The Viking Colonization of England in the light of Place-Names 1

"W DO NOT KNOW when the Scandinavians first traveled to the "Western Islands", i. e., the British Isles. Quite certainly Norsemen arrived there in the eighth century, perhaps earlier. The first time we hear about Viking expeditions to the West is 787, by Norwegians, at which time Norwegian Viking fleets attacked several places along Ireland's coast, originating from the Hebrides and other islands north of Scotland where Norwegian colonies had then already been established. In the early part of the ninth century Ireland and the Isle of Man were invaded by increasing hosts of Scandinavian Vikings, and, by this time, ships directly fronl Norway also participated in the warlike expeditions. Soon the Norsemen began to establish themselves in the occupied land. In about 840 a Norwegian king ruled northern Ireland, building strongholds and communities, among them Dublin. In the process the Christian Irish churches were replaced by Scandinavian heathen temples. To what extent Danes took part in these early journeys to Ireland we shall never know, but their numbers were hardly scarce, since the Danes, called by the Irish Danair, seem to have held supremacy over the east coast of Ireland at least for a shorter period of time. But they were soon driven out by the Norwegians, who then established Dublin as the center of a Norwegian colony in Ireland, which lasted until 1170, when the island was conquered by the English. Swedes do not appear to have been involved in the invasion of Ireland to any large extent; instead they were at that time more engaged in expeditions along the rivers of the European continent including present-day Russia. In Ireland the Norsemen annexed first and foremost the coastal districts, and then penetrated farther into the inner parts of the island. They also settled in the southwestern area with Limerick as the center of their dominion.

To what extent Danes took part in these early journeys to Ireland we shall never know, but their numbers were hardly scarce, since the Danes, called by the Irish Danair, seem to have held supremacy over the east coast of Ireland at least for a shorter period of time. But they were soon driven out by the Norwegians, who then established Dublin as the center of a Norwegian colony in Ireland, which lasted until 1170, when the island was conquered by the English. Swedes do not appear to have been involved in the invasion of Ireland to any large extent; instead they were at that time more engaged in expeditions along the rivers of the European continent including present-day Russia.
In Ireland the Norsemen annexed first and foremost the coastal districts, and then penetrated farther into the inner parts of the island. They also settled in the southwestern area with Limerick as the center of their dominion. 1 The following account of the origin and progress of the Scandinavian infiltration in England during the Viking Era and the subsequent centuries (A.D. 800-1200) does not normally include primary documentation of recorded instances of pertinent place-names in ]jIiddle English sources. This material can easily be found in the respective county volumes of the English Place-Name Society's publications and other works referred to in the text. recaptured the whole area south of the Humber, but the Vikings continued their invasion, and in 1013the Danish king Svend Tjugeskjeg ruled almost the entire Danelaw. The hostilities were soon resumed, but now with less success for the invaders, so that in 1042 Edward the Confessor could be elected king in England. The Scandinavian hegemony had been broken although not completely eliminated. Shortly thereafter the Norman invasion under William the Conqueror once and for all put an end to the Scandinavian dominion over England.
The Scandinavian blood in England soon became diluted, when the colonists of both sexes married into the indigenous population. During the twelfth century the Scandinavian nationalities became even less conspicuous through intermarriage with the immigrated Normans, Germans, Flemings and others. During this century we still find Scandinavian names of farmers in the entire Danela w, although children of the upper social ranks were often given Norman names. By the thirteenth century Scandinavian names are to be found only among the lower classes. Thus, the once powerful Scandinavian aristocracy was now reduced to a rather low stratum of society.
Invasions by Norwegians took place also on the western shores of England. In the early part of the tenth century Scandinavians from Ireland, the Isle of Man, and other Norwegian colonies, joined by native Irishmen, raided the northwestern coast of England from Cheshire in the south to the Scottish border in the north. In 901 a Norse invasion and settlement occurred on the Wirral Peninsula and in Chester, which had been established by the Romans and at that time belonged to Wales. The infiltration of this rather small area seems to have been quite thorough. Farther north in about 900, Norwegians started to penetrate Lancashire and Cumberland, a movement which continued through the tenth century and probably also during the first half of the eleventh century. The result was a thorough colonization of these districts. From Cumberland the Scandinavians then invaded and occupied Westmorland and the western parts of Yorkshire.
There were Norwegians in Ireland after 800, and those who now penetrated into northwestern England were probably to a large extent of the second and/or third Norwegian generations in Ireland. They had become more or less Celticized, with their language influenced by the Irish tongue.
But there are reasons to believe, based among other things upon the evidence of place-names, that Danes also settled at least in a small area of Lancashire, and that they perhaps lived there before the Norwegians arrived. It also seems certain that the Danes began a limited colony in southwestern Westmorland, but when this colony originated remains undetermined.
Some scholars have maintained that the Scandinavian infiltration of northwestern England was achieved rather peacefully, without any serious battles, and that it was an immigration rather than an invasion by force. 3 But it would be surprising if this statement were true, since everywhere else the Scandinavian occupation was preceded by fierce fighting. An invasion can hardly be a peaceful enterprise unless the aggressors are so completely superior in number and military strength that resistance becomes futile. The Norwegians were quite certainly not in that position during the early stages of their invasion. The assumption of a peaceful penetration is probably based on the fact that very little information about the Scandinavian aggression in northwestern England is to be found in contemporary local sources. But it should be noted that historical documents from the tenth century concerning this region on the whole are extremely few, and the lack of information about the Norwegian invasions may be accidental.
Thus, the Scandinavian dominion in England was achieved through invasions from two directions, from both the east and the west. The conquest from the east was performed mostly by Danes, joined by considerable numbers of Norwegians and quite a few Swedes. The western colonies were established predominantly by Norwegians and Irishmen.
This extremely brief account of the Scandinavian colonization of England, in which many important events have been omitted, is intended to sketch the most fundamental historical facts as a background to the following exposition. The medieval Anglo-Saxon sources contribute often rather detailed information about the points of time and the places of the Scandinavian descents and how they penetrated into the country. Yet the historical documents and archaeological finds usually do not give an accurate conception of the intensity of the Scandinavian penetration in the various districts. Nor can we derive from these sources any reliable information about the numerical relations among Danish, Norwegian and Swedish colonists. The question whether they simply usurped the old farms and villages or erected new farmhouses and brought new land under cultivation is not answered by these documents. The only way to obtain such knowledge is through analyses and statistical comparisons of place-names and, to a small degree, also of personal names. The value of the latter as criteria, however, must be judged with great caution. The following investigation will then deal almost entirely with the knowledge we can acquire from the place-names, especially the habitational names.
If a place-name contains Scandinavian words which did not exist in the English language before the colonization, it is obvious that the settle-ment was inhabited by Scandinavians. But we must remember that the name of a farm was not given by the colonist himself but by his neighbors in the surrounding area. Therefore a Scandinavian name of a farm or a hamlet indicates with certainty that a Scandinavian settlement existed there.
A great number of purely Scandinavian place-names existed in the Danelawabout 1100, but the frequency of such names varies greatly by districts. By mapping the Scandinavian habitational names we can find out where they were most concentrated and where they were rather few, and on this basis we may draw conclusions about the intensity of the Scandinavian infiltration.
Another circumstance should also be observed. If we find such a placename as N ormannebythere are several of them in the Danela wthis name tells us that, in the area where the so-named farm existed, Norwegians must have been in a minority in comparison with other Scandinavians (and natives). Otherwise the name would not distinguish the farm from those surrounding it. For the same reason the name Daneby informs us about a small number of Danish settlers in a predominantly Norwegian environment.
A few examples of typically Scandinavian names in the Danelaw may be adduced: Oakdale in Yorkshire (now partly Anglicized, Aikedale 1208; ON eik "an oak tree" and dalr "a valley"); Hesket(h) in Chestershire, Lancashire, Cumberland and Yorkshire (Hesteskeith, etc. in early sources; ON hestaskeio "a horse race course"); Ruswarp in Yorkshire (earlier Riswarp; ON hris "shrubs, twigs" and varp "a cast-up heap"); Gamelsby in Cumberland (Gamelesby 1285; ON Gamall, a common personal name and biJr "a farm or hamlet," about which see below); Hillbeck in Westmorland (Hellebec(k) c. 1170; ON hella "a flat rock" and bekkr "a creek"4). However, it must be said that most of the Scandinavian names are far from as transparent as the ones now mentioned. Anglicization, folketymological and other transformations often blurred the Scandinavian provenance quite early. Moreover, a huge number of names cannot be classified with certainty as Scandinavian or English, because they contain words which are similar in both languages, e.g. OE hlip: ON hUo "a slope," OE hol(h): ON hol "a hole," OE tre(w): ON tre "a tree," OE tun "a farm or hamlet": ON tun "an enclosure, a homestead," etc. When such words occur as the second element in compounds and the first element is an Old English word, the name is most likely English, and if the first member is an Old Norse word, the name is probably Scandinavian. A couple of other examples: Ramshead in Lancashire (Ramesheued c. 1204, now lost) contains according to Ekwa1l 5 OE ram "a ram," perhaps used as a personal name, and OE heafod or ON hr;fuo "a hill." This interpretation is possible, but since the name occurs in an area where Scandinavian names are very common, I consider it more likely an ON *Hrafnshpfuo "a raven's head," especially since in the same parish we find Hawkshead (Houkeshout c. 1235), which definitely is an ON *Haukshpfuo "a hawk's head." Ekwall believes that ON haukr here is used as a personal name, but in my opinion both names rather denote formations, stones, crags or the like, that reminded the name-giver of the shape of birds' heads. Such names of comparison are extremely common in Scandinavia and quite frequent in Danelaw districts with strong Viking penetration. In this case one name seems to be a counterpart to the other.
Place-names in England evidence certain phonological and inflectional criteria of Scandinavian provenance. The most important ones are the following.
In certain positions sk had in Old English become a fricative sound, or was at least close to it, before the Scandinavian invasion. But this sound change had not taken place in the Scandinavian languages. Therefore Askwith in Yorkshire (Ascwid 1086 DB, Askewith(e) etc. 1189ff.) seems to be a Scandinavian name. According to EPNS 34, p. 61, it may also be a Scandinavianized form of an OE asc-wudu (or earlier -widu). This is possible, but since the place nomenclature of the pertinent region shows heavy Scandinavian infiltration, it appears most likely that the name is of Old Norse origin.
Another certain criterion of a Scandinavian name is an inflectional form which did not exist in Old English. Most frequent among these forms is the Old Norse genitive ending -ar, which, however, is recognizable mostly in Norwegian names, since in Old Danish, Proto Norse -aR developed into -a or -EE before consonants. Examples: Stangrah in Cumberland (Stangerhovet 1180-1210); the first element is ON stangar, gen. sing.

A88ar Janzen
The Anglo-Saxon genitive ending -e8 corresponds with Old Norse -8, e.g. Braceby in Lincolnshire (Bre(i)zby 1086 DB, containing the personal name ODan. Breth, ONorw. Breior). This distinction is, however, rather rare even in the oldest sources, because in most cases the Old Norse -8 was replaced by -e8 by the scribes.
Another rarely occurring Old Norse case form we find in Thingwall in Cheshire, Lancashire and Yorkshire (Thinguelle etc. 1086 DB etc.), in which the last element is the Old Norse dative form velli of vpllr "a meadow." Several typically Scandinavian second elements in compound habitational names are common in the Danelaw, e.g. byr (in ONorw. also btJr) "a farm or village," jJorp "a secondary settlement," topt, toft "a building site," jJveit "a clearing," 8cetr "a shieling," often difficult to distinguish from setr "a house," bop (ONorw. bM) "a booth."lo By far the most frequent of these second elements is -by, which <?ccurs in all parts of the Danelaw where Viking settlements were established. The word was extremely rare in Old English, and it is therefore very likely that all early names containing -by are Scandinavian. It seems to have been a popular fashion among the Vikings to give the farms and hamlets which they usurped names ending in -by, usually with the new owner's name as the first element. But -by soon was incorporated into the English language, and therefore names with the second element -by, which contain a first element consisting of an English personal name or appellative cannot be considered criteria of Scandinavian settling.
When the Scandinavians conquered land, they usually did not exterminate or drive away the native population; instead the farms and villages passed over into their hands, and the natives were needed as workers or rent paying tenants. These old habitations had, of course, old native names, but the Vikings replaced them with new ones, often using the second element -by. In several cases it can be demonstrated that such alterations took place. The best known example is probably Whitby in northeastern Yorkshire, which earlier was called Streoneshalh. But not all names ending in -by are of this type. When the number of Scandinavians increased, either through natural growth or continued immigration or both, new land had to be cleared and cultivated. The new farms that then arose were also often given names ending in -by. Frequently we find places with pre-Viking names surrounded by groups of farms with Old Norse names, many of which end in -by.ll The second element -thorpe is quite common in the Danelaw. There existed, however, also a cognate OE prop, but in most cases the two words can rather easily be distinguished from each other .12 Since there are many more names ending in -thorpe in the Danela w than in the southern parts of England, it seems probable that most of the thorpe-names in the Danelaw are Scandinavian, even when distinctive old forms are not available. The wordporp was hardly used in Norwegian place-names at the time of the early Viking settlement in England, and we therefore conclude that -thorpe in the Danelaw normally indicates Danish habitations. It is interesting to note that farms with names ending in -thorpe frequently are located out in the periphery of the districts in lowlands or in other less fertile land. These farms are often of rather late origin and denote secondary colonization emanating from already established settlements.
Sometimes the Vikings seem to have substituted porp as they did -by for older English elements, e.g., Tholthorpe in Yorkshire, which was Duruljestune in 972 and Turuljestorp in 1086 DB.13 The elements -sat(e), -set(e), i.e. ON sf£tr or, sometimes perhaps setr, and -thwaite, i. e. ON pveit, occur in names of settlements usually situated in woody areas, obviously denoting secondary colonization.
It may be mentioned in passing that some second elements in placenames, which are common in Scandinavia, are completely or almost missing in the Danelaw, e.g., -vin, -heimr, -staoir. The reason for this fact is that these words had more or less ceased to be productive in name giving when the Viking colonization began. Other second elements are rare or missing, because they had not yet become frequently used in Scandinavia at this time, e.g. ruo, bhle, or because they were little used in districts from where the colonists came, e.g., rum, which is limited almost exclusively to southeastern Sweden. Thus, the Old Norse placenames in England constitute an important auxiliary means in dating certain types of place-names in Scandinavia. They give us, to some extent, termini ante quem and post quem. It should also be noted that Old Norse elements frequently occur in field names, many of which have become habitational names. But because of a lack of old name material, very few field-names are included in the earlier volumes of EPNS. However, when this material is published in future county volumes, it will hardly change the picture of the varying intensity of Scandinavian infiltration we derive from the habitational names.
The Scandinavians also peacefully bought land, which is evidenced, for instance, by the place-name, Oopeland, Cumberland, (Oaupalandia c. 1125), which is ON kaupaland "bought land."14 12 Ibid. 26 Up to now I have considered only purely Scandinavian place-names in England. But there is a huge number of compound names that are so-called hybrid formations, in which one element is English and the other Scandinavian, e. g. Olosegil in Cumberland (OlouisgilI340; ME cloh "a cleft" and ON gil "a ravine"),15 the lost Kerlingdimpil, Kerlingedimpel 1200 in Lancashire (ON kerling "an old woman" and ME dimple "a dip in the surface ofland").16 Such names are of course also important criteria of Scandinavian settling. They seem in most cases to be rather late and not to have been formed until the Scandinavians had adopted English words in their language and when Old Norse words had entered into the English vocabulary. Early naturalized Old Norse words were for instance bekkr "a creek," holmr, holrni "an islet," lundr "a grove." Thus, hybrid formations do not necessarily indicate Scandinavian settlements, and their value as criteria of Viking settlements varies considerably in different districts.
The hybrid formations are, ho,vever, often difficult to distinguish from another type, viz., original Old English names which were Scandinavianized or original Scandinavian names that were Anglicized. It happened quite frequently that Old English names were wholly or partially given a Scandinavian character through substitution of sounds, so that they fitted into the Old Norse sound system. A surprisingly large number of English place-names have still today Nordicized forms. Only one example: initial sk-changed in Old English but not in Old Norse into a fricative sound. Therefore OE Scelton, Sc(h )elton(e) etc. (*Scelj-tun) was by the Vikings pronounced Skelton, and that is the present form of some names in the Danelaw. 17 But Shelton, which is the result of English development without outside interference, also exists in areas where the Scandinavian influence was insignificant or none at all.
But not only were sounds substituted. A whole Anglo-Saxon word could be exchanged for a cognate or similar, more-or-Iess synonymous Old Norse word. In some names ON vior "a wood" replaced OE wudu, e.g. Beckwith in Yorkshire (becwudu c. 972, Becvi 1086 DB).18 Frequently OE stan "a stone or rock" was replaced by ON steinn, or vice versa, a process that may be either sound or word substitution.
Changes such as the ones now mentioned give us the information that, in areas where they are conspicuously numerous, the Scandinavian population must have been considerable in number and socially influential.
But it often happened that the Scandinavianized forms were later re-Anglicized, and we can discern the early Nordic influences only in old sources.
After the Scandinavian colonization had ceased, it often happened that original Old Norse names were Anglicized through substitution of English sounds or whole words. ODan. flask "a swamp" often occurs in the ME forms flasshe, flosshe, which may be due to substitution of ME sh for ON sk or perhaps to influence of the synonymous French flache. 19 ON N yb01e "the new homestead" was changed into N ewball. 20 Also in such names only old sources can reveal the Scandinavian origin.
The amalgamation of the English and Nordic languages that developed in the Danelaw soon became so sweeping that the original distinction between Scandinavian and English place-names to a large extent disappeared, and in numerous cases it cannot be determined whether a given name is English or Scandinavian. Then the original forms can be established only if the names are attested in sources that were written before the intermingling of the two languages reached a thorough-going stage. Fortunately, we are in possession of a priceless document in which the habitational names in the Danelaw are listed, viz., the remarkable real estate register which is known as the Domesday Book. The precursory survey was made by order of William the Conqueror and was completed in 1086, the year usually given as the date of the manuscript. There is, however, reason to believe that the existing two Exchequer volumes were composed considerably, perhaps as much as half a century, later. 21 In this source all names of farms and hamlets and their owners in the major part of England are recorded. All names ending in -by and most of those ending in -thorpe are accounted for, which is especially important, because these names represent the earliest phase of the Viking colonization of the Danelaw. But at the same time there are few names containing -jJweit and -toft and not one with -vra "a nook, a corner of land" and -scale (ON skali "a hut"), which are common in later Scandinavian names in England. Obviously, such names, and the farms they denote, originated after 1100. Names of lesser habitations are usually not recorded until later, when the fusion of languages had been in process for a shorter or longer period of time. It is important that in the Domesday Book the Scandinavian influence on the phonology of the English place-names and the English influence on the Scandinavian ones is rather slight, which enables us to establish the provenance of the names.
As was earlier indicated, the place-~ames offer the only reliable means for acquiring knowledge of the intensity of the Scandinavian penetration in the various districts of the Danelaw. In the section of this paper that follows an attempt will be made to use the place-names as a basis for such an estimation. Only the habitational names will be considered, since the field-names adduced in the published county volumes of English placenames are rather few in the earlier volumes while numerous in the most recent ones. In the beginning of this survey I do not distinguish between Danish and Norwegian names, because it would probably disturb the general impression of the scope and depth of the infiltration. But at its conclusion I shall consider the possibility of separating Danish from Norwegian names. The following survey will be extremely brief and will present only the major results of the investigation, but it is based on all criteria now avail able, even if they are not accounted for or referred to. I shall begin with the eastern part of England and proceed from the south to the north. Thereafter the infiltration of the northwestern counties will be discussed.
In the place-names of the counties in the Southeast, that is in the surroundings of London, there are very few traces of Scandinavian settling. In Essex, north of London, Scandinavian names are extremely rare, but in the coastal area an earlier existence of a Viking colony may be discerned. West of Essex the Scandinavian elements in the place-nomenclature are hardly worthy of mention. But we know from historical sources that the Vikings invaded the counties northwest of Essex, viz., Cambridgeshire, Huntingdonshire and Bedfordshire, although Scandinavian place-names are now practically non-existent there. It would appear that the Vikings did not settle permanently in this area. In the next county to the west, Northamptonshire, we find appreciable traces of Scandinavian colonization only in the northern parts, where there are about a dozen names containing -by and half as many -thorpe. The placenames in the tiny county of Rutlandshire show that Scandinavian colonists existed here, but not in large numbers.
It is obvious that, in the major part of the southeastern area now surveyed, the Viking penetration was very light, even if the Danes were for some time the rulers, evidenced by the fact that they changed the laws after their standards.
Proceeding from Essex northwards, we do not find any notable degree of Scandinavian penetration in the county of Suffolk, but in Norfolk we observe quite a different situation. About 500 habitational names are listed in the Domesday Book from this county, and at least 200 of them are Scandinavian. Most of the latter names are located in the eastern part of the county, while there are rather few in the interior region. According to historical sources the Vikings became firmly settled in East Anglia, i.e., Suffolk and Norfolk. The place-names reveal that the center of this colonization was the eastern part of Norfolk.
Farther north we come to districts where the Scandinavian infiltration must have been very thorough. In Leicestershire the density of Scandinavian place-names varies considerably in different areas. The concentration is heaviest east and northeast of Leicester, where a fertile farm land extends. There are 58 names ending in -by, while thorpe names are few, which indicates that secondary settlement in this area did not develop on a large scale. The heavy concentration of Scandinavian place-names in some parts of Leicestershire is remarkable, since this county undoubtedly was densely populated before the Vikings invaded it. Probably Scandinavian names replaced the old native ones to a large extent.
In the southern part of Lincolnshire, Scandinavian elements among the place-names are not particularly conspicuous, but in the northern half of the county we find the Scandinavian penetration reaching an extremely high degree. Purely Nordic names are in certain districts unbelievably frequent, even outnumbering the English names. Hybrid formations, in which usually the first element is Scandinavian and the second English, are also common. Thus, if one judges from the place-names, it looks as though in some places there lived more Northerners than native people. There are no fewer than 260 names ending in -by and rather many in -thorpe with Scandinavian personal nanles as the first element, indicating substantial secondary settlement. The thorpe names occur mostly in the coastal lowlands, where by names are rare. Apparently the less fertile low-lying land was not so attractive to the early invaders, who settled in the already cultivated regions. The local distribution of the place-names also appears to indicate that the invasion in this area came not as much from the coast as from the south and from York in the north. The Vikings penetrated into the central parts of Lincolnshire, whence the Roman military roads branched out in all directions. There seems to be a marked connection between these roads and the frequency of Scandinavian placenames. The names ending in -by are located especially along these roads and on the land between them.
In the county of Nottinghamshire, west of Lincolnshire, the placenames inform us that the Viking infiltration was rather strong, although not at all as dominant as in northern Lincolnshire: there are only some 20 names containing -by. Even if Nottinghamshire is only half as large as Lincolnshire, the difference is still striking. A far from small number of names are hybrid formations.
But especially in Nottinghamshire and adjacent counties we have to consider another type of place-name as a criterion of Scandinavian settling. In Nottinghamshire there are about 20 names, all recorded in the Domesday Book, with the second element -ton « OE -tun or ON -tun "an enclosure, a farmstead") and Old Norse personal names as the first element, e.g. Asloekton (Aslaehetune 1086 DB; ON Aslakr), Thurgarton (Turgarstune 1086 DB; ON porgeirr, in the Anglicized form purgar). These names present a special problem: is the second element English or Scandinavian 1 This is not possible to determine with certainty, but when we realize that -ton is by far the most common element in English place-names and that -tun is rare in Danish place-names (the colonists in this area were mostly Danes), it seems most likely that these names originally were native English names of established farms and hamlets with an Old English personal name as the first member but that the Danes substituted the names of the new Danish owners for those of the dispossessed English farmers. 22 Therefore, these hybrid names in -ton with a Scandinavian personal name as the first element ought to be considered equivalent to the names ending in -by, when we try to determine the intensity of the early Viking infiltration. These names also give us an idea of how the Norsemen took possession of the old farms.
The distribution of the Nordic place-names in Nottinghamshire shows on the whole that the Scandinavian colonization was spread rather evenly within the county.
In Derbyshire, west of Nottinghamshire, Scandinavian place-name elements are considerably less frequent. We find only ten names ending in -by, to which may be added a dozen hybrid names ending in -ton with an Old Norse personal name as the first member. But in the southern part of the county, close to the border towards Leicestershire, there seems to have lived a more compact Scandinavian population.
In the huge county of Yorkshire the Scandinavian infiltration in some areas appears to have been as thorough as in northern Lincolnshire. Yorkshire was early divided by the Viking rulers into three thrithings, now called ridings, viz., the East Riding, West Riding, and North Riding. This system of dividing districts into thrithings existed in some parts of Scandinavia. 23 f:Ĩ .tJ The Nordic colonists also introduced the division into so-called wapen takes, a clearly Scandinavian word, replacing the old division into hundreds. Furthermore, the Vikings carried another reform into effect, viz., the establishment in some areas of unities denoted by a word attested in Old Swedish as byalagh "the community of a village." The name Bierlow « ON By(j)ar19g) bears evidence of this reform. Also the division into so-called carucates (areas of 60 to 160 acres) as substitute for the old hides was chiefly established by the Vikings. Wapentakes and carucates were 22 See EPNS 17, p. 18f.; A. H. Smith in EPNS 26,p.192f. 23 Riding is the ON prilJjungr "a third." As the second element in compounds with the three names of the cardinal points as the first member the initial pwas fused with the final consonant of the first element. From these compounds the form riding was abstracted. The modern form ought to be *ridding, but the word was adopted in later times from Middle English and pronounced according to modern phonology. introduced not only in Yorkshire but in most of the Danelaw, but thrithings and "byalagh" remain characteristic of Yorkshire. That the Scandinavian infiltration and influence was so radical in Yorkshire is hardly surprising, since York was for quite some time the administrative center of the whole Danelaw.
The Scandinavians dispersed into most parts of Yorkshire, but the heaviest permanent infiltration took place in the whole East Riding and the eastern parts of the West and North Riding. Here we find large numbers of names containing -by and -thorpe. Of the names recorded in the Domesday Book from the East Riding, two out of five are Scandinavian, and, if we consider also names of lesser habitations and field names, the Nordic penetration stands out even more strongly. In certain areas the Norse names are more frequent than the native ones. In the Whitby district of the North Riding the concentration of Scandinavian names is more dominant than anywhere else in England. Almost all habitational names within this area recorded in sources from the eleventh through thirteenth century are Scandinavian.
I cannot refrain from mentioning in passing a detail which is of no importance for the over-all picture of the Scandinavian influence in Yorkshire, but which is very interesting nevertheless. There are some incidents when a field-name ending in -how, i. e. ON -haugr "a hill, a mound" contains the same personal name as the first element of the name of the farm where the mound is located. Thus, a lost name Leggeshou denoted a mound on a farm named Legsby (Lagesbi 1086 DB, Leggesbi 1202) and the likewise lost Katehou was the name of a similar formation on the farm Oadeby. It hardly involves a great piece of audacity to assume that the first Scandina vian owners of these farms were buried in the mounds, because it must have taken place while the colonists were still heathen, and they were Christianized very early. Thus, such field-names belong to the earliest stratum of Norse place-names in England, during the early decades of the tenth century.
From Yorkshire the Scandinavian invaders continued northwards into Durham and Northumberland, as well as southwards into Lincolnshire, where they were joined by Vikings who came directly from their homelands. In Durham we find a considerable number of Scandinavian names only in the southern part, but not to the same extent as in Yorkshire. In Northumberland there are rather few, which is a surprising fact, since we know that the coast of this county suffered extensive raids, and that Scandinavians ruled there for some time. 24 But evidently the invaders did not settle in this district in large numbers, and we may conclude that the attacks were mostly forays. This brief survey of Scandinavian place-names in eastern England has yielded the information that the Viking penetration was heaviest within an area comprising the western half of Lincolnshire, the northeastern part of Leicestershire, the eastern parts of North and West Riding and the whole East Riding of Yorkshire. Now we turn our attention to northwestern England, where the Scandinavian intruders, at least during the early stages of the colonization, were accompanied by Irishmen. As was stated above in the historical introduction, Scandinavian settlements existed in Cheshire as early as the beginning of the tenth century. But no influence worthy of mention on the place-nomenclature of this county can be found except on the Wirral Peninsula, which projects from Chester.
In Lancashire, north of Cheshire, on the other hand, pure and hybrid Scandinavian place-names become quite frequent, especially in the coastal districts, and we discover regions with particularly strong Scandinavian infiltration.
But the heaviest Scandinavian penetration of northwestern England took place in Cumberland; here we find some centers with strongly concentrated Viking colonization, where Old Norse place-names became at an early time almost as numerous as the native ones.
From Cumberland the Norse and Irish colonists spread into the mountainous and woody Westmorland, where they settled in the western half of the county. In the so-called Lake District the English names of older habitations may be fewer than the Scandinavian ones. Some of the intruders continued also into at least the westernmost parts of the West and North Riding of Yorkshire; see more about this below. The map is intended to show the concentration of the Scandinavian infiltration in these various parts of the Danelaw. Now we ought to consider to what extent the place-names can give information about the local distribution of Danes and Norwegians in the colonized areas. We know that the majority of the Vikings landing on the east coast were Danes and on the west coast, Norwegians, but many Norwegians took part in the colonization of eastern England, and we shall find Danes in the Northwest as well. The numerical relationship cannot be established through historical and archaeological material, but the place-names give us at least some elucidation of this problem.
Perhaps we should first pay some attention to the fact that there were also Swedes among the Scandinavian colonists. But unfortunately the place-names from such an early time as the Viking Era do not allow us to distinguish between Swedish and Danish elements. Only personal names can sometimes give us knowledge about the nationality of the settlers.
Most Old Norse personal names in the Domesday Book are common to all Scandinavian countries, but some are purely Norwegian or purely tfo rlolt<;l& a.n. iIW7:"> '" "" Ii0 I Danish, while there are also some which are exclusively or chiefly Swedish. We find in the Domesday Book and in a couple of later sources, for instance, two persons from Norfolk named Stergar (Sterger, Stirgaro, ablative), which according to von Feilitzen 25 is the OSwed. Styrger. No ONorw. *Styrgeirr is known. ODan. Styrger has been recorded only twice. 26 Since the name seems to have been used more frequently in Sweden, there is a better chance that the Stergars in Norfolk were Swedes rather than Danes. The Domesday Book also lists a person named Bosteinnus, who quite certainly had the OSwed. name Bosten. 27 The name is unknown in Denmark and Norway. A farm in the northernmost part of Yorkshire is called Tidkinhow, pronounced tiTjkin'Ju and attested as Tidkinhowe 1575. The last member is undoubtedly ON haugr "a hill, a mound." According to EPNS 5, p. 148 the first element is "a late petfornl of some such personal name as OE Tydi, Tidi." I have, for various reasons, declined this interpretation and suggested that the place-name contains the OSwed.
personal name Tiftkumi 28 For the change m > n, cf. for instance, Arnegrin « ON Arngrimr), Ulgrin « ON Uljgrimr) in Domesday Book. The unstressed u developed into ME e and further to i, either through vowel assimilation or in association with the English noun kin. If my interpretation is correct, the place-name is extremely interesting, because Tiftkumi seems to have been used only in the Malar district in central Sweden around the present capital of Stockholm, and we would then in this case be able to identify a Swedish settler from a very limited area.
In Lincolnshire there are three farms or villages called Hedebi. It is tempting to believe that these places were named after the Swedish colony of Hedeby in southern Jutland, which existed during the first half of the tenth century. If that is true, the owners were most likely Swedes. Such scattered names as these give us, however, not the slightest knowledge about the strength of the Swedish contribution to the Scandinavian colonization of England.
There are some criteria which make it possible to distinguish between Danish and Norwegian place-names in England. The following are the most important.
1. The Scandinavian place-names in England sometimes contain words that were exclusively West-Scandinavian and therefore did not belong to the Danish (or Swedish) vocabulary, or were at least very rare, at the time when the Danelaw was colonized. Such words are skdJi, m. "a hut," in Middle English sources usually written scale, which occurs in a large number of place-names, chiefly in the Northwest, which is to be expected, since the Norwegians were there in a majority, but also in the eastern districts, where it indicates Norwegian settlements; gil, n. "a ravine," which also frequently occurs in the Northwest, but hardly in the East; ergh "a seieling, a hill-pasture," borrowed from Irish or Gaelic (Middle Irish airge, Gaelic airigh), also common in the northwestern counties but later spreading into more eastern and southern districts, probably at least partly from York. 29 2. If the first element of compound names in old sources ends in -ar before an initial consonant of the second element, as was mentioned above, the ending indicates in most cases West Scandinavian origin.
3. Assimilations of the consonant combinations mp, nk, nt to pp, kk, tt are essentially West-Scandinavian, e.g. ONorw. brekka: ODan. brink "a slope" (for instance in Breck in Lancashire, the West and North Riding of Yorkshire, but Mickelbring, Bowbrings in Lancashire and the West Riding of Yorkshire).
4. Such vocalic differences as ONorw. 11, corresponding with ODan. 0, for instance ONorw. bUb "a booth" (in English sources normally written buth, bouth) but ODan. both (in English sources spelled both) and ONorw. o corresponding with ODan. u, for instance in ONorw. holmr "a small island," but ODan. hulm. 5. As was intimated above, the majority of names ending in -thorpe denote Danish (or Swedish) settlements, since porp is extremely common in East-Scandinavian place-names while in Norway it appears chiefly in the southeastern districts and even there is hardly a common element. Therefore, in the Danelaw areas, where thorpe names are very frequent, we have reason to believe that Danish settlers were in a majority. 6. On the whole, it seems that the extremely common second element -by in most cases signifies Danish settlements. But caution is necessary in separate instances, because in the first place byr existed also in Old Norwegian and occurs in Norwegian place-names, and secondly, the synonymous ONorw. b~r also might be the origin of -bi in English sources. In main-stressed syllables bhr would be represented by English *be, but as a weakly stressed second element in compounds it could develop into -bi, which demonstrably happened in Scotland. It is a fact that placenames ending in -by occur in particularly high number in areas that for other reasons are characterized by heavy Danish infiltration. 7. The difference between ONorw. topt and ODan. toft "a building site" can only to some extent be used as a criterion for distinguishing between 29 About the occurrence of these words in English place-names see, in the order in which the words were mentioned, A. H. Smith in EPNS 26, p. 123; 25, pp. 200, 157.
Danish and Norwegian place-names. But it is certain that the word, regardless of the phonological difference, was used in place-names more often in Denlnark and southern Sweden than in Norway and the Norwegian colonies in Orkney, Shetland and the Faroes. In English sources the form is almost exclusively toft. The word is contained in place-names chiefly in the eastern parts of the Danelaw, but it is infrequent in the Northwest. Thus, it seems quite certain that the majority of place-names containing toft indicate Danish rather than Norwegian settlement. But since toft was adopted into the English language as a legal term, names ending in -toft that are not recorded in very early sources may very well be English. 8. A rather common second element in Scandinavian place-names in England is -thwaite, i. e. ONorw. pveit, ODan. thwet "a clearing," but this word is hardly usable as a certain criterion of Danish or Norwegian habitation. We may, however, observe that this element is by far most frequently used in the Northwest, while it is rare in the eastern Danelaw counties. There are well over 200 in the northwest but just a few in the remainder of the Danelaw. Therefore it is possible that at least some of the thwaite names in the eastern districts indicate Norwegian settlement.
If we take into consideration the criteria now discussed and some others of less importance and not mentioned here, the place-names indicate that among the Scandinavian settlers in eastern England the Danes were in an overwhelming majority. We find considerable Norwegian elements only in the northeast in the Whitby area. Thus it seems that the Norwegian fleets that sailed to England in the beginning of the tenth century landed mostly in the Whitby district. In and around York the Norwegian population was also rather dense, although not predominant. The reason for this situation is undoubtedly to be found in the lively communication between York and Dublin. Obviously many influential Norwegians moved from Ireland to York, since Norwegian Viking kings ruled York from 915 to 950.
The place-names tell us that smaller Norwegian colonies existed in the eastern parts of the Danelaw. In Yorkshire and Lincolnshire, where the Scandinavian infiltration was most thorough, we find the name N ormanby, i.e., ON NoromannabiJr,16 times. In the eastern counties there are also several N ormantons i. e. ON N oromannatun, in which the last element may be Old Norse or originally English; see above. In the areas around these places, the Danish majority must have been strong, since such names would not have been distinctive in areas where the Norwegians were in a majority.
We also find quite a few names with Ire-as the first member, e.g.
irabyr, iratun. irar (plur.) designated not only Irishmen but also Nor-wegians that came from Ireland, and it is possible that some of these names were given because the new owners were Celticized Norwegians. It is interesting to observe that some of these names are to be found in the easternmost parts of Yorkshire. Did these invaders, Irishmen or Norwegians, cross northern England from the west coast, or did they sail around Scotland to land in Yorkshire 1 We shall soon try to find out if the place-names can give us any help in answering this question.
It should be noted also that some scattered Danebi, Denebi exist in the eastern counties. These names inform us that in these particular areas the Norwegians were in a majority.
The geographical distribution of the Scandinavian place-names in England indicates that the Danes were the first who colonized eastern England, sequestering the most fertile land of the plains and along the rivers. In such districts Norwegian names are few. Apparently the Norwegians who later invaded eastern England usually avoided the areas which were already held by Danes.
In northwestern England the Scandinavian place-names are to a large extent of West Scandinavian types, which of course is to be expected. We find large numbers of names containing such Western words as fell (fjall) "a mountain," foss "a waterfall," slakki "a valley," gil "a ravine," nes "a promontory," skltli "a hut," tjprn "tarn," holmr "small island." The first colonists came from Ireland, the Isle of Man, the Hebrides and Scotland. The place-names often reflect the Celticized language spoken by these Norwegians, but later Norwegians from Norway joined their Celticized fellow-countrymen. The place-names clearly show that the most thoroughly infiltrated areas were Cumberland, western Westmorland and some coastal districts of Lancashire. But even if Norwegian and Irish colonists dominated this region they were not in sole control of the land. In the southernmost part of Lancashire, south and southwest of Manchester, we find evidence of a strong Danish colony.30 Quite a few Danes seem to have settled also west of Kendall in southern Westmorland, because we find there a group of five names ending in -thorpe. 31 In Cumberland one finds very few names, if any at all, that necessarily or probably are Danish.
As was intimated above, the language of the Norwegians coming from Ireland showed obvious Irish elements. The most remarkable of these Irish features was a type of compound word which was alien to the Germanic languages, viz., so-called inverted compounds, in which the order of the elements is inverted compared with the Germanic system: Dun Patrie "Patrie's fort," Beesbari ea. 1200 "Snari's creek," Toftar-asmind 30 Ekwall,Lancashire,p. 245ff. 31 Ibid.,p. 247. thirteenth century "Asmund's lots," Rudswain ca. 1290,probably "Svain's clearing." An interesting name is Bek Troyte from the eleventh century, which recurs as Truttebeek in the fourteenth century. It seems, then, that the Irish compound type disappeared after the Norwegians had lived in England for some time.
Such Celticized place-names in northwestern England are extremely interesting and important for solving problems here discussed. They are fossils that inform us how the Norse language was strongly influenced by the Irish milieu at a time when our knowledge of the destiny of the Scandina vian languages in foreign countries is practically non -existent.
This type of compound is an Irish innovation, but there are equivalent or analogous phenomena in other Indo-European languages, e.g., Latin Oampus Marti(~t)s "the Field of Mars," the name of a large area northwest of the ancient city of Rome. However, in the Scandinavian placenames the last element always appears in uninflected form. Apparently the Norsemen did not comprehend the difference between the Irish case forms, or, if they did, they did not need the genitive ending in the compounds, because the placement of the second element was in itself a sufficient grammatical expression of the genitival function. The inverted compounds were used by the Scandinavians especially when the modifying element was a personal name and particularly when the compound word was used as a place-name. As is evident from a couple of compounds among the above examples, the Norwegians formed compounds according to the Irish method even when both elements were purely Old Norse words.
In Cumberland we recognize around 70 place-names of the inverted type, and nine in Lancashire. In Westmorland around 70 can be identified, almost all in the western half of the county, west of Maiden Way, which runs in a north -south direction; only a couple of certain incidents occur , in the eastern parts. 32 In the western half there are also quite a few names containing the Irish or Gaelic loan word ergh, but east of Maiden Way we do not find a single example. In the eastern part of the North Riding of Yorkshire can be found six ergh names, in the East Riding, five, and in Lancashire, two.
It has been said that the Scandinavian place-nomenclature in the eastern region of Yorkshire shows a conspicuous similarity to that in the Lake District in the West. Because of this similarity, all who have cared to express an opinion in this matter, even Ekwall, who was the first to point out the occurrence and importance of the inverted compounds in Scandinavian place-names in England, have claimed 33 that the Norse-men from Cumberland forced their way into Westmorland and continued into and through Yorkshire all the way to the eastern region of the country, leaving only a few settlers in the Western areas. This conclusion is partly based on the alleged occurrence of inverted compounds and on the existence of some ergh names in Yorkshire. If this opinion is correct, it would indicate a very strong invasion from Westmorland into Yorkshire, although historical sources seem to give no direct information about the matter.
When one looks at the map of the land concerned, he may be a little skeptical of this generally accepted explanation of the Norwegian migration into eastern Yorkshire. At least I am. Westmorland is, especially in the eastern parts, a mountainous and woody land, which at the time of the Scandinavian invasion must have been very sparsely populated. It is separated from Yorkshire by the high Pennine Divide, a mountain ridge around 800 meters or 2,600 feet above sea level, which must have been rather difficult to cross. If the Norwegians really penetrated from Westmorland into Yorkshire in great numbers, it is remarkable that there are practically no inverted compounds and no ergh names in the eastern half of Westmorland. I would, therefore, conclude that comparatively few Norwegians settled in this region.
But scholars have found a few inverted compounds in Yorkshire, and ergh names definitely occur in the eastern parts of the North and East Riding. However, Ekwall has shown that one of the alleged Irish compounds in Yorkshire is an English name with the normal Germanic order of the elements. 34 I believe that I have proven that all the other names in Yorkshire which have been explained as Irish compounds are purely Scandinavian names with normal word order. 35 Thus, there are no inverted compounds in the place-nomenclature of Yorkshire. But what about the ergh names 1If they originated with the Norwegians from Westmorland, it would be surprising that the invaders left so few traces in the large region between western Westmorland and eastern Yorkshire. Actually, the occurrence of ergh in the East is no certain criterion of settlement by Celticized Norwegians. It is more likely that the word was brought from Ireland to York, from where it could rapidly spread into the surrounding countryside. A good parallel to such a phenomenon is the Irish word cross, which soon became a common word throughout the Danelaw and therefore does not indicate Norwegian penetration. Strangely enough, neither ergh nor cross occurs in western place-names that consist of inverted compounds.
In the westernmost parts of Yorkshire we find some place-names containing such typically Norwegian words as gil, skali, brekka, even ergh, 34 Ekwall, in Studies tillagnade Axel Kock,p. 218f. 35 Janzen in Namn och bygd 1960, p. 43:ff.;cf. EPNS 36, p. 53. proving that the Norwegians to some extent must have crossed the Pennine Divide. Also the western area of Durham and perhaps Northumberland was reached by Norwegians from the west.
In my opinion, the theory of a strong Norwegian invasion into eastern Yorkshire from the west over the Pennine Divide is based on dubious or wrong presumptions, although I do not deny the possibility that such a migration might have occurred on a small scale. The Norwegian settlers in eastern Yorkshire arrived there chiefly from three directions: 1) directly from Norway; 2) by sailing from Ireland and rounding Caithness in Scotland, and 3) through transplantation from York. Some ergh names in eastern Yorkshire could have been given by Norwegians coming from Ireland.
In earlier times the settlements and cultural currents proceeded along river valleys, not over mountains or through forests. This is the way it happened in Scandinavia, and there is hardly any reason to believe that the Scandinavians in England changed their habits and penetrated in huge numbers through very rugged terrain.
A complete and accurate investigation of the problems discussed here cannot be conducted until the place-names of all counties have been published and the older county volumes, especially that of East Riding of Yorkshire, have been rewritten on the basis of the new rich source material which was not available in earlier days.
It is not possible to give a reliable chronology of the Scandinavian place-names in England in the same way as one can in Scandinavia, since the colonization of the Danelaw lasted only a couple of centuries. But a few observations may be made. The large settlements with Scandinavian names are normally older than the minor ones. Also farms on the fertile plains and valleys were annexed before secondary settling took place in the less fertile areas, although in this conclusion we are hardly helped by the place-names. We may, however, as was earlier said, make a clear distinction between the older names containing -by and the younger ones containing -thorpe.
As to the place-names in northwestern England, we may assume that the inverted compounds can hardly have been common until the Old Norse language in Ireland had been rather strongly Celticized. Thus, if we proceed with great caution, we may adhere to the hypothesis that regions that have few or no inverted compounds, for instance Cumberland, might have been settled earlier than areas where the inverted place-names are frequent.
In isolated cases it is possible, within certain limits, to date some placenames. Just one example: in Lancashire, Stanraysinum, name of a place just north of Lancaster, occurs in a manuscript from 1425-1450. As Ek-wall suggests,36this name must be ON Steinhreysinum, dative plural of stein-hreysi "a cairn." The definite article seems to have been used in place-names in Norway and Iceland not until ca. 1100. This name, then, gives us a few interesting pieces of information, namely that Norwegian colonization in Lancashire was still going on about 1100, that the settlers probably came directly from Norway, and of course that the Norwegian language was spoken in Lancashire about 1100.
The Scandinavians that came to Ireland and England during the early period of colonization were heathens, while the native population remained Christian. The interesting question presents itself, whether any Scandinavian place-names show evidence of pagan Scandinavian worship in England. The answer is "extremely few." Alexander Bugge 37 adduces some Scandinavian place-names in England, which in his opinion contain names of pagan Scandinavian gods. But none of them is self-evident. A few examples: in Torweswe in Lincolnshire the last member is ON ve "a temple," but the first element is probably not, as Bugge believes, the name of the god parr, but rather the common Old Norse personal name parir. Thus the place-name hardly means "the temple consecrated to the worship of Thor," but more likely "Thorer's temple." But neverless the name must refer to a heathen sanctuary. The same may be said about Hovby, probably containing ON hof "a temple" and meaning "the farm or village with a temple," but hof could also be English and refer to a Christian temple. Othenesberg exists in two places, in Yorkshire and in Norfolk, According to Bugge this name contains the Old Norse name Ooinn, but, as Ekwall has suggested, earlier forms with Ou-rather point to the personal name ON Auaunn. Grimesarg undoubtedly contains ON hprgr, OSwed. hargher, which in place-names denotes a heap of stones, usually but not necessarily serving as a sacrificial altar. The name could simply mean "Grim's heap of stones." The impression we get from a survey of the place-names in the Danelaw is that the population about the year 1100 was extremely heterogeneous. The indigenous Celtic people had earlier been intermingled with Angles and Saxons, and during the Viking Era Scandinavians and more Celts infiltrated the land more or less thoroughly, and soon thereafter the Normans, Flemings, Germans and other nationalities immigrated in large numbers. From a racial point of view the English people might be the most heterogeneous within the European territory. The place-names offer a silent but conspicuous evidence in support of this conclusion.