As the major authority in the field, Christina Larner’s analysis of the chronological and geographical pattern of the hunt serves as a vital starting point to any discussion. In Enemies of God, Larner divided the hunt into four basic categories based upon intensity: National hunts (1590–1; 1597; 1629–30; 1649; 1661–62) ‘in which cases came from all over the non-Gaelic-speaking areas and even occasionally from the Gaelic areas of Invernesshire and Ross-shire’; ‘almost national hunts’ in which several regions would become involved; small panics; and isolated cases, in which individuals would be pursued for acts of malefice and not specifically for ‘ideological non-conformity’. The smaller categories were always subsumed into the larger, so that the large national hunts included examples of each.6 Larner illustrated the variation in the intensity of the hunt over time with a graph.7 The most obvious factor for these variations, as noted by Larner, was the ‘rise and fall of the level of official interest’ in witch-hunting displayed by the elite.8
The geographical dimensions of the witch-hunt in Scotland were also explored by Larner. Enemies of God includes a map showing which areas of Scotland had intensive hunts, which had many cases, few cases or no cases.9 The difference in intensity between the various regions of Scotland was discussed. In particular, Larner notes the general lack of witch-hunting in the Highlands:
In the Highlands, especially those parts outside the Kirk session’s, system and within the dominion of the clans there was no witch-hunting, or none that reached the records. Gaelic-speaking areas in general provided very few cases although Tain in Ross-shire was an exception to this . . . Towards the end of the 1661–2 hunt there were several cases in Strathglass in which the landlord used accusations of witchcraft as a means of evicting some unwanted tenants. On the whole, though, Gaelic patronymic names such as those of Mary Nein Goune Baike of Strathglass and Marion Nein Gollimichaell of Tain are rare in lists of suspects.10
This chapter will return to the entire subject of the Highlands and the witch-hunt.
The Scottish witch-hunt, according to both the map and text of Enemies of God, was centred on Fife, the Lothians, the eastern Borders and a small area around Aberdeen. Proximity to the centre of government is suggested as one possible explanation for this.11 Larner also notes that within ‘these general areas there were certain small towns and villages which appear again and again’. Tranent and Prestonpans in particular are singled out as two places ‘which featured in both in the first witch-hunt and in all the major hunts’, and Inverkeithing, Dumfries, and Aberdeen are also given examples of places with long records of involvement in the witch-hunt. Larner concludes:
There seems to be a self-perpetuating element in witch-hunting. Where there were local memories of actual burnings it was relatively easy to stimulate them again.12
The prominence of fishing villages such as Bo’ness, Largs, and Pittenweem in the list of those places where witch-hunting was frequent is noted.13 In assessing this pattern, Larner comments that there is no obvious link on a national level between famine, pestilence, war and other demographic disasters and the incidence of witch-hunting: ‘Whether there may be at a local level is an issue yet to be explored’.14 There is thus no obvious explanation for the witch-hunt which arises out of an examination of this data.
Christina Larner is the only author to have attempted to provide a comprehensive analysis of the witch-hunt in Scotland. It is understandable, therefore, that her interpretation of the geographical and chronological dimensions has dominated the discussion since Enemies of God was published. For example, Ian and Kathleen Whyte have suggested that the prominence of East Lothian in witch-craft cases can be explained because of the ‘close proximity to the justiciary in Edinburgh’. They concur in the notion that once an area had some experience of accusations for witch-craft, it was easier in future for a panic to begin.15 Writing on Fife, Raymond Lamont-Brown argues that in ‘no other place in Scotland were witches hunted with such fervour as in Pittenweem’.16 In his survey of European witchcraft, The Witch-hunt in early modern Europe, Brian Levack has used statistics from Scotland, derived from the Sourcebook, to make the argument that witches were more likely to be executed if tried in the regions than by the central government.17
What has gone largely unnoticed in the statements made since Enemies of God was published, was the tentativeness and limited nature of the discussion within the book itself of the geographic and chronological patterns of the Scottish witch-hunt. Two suggestions for further study in as many pages should alert us to the reality that an author is valiantly exploring new territory – not that extensive research has answered all possible questions.18 Among the simple questions Enemies of God did not address was the variation in witch-hunting between particular shires. Put simply, which shire witnessed the most cases? Which shire had the least? The map in Enemies of God failed to note that Ayrshire was an area with a substantial witch-hunt. Despite the fact that any discussion of the shape and nature of the Scottish witch-hunt must at least begin by dealing with the list of known cases, the Sourcebook of Scottish Witchcraft has remained a virtually untapped source of information and material since the publication of Enemies of God.
One reason for reluctance on the part of historians to use the Sourcebook may have arisen from the cautions and warnings made by both Larner and Bruce Lenman.19 Both raise valid points. The Sourcebook does have many drawbacks. Information is incomplete, infuriatingly inadequate in terms of data on social class and position, irregular, and at times duplicated. It must be used with great caution, particularly in those areas where large quantities of data are unknown, especially in terms of the ‘fate’ of those accused and the number of executions.20 One area where the Sourcebook is quite good, however, is in the area of chronology and geography. There are many instances when the precise date is missing, yet only in 47 cases do we not know the year. Similarly, there are only 105 cases for which we have no geographical information whatever. We can currently position 89% of the cases within a particular shire.21 To begin the re-examination of the pattern of the Scottish witch-hunt it was necessary to re-enter the data contained in the Sourcebook on a computer database, standardize some of the information, and then begin to analyse it. (For a detailed discussion of the Sourcebook, the creation of the Scottish Witch-hunt Database and a discussion on geography and the creation of the maps, see the Appendices.)
Chronology
A re-examination of the chronological patterns of the Scottish witch hunt (See Graph 1) makes it clear the extent to which witch-hunting in Scotland was a seventeenth-century phenomenon. Graph 1 is similar to the one found in Enemies of God (the one important difference will be discussed in a moment). The major hunts, or ‘peaks’, can clearly be seen: 1590; 1597; 1629–30; 1649–50; 1658–59; and, 1661–62. It is important to remember that Graph 1 portrays the number of ‘cases’, not the number of witches. Some individuals appear more than once: for example, Elspeth Thomson, whose story has already been told, is listed twice in the Sourcebook under different spellings. That same year another Elspeth Thompson’s name appears in Banff, again twice.22 The opposite is also true: one case may represent many individual accused witches. References to ‘many witches’, ‘some witches’, ‘a great number of witches’, or some similar phrase, makes any attempt at accurate quantification impossible. Graph 1 attempts, in a minor way, to take account of these ‘multiple cases’ by noting when they occurred and taking some estimate of their significance by assuming that each multiple case involved two additional witches (noted in black on Graph 1). The preponderance of these cases in 1649 is worth noting. The dominance of cases in the seventeenth century is apparent.23
Another problem in trying to portray the Scottish witch-hunt on a graph relates to the 1590s. This decade saw the first major persecution of supposed wi
tches in Scottish history; however, it was not until 1597 that government regulations required a commission before a local witch could be executed.24 The result is that we have no idea how many witches may have been brought before local courts prior to the enforcement of this regulation. The rather mysterious comment that ‘a great number of people’ throughout Scotland were executed for witch-craft in 1597 only adds to the confusion.25 What was meant by a great number? What regions were involved? Christina Larner attempted to take the poor quality of the records into account by noting the estimates in a dotted section above both 1590 and 1597.26 Any estimate is problematic, particularly when we are dealing with such imprecise phrases as a ‘great number’. The difficulty comes in regard to the data graphed in Enemies of God for 1597. The graph shows 200 known cases with another 100 estimated, for a possible total of over 300. The Sourcebook however records only sixty-one cases in 1597.27 An additional twenty-two have been added as a result of the intensive research on Fife.28 That there were cases in this year of which we have as yet no record is certainly true: the issue becomes the size and extent of these cases. It is in the estimate of the extent of the witch-hunt in Scotland that Graph 1 differs significantly from the one found in Enemies of God.29
Graph 1 – Cases of Witchcraft in Scotland, 1560–1730.
Source: Scottish Witch-Hunt Data Base (SWHDB)
Our concern for the major years of persecution distorts our understanding of the period. As the graph makes clear, Scotland was not always in the midst of a great witch-hunt. Indeed, it is the major hunts which stand out as somewhat unusual, arising unexpectedly and being short-lived. Larner suggests that the most obvious ‘immediate causal factor’ for these variations was ‘the rise and fall of the level of official interest in the apprehension and conviction of this particular type of criminal’. As such, the political events with which she associates these events are to be considered as ‘pegs rather than explanations in themselves’.30 Yet, we still need to consider what factors may have contributed to ‘the rise and fall’ of elite interest in persecuting witches.
The timing of the major hunts coincides closely with significant events, some political, some intellectual and social, of which we are already aware. The royal initiative and possible political motives of the events of 1590 and North Berwick kirk have been thoroughly explored.31 The 1629–30 outbreak parallels a major ‘continental witch panic’.32 Persecutions began in 1649 shortly after the radical Presbyterian faction took power in January of that year.33 The increase in interest in 1658 and early 1659 coincided with the collapse of English authority within Scotland, and was followed, after a period of judicial paralysis, by a severe persecution at the time of the Restoration in 1660.34 Certainly other significant political events occurred in Scotland, notably the National Covenant (1638) and Bishop’s Wars (1639–40), with there being no contiguous increase in witch-hunting. Still, the reality is that significant political events coincided with witch persecution. A political event has also been seen as significant in the decline of witch-hunting. The sparseness of cases from the 1650’s is usually explained by the leniency of English judges,35 during the period of Cromwellian occupation. This could have been the result of distrust and an unwillingness on the part of the occupying authorities to strengthen the power of the kirk. Indeed, Geoff Quaife has suggested the occupation marked something of a pause or an interlude between the hunts of 1649–50 and 1658–62.36 The possibility of a link between these political events and witch-hunting needs to be considered carefully.
Geography
When we move from considering the chronology of the Scottish witch-hunt to the geographic distribution of the cases, different issues emerge. Our concern is no longer with attempting to understand the pattern over time, but with discerning why a particular region of Scotland had a higher incidence of witch persecutions than another particular region of Scotland. Table 1 catalogues, in order from highest to lowest, the number of cases in each Shire of Scotland throughout the entire period. It was in Haddington (East Lothian), that the most intense witch-hunting took place. Edinburgh and Fife also witnessed severe hunts. Surprisingly Linlithgow (West Lothian) did not experience such intense persecution.
The varying intensity of the witch-hunt in Scotland is obvious from this information and from Map 2. It comes as no surprise that certain areas produced far more witches than did other areas; however, what must be recognised is that these patterns defy simplistic interpretation. Proximity to the national capital does not seem to have been a factor.37 Linlithgow, closer to the central administration than Aberdeen, Perth, Berwick, Lanark, Ayr, Dumfries and Fife, was not a major centre of the witch-hunt. Kinross, squeezed between Fife and Perth, witnessed only eight cases. Similarly, Kincardine and Selkirk remained minor players, though surrounded by Shires in which the witch-hunt was far more severe.
Map
Shire
no.
1
Haddington
520
2
Edinburgh
325
3
Fife (SBSW)
280
(SWHDB)
(420)
4
Aberdeen
158
5
Ayr
127
6
Berwick
125
7
Perth
118
8
Lanark
116
9
Linlithgow
101
10
Dumfries
99
11
Peebles
79
12
Forfar
74
13
Renfrew
69
14
Ross
65
15
Orkney
50
16
Stirling
48
17
Roxburgh
47
18
Nairn
46
19
Moray
43
20
Bute
42
21
Caithness
35
22
Inverness
35
23
Clackmannan
32
24
Kirkcudbright
32
25
Dumbarton
21
26
Selkirk
21
27
Shetland
19
28
Banff
10
29
Kinross
8
30
Wigtown
8
31
Sutherland
5
32
Argyll
3
33
Kincardine
2
* * *
Unknown
317
other
6
Total
3089
Table 1 – Witchcraft cases in Scotland listed by shire, 1560–1760. Map number refers to the codes on map 2.1. Highland areas are included in the following shires: Ross, Nairn, Moray, Caithness, Inverness, Sutherland, and Argyll.
Source SWHDB; see Appendix B for details.
Although the intensity of the witch-hunt varied from place to place, cases appeared in all of the Shires of Scotland – including the Highland Shires. The Highlands were never a major area of witch persecution. It needs to be stated, however, that certain areas of the Lowlands could make similar claims. These cases also come from a wider area of the Highlands than has been recognised.38 As can be seen by Map 2, cases occurred in different parts of th
e Highlands, including the West coast and the Isle of Skye. So far no cases have been located in the Outer Hebrides. Still, it needs to be recognised that the characterization of the Highlands as an area where the witch-hunt did not occur is far too simplistic. (Nor, if one reads closely, is this what Larner said.)
Map 2 – Scotland, 1560–1760. Witchcraft cases by shire. Numbers refer to the codes as listed in Table 1.
Chronology and Geography
While a separate analysis of chronology and geography is very enlightening it must be remembered that such a separate analysis is somewhat misleading: witchcraft accusations did not occur in time or place, they occurred in a time and a place. This more complex task of combining the when and the where of each case gives us a far more complete picture of the hunt. Within the study of the European witch-hunt, various attempts have been made to combine these two axes, yet no one model or approach has won consensus.39 Nor is a particular approach immediately obvious. With which axis, time or place, do we begin? What unit of time, year or decade, should we use? Are Shires an appropriate level at which to study the phenomenon, or should we move ‘down’ one level to that of parishes? The answer to the latter question is, of course, ‘yes’: parishes are a better level to study the witch-hunt than shires. Unfortunately, the data are such that this is, for the moment at least, impractical for Scotland in its entirety. Choices must be made. The unit of ‘year’ seems most practical, yet even here attempting to represent this within a chapter poses some difficulties, unless one wants to leaf through 200 maps. We will take a brief look at each of these two axes, beginning with the shires of Scotland, in order to develop a clearer sense of how they shed light on the Scottish witch-hunt.
Witches of Fife Page 4
