, Sailing directions for the Bristol Channel (London, 1879), 133-6.
28. In making this point depending on an assumed error involving the first consonant, we should note that Williams’ chart, the earliest document to mention the name, contains three clear errors in the immediate vicinity: Porthet for Porshet, the name of the town; Blackmore Pt for Blacknore Pt; and Kings Road for King Road.
29. Much of this point is now occupied by Royal Portbury Dock.
RICHARD COATES and A.J.W.
コメント