Family Names

Family Names  
Patrick Hanks and Harry Parkin
The Oxford Handbook of Names and Naming
Edited by Carole Hough

Print Publication Date: Jan 2016 Subject: Linguistics, Semantics, Pragmatics

Online Publication Date: Mar 2016 DOI: 10.1093/oxfordhb/9780199656431.013.9

Abstract and Keywords

This chapter begins with an introduction to the world’s personal naming systems, leading
into an account of the development of hereditary surnames in England, Wales, Scotland,
and Ireland, and the effects of international migration on local name stocks. Widespread
popular and scholarly interest in the history of family names is then highlighted in an ac­
count of the major family name dictionaries. The importance of regional and localized
study, for our understanding of regional variation in surname patterns, is stressed, and
such analysis is suggested as being the next logical step following the work of the English
Surnames Series. Multidisciplinary approaches to surname study are also introduced,
through a summary of work on DNA and surname inheritance, demography, and local his­
tory. Finally, a case is made for the future creation and statistical study of large surname
databases in order to advance our knowledge of surname changes, development, and dis­
tribution.

Keywords: family names, naming systems, surnames, European surnames, world surnames, geographical distribu­
tion, FaNBI, dictionaries

15.1 Origins of Hereditary Surnames

15.1.1 Personal Naming Systems of the World

Every human society has a naming system for identifying individuals within it (see chap­
ter 13). This normally consists of one or more given names (see chapter 14) and an addi­
tional name whose function is to identify the individual as a member of a family within so­
ciety. With very few exceptions, there are just three such systems of personal naming
throughout the world: the patronymic system, the binomial system, and the Arabic sys­
tem. The focus of this chapter will be on family names within the binomial system, but
first we give a brief account of the other two systems, both of which have contributed to
the development of family names within the binomial system in the English-speaking
world and in other European languages.

Page 1 of 24
Family Names

In the Arabic personal naming system a person’s name comprises up to five elements.
These are: kunya (a kind of aspirational nickname, for example Abu-Fazl ‘father of bounty’
and Umm-Abdullah ‘mother of Abdullah’, which could be adopted regardless of whether
any child called Abdullah actually exists), ism (given name), nasab (patronymic), nisba
(locative name), and laqab (distinguishing nickname such as al-Aswad ‘the Black’). Kunya,
nasab, nisba, and laqab have all been adopted as ‘surnames’ among people from the Is­
lamic world who have migrated to English-speaking countries and to other countries
where the binomial system of personal naming is prevalent.

While this is the standard system of Arabic personal naming, used throughout the Islamic
world, there is much variation in different countries, with the different elements being
used in different ways (for more information see Schimmel 1989; Ahmed 1999; Roochnik
and Ahmed 2003). One of these differences concerns the use of fixed family names.
Ahmed (1999) comments:

In some Muslim countries, e.g. Egypt, Iran and Turkey, family names are well es­
tablished, but in the Indian subcontinent a complete liberty in selecting names
means (p. 215) that there is no necessary continuation of the surname from father
to son. Also, there is little distinction between a surname and first name and they
are freely interchanged.

(Ahmed 1999: xiii)

The patronymic system was once the norm throughout most of Europe. People were
named according to their parentage, so that along with a given name, they would be iden­
tified by the given name of their father and very often by reference to previous genera­
tions too (see, for example, section 15.1.2.4, on Welsh surnames). The patronymic system
survived in Sweden well into the nineteenth century and still exists today in Iceland,
where people are typically known by a given name and a patronym. Thus, the son or
daughter of a man with the given name Sven would be Svensson or Svensdóttir
respectively. This system is also found in English medieval records such as the fourteenth-
century poll tax returns, where, for example, Alicia Robertdoghter is recorded in Rigton,
West Riding of Yorkshire, in 1379. However, no names of this -daughter type have sur­
vived in England today. This example shows how the patronymic system, which is not
hereditary, is distinct from the binomial system, in which an individual inherits a heredi­
tary surname1 as well as being given a forename at or soon after birth. The binomial sys­
tem is used today throughout the English-speaking world, in Europe, and in certain other
countries.

The binomial system has been established in most European countries since the four­
teenth century. Between the twelfth and the fourteenth centuries (and in some places ear­
lier) descriptive, non-hereditary bynames—typically derived from locations, relationships,
nicknames, or occupations—gradually became fixed within family groups and passed
down to subsequent generations. Throughout Europe there is remarkable uniformity in
the types of names used, with comparatively few local differences. An example of a local
difference is that family names of locative origin are very rare in Ireland but very com­
Page 2 of 24
Family Names

mon in England. This is predominantly due to differences in the historical development of

bynames and hereditary family names, which we will now summarize by giving a brief ac­
count of surname history in Britain and Ireland.

15.1.2 The Origin and Typology of Surnames in Britain and Ireland

15.1.2.1 English Surnames

There is no simple answer to the question when and why hereditary family names first
came into use in England. The history of their development is complex, with much varia­
tion in different parts of the country and different social classes, over several centuries.
However, some broad generalizations can be made. It was very rare for a person to be
(p. 216) recorded with more than one name before the Norman Conquest. Hey (2000: 51)

comments that ‘the Englishmen who were recorded in Domesday Book as the holders of
land before the Conquest did not possess hereditary surnames but were known simply by
a personal name, such as Alric, Thorald or Wulfstan’. However, in some pre-Conquest
records, ‘it was often found convenient to identify a man by describing him as son of his
father’. Therefore, it could be said that some people bore second names at this time, but
‘such names were not family names; they died with the man’ (Reaney 1967: 75).

The next step toward the adoption of hereditary family names in England was the use of
non-hereditary bynames. These names had a rather different semantic value from that of
surnames today. They were used to describe some aspect or feature of their bearer, dis­
tinguishing him (or her) from other people by reference to occupation, geographical loca­
tion or origin, relationship to another person, or some physical or behavioural character­
istic.

Bynames and surnames are classified under one of the following four broad categories:
locative names, nicknames, occupational names, and relationship names. Each category
can be further subdivided. Thus, locative bynames can be either topographical (derived
from a feature of the landscape, e.g. Hill, Ford, Marsh) or toponymic (taken from the pre-
existing name of a town, farm, or other habitation, e.g. Burford, Blakeway, Copplestone).
Many occupational names are straightforward and self-explanatory even today (e.g. Bak­
er, Smith) but others are fossils, from a term that is no longer used (e.g. Wright,
Chandler). Some occupational names originated as metonymic nicknames, for example
the surname Cheese denoted a maker or seller of cheese. The surname Wastell, denoting
someone who made or sold fine cakes, is a metonymic nickname from a Norman French
word that is the equivalent of modern French gâteau. Status names such as Knight and
Squire are usually classified as a subdivision of occupational names.

Bynames were coined mainly in Middle English—the vernacular language of the time—al­
though names of Norman French origin were also adopted. The adoption of bynames fol­
lowing the Norman Conquest may have been accelerated by an increase in medieval bu­
reaucracy. Hey (2000: 54) attributes the development of hereditary surnames at least in
part to the fact that ‘whereas the Anglo-Saxons and Vikings had used a wide range of per­
sonal names, the Normans favoured very few’, some of which are still strikingly frequent

Page 3 of 24
Family Names

today, such as the traditional male names John, Robert, and William and the female names
Juliana, Isobel, and Elizabeth. The smaller stock of given names in use among the Nor­
mans and the gradual abandonment of most Anglo-Saxon given names meant that a larg­
er number of people were known by the same name, so there was a need to distinguish
between individuals in some other way than the use of a sole given name. Bynames were
used for this purpose. As each byname was particular to the individual, it would not have
been passed on to any offspring. This non-hereditary characteristic meant that any one in­
dividual might be known by two or more bynames. An example is ‘Ricardus filius Walteri,
de Cliue’ (Reaney 1997: xii), recorded in a Worcestershire assize roll from 1221. This
court record identifies the individual both by his parentage and by the location (Cleeve)
from which he came.

(p. 217) Throughout this period (eleventh–fourteenth centuries) hereditary surnames were
gradually coming into use, but they were by no means stable. There is ‘evidence
that . . . nicknames and “bynames” continued to replace or modify established surnames
into the nineteenth century at least’ (Redmonds 1997: 96). On the other hand, it is clear
that certain names began to be passed from father to son from soon after the Norman
Conquest and that this practice established itself as the norm by the end of the four­
teenth century. Thus some people were known by what we today would call surnames,
while others during the same period were known by non-hereditary bynames. Some of
these bynames came to be transmitted along family lines and so established themselves
as hereditary surnames, while others died out during the medieval period. Sturges and
Haggett (1987) have shown, by purely statistical modelling, given reasonable assump­
tions about the number of marrying sons in each family, that there is a general tendency
for common surnames to become more common, while rare names become rarer and
many of them die out.

The use of hereditary names in England was highly socially stratified from the beginning,
soon after the Conquest, and was influenced by their use in Normandy, where ‘some of
the more important and wealthier noble families . . . already possessed hereditary sur­
names’ (McKinley 1990: 25). Indeed, it was the wealthy landholders who were the first to
adopt hereditary surnames in England, ‘in the two centuries or so after the
Conquest’ (McKinley 1990: 28), while other social classes continued to use non-heredi­
tary bynames. These landholders typically used toponymic names—that is, they were typi­
cally identified by the names of estates from which they came.

While the development of surnames was by no means uniform throughout the country,
most authorities agree that hereditary surnames were in the majority in the south of Eng­
land by about 1350 and by 1450 in the north (Reaney 1967: 315; McKinley 1990: 32; Hey
2000: 53). Some hereditary surnames ‘had genuinely late origins, evolving in parts of
northern England well into the 1700s’ (Redmonds 1997: 57). The development of heredi­
tary surnames in England was a complex, long-drawn-out process.

Page 4 of 24
Family Names

15.1.2.2 Irish, Manx, and Scottish Gaelic surnames

The Scottish Gaelic language and the Irish language are closely related and this was even
more true in the Middle Ages, at the time of surname formation. Scottish Gaelic is spoken
in the Highlands and Islands and was, until the sixteenth century, in Galloway in the
south-west. A traditional view is that Scottish Gaelic was brought to Scotland from Ire­
land in the fourth–fifth centuries AD, but it seems more likely that there was continuous
interchange from earliest times around the Irish Sea, for example in the ancient maritime
kingdom called Dál Riata, which extended from northern Ireland up into the Hebrides.

Many similarities can be perceived between the Irish Gaelic names of Ireland and the
Scottish Gaelic names of Scotland. Attempting to distinguish between the two risks mak­
ing a false distinction. Nevertheless, many surnames can be identified as distinctively Ir­
ish, while a smaller number are distinctively Scottish. In particular, the latter include
names in the clan system, a distinctively Scottish social institution according to (p. 218)
which people were associated by birth, servitude, or locality with the hegemony of a clan
chief, either taking the clan name as a surname or taking a surname of a ‘sept’ (a subordi­
nate group) of one of the major clans.

Initially, Gaelic patronymics were formed by use of the prefix Mac and (in Ireland) by Ó
‘grandson of’, giving patronymics such as Mac Cárthaigh ‘son of Cárthach’ and Ó Conall
‘grandson of Conall’. Non-hereditary names of this form ‘will be found in the records re­
lating to centuries before the tenth’, with their use as hereditary surnames having come
‘into being fairly generally in the eleventh century’ (MacLysaght 1985: ix). Names that
became hereditary yielded anglicized forms such as McCarthy and O’Connell. Woulfe
(1923: 15) observes that ‘Irish surnames came into use gradually from about the middle
of the tenth to the end of the thirteenth century’.

After the convention for prefixing names with Mac and Ó had become common, further
changes in Irish surnames took place. Some included the words giolla and maol, meaning
‘follower’ or ‘servant’, ‘in the sense of devotee of some saint e.g. Mac Giolla Mhártain
(modern Gilmartin or Martin) or Ó Maoilbhreanainn (modern Mulrennan) from St. Martin
and St. Brendan’ (MacLysaght 1985: ix). Surnames deriving from occupations and nick­
names were also formed, such as Mac Giolla Easpaig ‘son of the servant of the bishop’
and Mac Dubhghaill ‘son of Dubhghall’, a personal name meaning ‘dark stranger’. Most
Irish surnames acquired one or more anglicized form in the sixteenth century. For exam­
ple, the two names just mentioned yielded the anglicized forms Gillespie and McDowell.
Many Irish surnames yielded two distinct sets of anglicized forms, due to the phonetic
phenomenon of lenition. For example, in the Irish surname Mac Daibhéid ‘son of David’,
the D- came to be pronounced as a gutteral voiced fricative, yielding the anglicized sur­
name McKevitt alongside the more etymological form McDevitt. The same phenomenon in
Scottish Gaelic yielded both McWhan (lenited) and McSwan (unlenited) as anglicizations
of Mac Suain ‘son of Sveinn’. Similarly, Mac Domhnuill ‘son of Donal or Donald’ is the
source of both McDonnell (McDonald) and McConnell. In a further development, the
patronymic Mac- was often dropped or reduced to a residual C-, yielding anglicized sur­
names such as Connell and Donald. Patronymic prefixes in Ireland ‘were very widely
Page 5 of 24
Family Names

dropped during the period of submergence of Catholic and Gaelic Ireland which began in
the early seventeenth century’ (MacLysaght 1985: x). Some Irish names were translated
to give English equivalents, with the Irish Mac a’ghobhainn ‘son of the smith’ sometimes
being anglicized as Smith and Mac an tSionnaigh ‘son of the fox’ as Fox. Sometimes, Irish
names were mistranslated due to folk etymology, as in the case of Bird, which, as an Irish
name, represents quite a large number of Irish names that happen to contain the letters
éan, for the Irish word éan does indeed mean ‘bird’, although this has nothing to do with
the surnames Ó hÉanna (Heaney), Ó hÉanacháin (Heneghan), or Mac an Déaghanaigh
(McEneaney), which are among those for which Bird has been adopted.

The development of Irish surnames into their modern forms was sometimes even more
complex, as MacLysaght (1985) shows in a discussion of Abraham as an Irish surname:
(p. 219)

Of course that is Jewish elsewhere, but in Ireland it is the modern corrupt or dis­
torted form of an ancient Gaelic surname, Mac an Bhreitheamhan (son of the
judge). It was first anglicized MacEbrehowne, etc., which was shortened to MacE­
brehan and MacAbrehan, later MacAbreham and so to Abraham. Other anglicized
forms of this name are Breheny and Judge.

(MacLysaght 1985: xii)

The prefixes Mac and Ó in Irish surnames re-emerged in an anglicized form in the late
nineteenth century. MacLysaght (1985: x) suggests this began as a result of a ‘revival of
national consciousness’, and comments that there was a steady increase in the number of
people adopting O in the name O’Sullivan from 1866 to 1944. Similarly, Yurdan (1990: 3)
notes that ‘during the renaissance of interest in things Irish during the period 1930–60,
the “O”s and “Mac”s were reinstated to their former positions’. Since the 1960s there has
been an equally noticeable resurgence in the use of Irish-language (Gaelic) forms of fami­
ly names in Ireland.

As noted by Hanks and Muhr (2012), there has been considerable exchange of surnames
between Britain and Ireland for almost a millennium. In the twelfth and thirteenth cen­
turies, English kings and Norman barons brought family names such as Butler, Clare,
FitzGerald, and Bermingham to Ireland, and in the sixteenth and seventeenth centuries
many other family names of English, Welsh, and Scots origin became established there. In
the early seventeenth century, King James I of England (and VI of Scotland) encouraged
the settlement of ‘plantations’ in Ireland, particularly northern Ireland, as a result of
which the family names of Scottish Border reivers and others (Nixon, Armstrong, Paisley,
etc.) became established in Ireland, mainly northern Ireland. In the nineteenth and twen­
tieth centuries, the flow was reversed and most Irish surnames, in their anglicized forms,
became established in Britain: notably in south Lancashire, Lanarkshire (Glasgow), the
coal-mining region of south Wales, and the industrial west Midlands.

Page 6 of 24
Family Names

While many Irish and Scottish Gaelic hereditary surnames were in existence as early as
the eleventh century, non-hereditary names persisted, as can be seen in this late example
noted by Black (1946: xxv): ‘Gideon Manson . . . died in Foula in March, 1930. His father’s
name was James Manson (Magnus’s son) and his grandfather was called Magnus Robert­
son.’

Before leaving the topic of Gaelic family names, we should note that there are approxi­
mately 200 distinctively Manx family names in Britain today. Many of these begin with an
initial C- (Clague, Cretney) or Q- (Quirk, Quinney), residues of Gaelic Mac-.

15.1.2.3 Surnames in Scots-speaking Scotland

Scotland is a country with a rich variety of linguistic and cultural heritages. In addition to
Gaelic, Scottish family names are also of Cumbric origin,2 Scandinavian (also known
(p. 220) as Old Norse), and Anglian (the northern dialect of Old and Middle English). The

latter in Scotland developed into distinctively Scottish varieties of English, sometimes

called Lallans (the language of the Lowlands), the Scots leid (the Scottish language), or
simply Scots, which is the term we shall use here.

Hereditary surnames first occurred in Scots-speaking regions at around the same time as
in England, and many were ‘introduced into Scotland through the Normans’ (Black 1946:
xiii), usually with names of toponymic origin. Following this, the ‘spread of surnames in
Scotland seems to have been slow’ (McKinley 1990: 45). While most landholders seem to
have ‘acquired surnames . . . by about 1300’ (McKinley 1990: 45), it seems that ‘the gener­
al spread of hereditary surnames was not complete in the Scots-speaking regions until at
least the sixteenth century’ (McKinley 1990: 46). The establishment of hereditary sur­
names in the country occurred later than in England.

An important influence on the development of Scots family names was the importation of
a Norman bureaucracy in the twelfth century, for which the person most responsible was
King David I (reigned 1124–53). David had been brought up at the English court of King
Henry I and had married Maud, Countess of Huntingdon. When, at the age of 39 or 40, he
unexpectedly succeeded to the throne of Scotland, he took with him a cohort of Norman
retainers from eastern England with surnames like Lindsay, Ramsay, Sinclair, and Hamil­
ton.

Scots surnames can be classified using the same typology as for England (see section
15.1.2.1). Black (1946: xxix) notes that ‘contrary to the common view, I have found few of
our [Scottish] surnames to be derived from nicknames’.

15.1.2.4 Welsh Surnames

The development of hereditary surnames in Wales was very different from the English, Ir­
ish, and Scottish patterns. Even though Norman lords acquired land in Wales soon after
the Conquest, ‘neither this, nor the increasing use of hereditary surnames by English set­
tlers in Wales, seems to have had much influence among the Welsh population’ (McKinley
1990: 41). Even by 1500, hereditary surnames were still rare in Wales. The Welsh

Page 7 of 24
Family Names

patronymic naming system involved using Welsh mab ‘son’ to create names in the form of
‘X mab Y’. The word mab would have become fab due to grammatically triggered lenition,
which subsequently became ab because ‘the Welsh f sound was probably bilabial and
therefore more easily lost’ (Morgan and Morgan 1985: 10). Generally, ab occurred before
names with initial vowels, and ap before those with initial consonants, resulting in names
such as ‘Madog ab Owain’ and ‘Madog ap Rhydderch’ (Rowlands and Rowlands 1996: 8),
although not all recorded names conform to this rule.

In Wales, not until the mid-sixteenth century did ‘the change to settled surnames begin to
filter through different levels of society’ (Rowlands and Rowlands 1996: 25), resulting in
the loss of ab or ap in a number of names. This explains why such a large proportion of
surnames in Wales today are derived from given names. Jones and Williams are typically
Welsh names: the English genitive -s apparently replaced Welsh ab/ap in many cases, with
such genitive -s names having ‘been common in Wales since at least (p. 221) the sixteenth
century’ (McKinley 1990: 226), coinciding with increased adoption of hereditary sur­
names in the country.

The Welsh patronymic form is still retained, to some extent, in certain hereditary sur­
names, where ab/ap has become incorporated with the following name through metanaly­
sis, ‘thus Thomas ap Howell would become Thomas Powell’ (Rowlands 1999: 166–7). This
was, and is, most common in areas of ‘greatest and earliest English influence’ (Rowlands
1999: 167), close to the English border. In other parts of Wales, the Welsh patronymic sys­
tem appears to have been retained much longer, with names in ap occurring as late as the
eighteenth century ‘in upland Glamorgan parishes and in western Monmouthshire’ (Row­
lands and Rowlands 1996: 25–6). There are some personal names found today in Wales
with the form X ap Y. These can be attributed to ‘renewed national awareness and grow­
ing interest in the past’, leading to a revival of patronymic names ‘in the second half of
the twentieth century’ (Rowlands and Rowlands 1996: 34).

15.1.3 The Effects of Migration on the World’s Family Name Stocks

While each country and indeed each region has its own histories and patterns of family
name development, worldwide migration has meant that present-day name stocks tend to
be much more ethnically and culturally diverse than they were a few decades ago. There­
fore, any attempt to survey current family name stocks in any one country generally re­
quires a wide variety of linguistic expertise. In the UK in particular, a reasonably compre­
hensive account of present-day surnames requires not only traditional expertise in Old
and Middle English, Latin, Anglo-Norman French, and the Celtic languages but also ex­
pertise in Yiddish and Hebrew, other modern and medieval European languages, Arabic,
Persian, Turkish, Indian languages, and Chinese, among others. Only in this way can a
reasonably comprehensive account of modern family names in countries such as Britain,
Australia, and America be developed. Family name dictionaries and surveys have been
compiled in several but by no means all countries of the world.

Page 8 of 24
Family Names

15.2 Studies of Family Names in Britain and

Ireland
The most reliable traditional introductions to the study of surnames in Britain are those
by Reaney (1967) and McKinley (1990), which offer a philologist’s and a historian’s per­
spective respectively. Because they are essentially national surveys, they have little to say
about surnames outside Great Britain, while at the other end of the spectrum they can do
little more than exemplify fine-grained local details and regionally specific patterns of
surname development and distribution that are now recognized as (p. 222) an essential
component of the study of surnames. McKinley (1990) himself identifies the problem thus:

It is impossible to examine the surnames present in several counties, from differ­

ent parts of England, without being struck by the very sizable differences which
existed in the Middle Ages, and which in large measure persisted into later peri­
ods, between the different English regions.

(McKinley 1990: 20)

An ideal introduction to the study of surnames would be interdisciplinary, bringing to­

gether the expertise of historians, historical linguists, demographers, statisticians, ge­
nealogists, and (more recently) geneticists. Increasingly, this interdisciplinary approach is
beginning to be adopted, but at present no such survey exists. An international perspec­
tive beyond English can be gleaned from the 108 pages of introductory essays in the Dic­
tionary of American Family Names (DAFN 2003).

15.2.1 England

Compared with other onomastic fields such as place-names, family name research has re­
ceived relatively little scholarly attention in Britain until recently. The earliest work offer­
ing information about surnames is a chapter in Camden’s (1605) Remains Concerning
Britain, which includes an alphabetical list of 253 locative surnames, mostly the sur­
names of gentry of Norman French origin. Over two centuries were to elapse before the
next relevant work, namely Lower (1849), which outlines the chronology of hereditary
surname adoption. It organizes the discussion of surnames by categories, though these
are different from those that are generally used today.

The next important work is Bardsley (1875), which categorizes surnames using a typology
of five types: ‘Baptismal or personal names’, ‘local surnames’, ‘official surnames’, ‘occu­
pative surnames’, and ‘sobriquet surnames or nicknames’. Building on this, Bardsley
(1901) produced the first reasonably comprehensive inventory of English surnames.
Among other innovations, it makes a systematic attempt to support etymologies with ex­
amples of early bearers. In the early twentieth century, studies of English surnames were
published by Weekley (1916) and Ewen (1931, 1938) among others.

Page 9 of 24
Family Names

‘The standard work on the etymology or meaning of surnames’ (Redmonds et al. 2011: 4)
is P. H. Reaney’s (1958) Dictionary of British Surnames, published in a revised third edi­
tion as A Dictionary of English Surnames (Reaney 1997). Explanations are terse and
sometimes cryptic, but they are grounded in traditional scholarship. Most importantly,
they are supported by a wide selection of early bearers from medieval records. Recent re­
search has shown, however, that the Reaney (1997) dictionary must be used with caution.
Reaney was a great scholar, but we now know that some of his magisterial pronounce­
ments are simply wrong. For example, Redmonds (2014) has shown that Gaukroger is a
locative surname meaning, roughly, ‘cuckoo crag’ and not, as Reaney asserts, a nickname
meaning ‘foolish Roger’. Others of Reaney’s explanations (p. 223) are ‘fudges’, which blur
the issue to the point of being misleading. Typical is Ramshaw, which Reaney has merely
as a cross-reference to Ravenshaw. Etymologically these two surnames are indeed relat­
ed, but in fact Ramshaw is a toponym from a place near Bishop Auckland in county
Durham (a place not mentioned by Reaney), while Ravenshaw (the main entry in Reaney),
which is now rare or extinct as a surname, is from a place in Warwickshire. They explain
a cluster of eight different toponymic surnames (Ravenshaw, Ravenshear, Ramshaw,
Ramshire, Ranshaw, Renshaw, Renshall, Renshell) as ‘dweller by the raven-wood’, ap­
pearing to imply that they are variant spellings of a single topographic name—but the fact
is, there is no such thing as a ‘raven-wood’ and no one was ever named as a dweller by
one. The family names concerned are from different place-names, and these places were
named hundreds of years before surnames came into existence.

Such problems were compounded by the fact that, for many names, Reaney’s terse expla­
nations regularly give only an Old English, Old Norse, or Continental Germanic etymolo­
gy, bypassing intermediate steps such as Middle English and Old French. Reaney adopted
this policy mainly because of space constraints imposed by his publisher due to post-war
paper shortages, but it is particularly misleading because surnames were formed in the
Middle English and early modern English periods; there is no such thing as an Old Eng­
lish surname.

Perhaps Reaney’s greatest weakness was his almost complete failure to take account of
the statistical relationships between surnames and locations. It must also be mentioned
that literally thousands of well-established English surnames do not appear at all in
Reaney’s dictionary, which has been described as a dictionary of medieval surnames
arranged under their modern derivative forms. If Reaney had nothing to say about a
name, he simply omitted it. We mention these points, not to carp at Reaney’s achieve­
ment, which is remarkable by any standard, but in order to illustrate the enormous
amount of fine-grained detailed research that is needed before studies of surnames and
family names can take their place as adequately informative and reliable works alongside
place-name studies and works of historical lexicography.

One recent work that is better focused and stands up to scrutiny is Redmonds’ Dictionary
of Yorkshire Surnames (forthcoming). This is based on detailed evidence of many kinds:
medieval local records, local dialects (past and present), genealogical and genetic track­
ing, and contrastive geographical distribution. It will serve as a model for future county

Page 10 of 24
Family Names

and areal studies. Hopefully, in years to come, a range of comparable county-by-county

studies will be created, emulating Redmonds’ achievement for Yorkshire.

Two other works of surname lexicography that must be mentioned here are Cottle (1967)
and Titford (2009). Cottle’s is an admirably succinct and reliable work, which proved pop­
ular for over forty years. It contains a few entries that were not explained by Reaney and
an occasional dry witticism. For example Butlin, the surname of the founder of a chain of
holiday camps, is explained as being derived from Old French boute-vilain ‘hustle the
churl’; Cottle adds, ‘suggesting an ability to herd the common people’. Titford (2009) is an
expanded edition of Cottle’s work. It made extensive use of previous (p. 224) publications:
not only Cottle and Reaney but also Hanks and Hodges (1988) and DAFN (see section
15.6 below).

Rigorous scholarship is a feature of the Lund Studies in English, inspired by Professor Eil­
ert Ekwall, himself a great surname and place-name scholar. These works are far from
comprehensive, but their focus on surname typology makes them useful sources of partic­
ular early name bearers and etymological information. For example, Fransson’s (1935)
Middle English Surnames of Occupation 1100–1350 provides a list of occupational names
with early bearers and suggested etymologies; Löfvenberg (1942) explains a selection of
medieval locative surnames; and Jönsjö (1979) does the same for nicknames. Other rele­
vant works in this series are Thuresson (1950) and Kristensson (1967–2002). However,
these works are not without problems. Fransson (1935) studied names from only ten Eng­
lish counties. Jönsjö’s (1979) etymological explanations are sometimes ambiguous and his
treatment of names that share an element is not always consistent. McClure (1981b)
comments:

If one dimension of information is chiefly lacking in the comparative methods used

in Lund Studies of ME bynames it is that of local and biographical history. The
name is treated as ‘word’ rather than ‘person’, as a manifestation of linguistic
form rather than social life.

(McClure 1981b: 101)

Clark (2002: 116) makes a similar point, that ‘to study in purely lexical and etymological
terms a form recorded as a name, and sometimes solely so, may be to study something
that never, and certainly not in the given context, existed at all’. Nevertheless, the Lund
studies made an important contribution to the identification and understanding of English
surnames.

The English Surnames Series (ESS), funded by the Marc Fitch Foundation at the Universi­
ty of Leicester, set out to investigate surnames historically county by county. Only seven
volumes were published (Redmonds 1973; McKinley 1975, 1977, 1981, 1988; Postles
1995, 1998), but these have provided a wealth of detailed information on surnames in the
particular counties and regions studied. Clark (1995) recognized the importance of this
approach, noting that the works of the ESS

Page 11 of 24
Family Names

never lose sight of the special nature of naming, as distinct from common vocabu­
lary, and so proceed consistently in terms of social status, of domicile and land­
holding, of migration-patterns, of economic activity, of gender and familial rela­
tionships, of types of milieu, and of ramification of individual clans.

(Clark 1995: 384)

However, it is not necessarily the case that county-based research is suitable for investi­
gating regional surname patterns. Postles (1995: 4), in his ESS volume for Devon, con­
cedes that ‘counties can never be’ regional societies, while Redmonds (2004b: xiv) has al­
so commented that ‘many of the counties are made up of several distinct regions, and
these can be linked to marked differences in their topography, history and language’. Fu­
ture surname research could benefit from focusing on socially, topographically, (p. 225)
culturally, and linguistically distinct regions, perhaps investigating particularly localized
patterns of development, as Hey (2000: xi) has suggested:

The research that will forward our understanding of how surnames arose and
spread will need to be focused on particular parts of the country, looking at how
groups of names were formed at different times in particular local communities.

(Hey 2000: xi)

In 2009–10 Oxford University Press and the Arts and Humanities Research Council of
Great Britain were persuaded to initiate an ambitious research project called Family
Names of the United Kingdom (FaNUK)). Eventually, this found a home at the University
of the West of England under the direction of Richard Coates, with Patrick Hanks as lead
researcher. It is due to be published in 2016 as the Dictionary of Family Names in Britain
and Ireland (FaNBI). The entry list is based on a comparison of 1881 Census data with a
more recent inventory based on 1997 electoral rolls, so that in principle almost every sur­
name in the UK, no matter how rare, can be considered. People often ask, how many sur­
names are there in the UK? Unfortunately, a simple answer cannot be given, because
among the hundreds of thousands of very low-frequency items, genuine surnames (most
of which are recent immigrant names—i.e. names that came to the UK after 1945) merge
imperceptibly into misprints and transcription errors. FaNBI contains entries for all fami­
ly names with 100 or more bearers in the UK in 1997, regardless of ethnic or cultural ori­
gins. To these were added entries for names that are in other British surname dictionar­
ies and ‘established names’ that are of particular historical or philological interest. ‘Es­
tablished names’ in this context is a term that contrasts with ‘recent immigrant names’.
In practical terms, established names are those found in both the 1881 census and the
1997 data.

The result is a headword list of over 45,000 family names. There are almost 20,000 main
entries and over 25,000 current spelling variants, together with innumerable examples of
historical spelling variants. The spelling of family names in the UK is much more volatile
than the spelling of place-names or English vocabulary words. Particular spellings of a
widespread name sometimes come to be accepted as conventional in different families or

Page 12 of 24
Family Names

in different local areas. There are at least three ways in which FaNBI differs from previ­
ous works: (1) early bearers, (2) information about geographical location, and (3) recent
immigrant names.

Following the lead set by Bardsley and Reaney, examples of medieval and post-medieval
early bearers are systematically included in FaNBI under each main entry, extracted from
sources such as medieval tax records, court records, wills, and parish registers, many of
which are now available for analysis in digitized form. These lists show the linguistic de­
velopment and geographical spread of each surname since the time of its first use, while
in many cases early forms provide evidence for etymological origins.

The main location of early bearers in Archer’s (2011) British 19th Century Surname Atlas
(see section 15.3.1) is summarized briefly but systematically for almost every FaNBI en­
try. An attempt is made to record the earliest known bearer in the main geographical lo­
cation with which the name is associated. In many cases, especially among locative
(p. 226) surnames, the geographical distribution of a surname correlates with the locality

in which it originated, and this can provide useful evidence for the identification of lost
place-names. The information about the main 1881 location also makes FaNBI a useful
genealogical resource, pointing family historians toward the county or counties in which
their research is most likely to be productive.

In addition, FaNBI provides a picture of immigration to Britain through the centuries. The
Celtic, Anglo-Saxon, Scandinavian, and Norman population stocks were augmented in
substantial numbers from time to time over the centuries. Flemish weavers migrated to
England, having been first invited in the fourteenth century by Edward III with the aim of
maintaining and improving the English wool and cloth industry. Huguenots entered
Britain during the seventeenth century, fleeing to avoid religious persecution. Sephardic
Jewish surnames from Spain, Portugal, and other Mediterranean countries arrived from
the seventeenth century onwards, and waves of Ashkenazic surnames from central and
eastern Europe arrived in the nineteenth and twentieth centuries.

Following the collapse of the British Empire in the second half of the twentieth century,
ethnic diversity in Britain greatly increased as many people holding (or acquiring) British
passports chose to migrate to England for economic and other reasons. As a result, many
names borne by recent immigrants have been pressed into service as family names in
Britain. Approximately 3,800 recent immigrant names with more than 100 bearers are
recorded in FaNBI and more than 1,600 of these are from the Indian subcontinent, with
Muslim, Hindu, Sikh, and other religious affiliations, each of which provides a rich set of
etymological and cultural traditions. Muslim names in the subcontinent are mostly of Ara­
bic etymology, with some Persian; Sikh names are derived from Panjabi, while Hindu
names come from many different Indian languages. The Indian family name Patel is the
32nd most frequent surname in FaNBI’s 1997 data with 95,177 bearers, followed by Khan
with 63,795. Muhammad has only 15,016 bearers, but that is because there are seven­
teen variant spellings in the dictionary (plus a lot more that are too rare to be included).

Page 13 of 24
Family Names

Over 400 family names in Britain are of Chinese origin, many of them being Hong Kong
romanizations of Chinese names in the Cantonese dialect, as opposed to the Mandarin
forms, which are regarded as standard in China itself. Other family names of Chinese ori­
gin arrived via Malaysia and Singapore. English surnames of Chinese origin are particu­
larly complex: a single ‘English’ orthographic form may represent up to twenty-two differ­
ent Chinese surnames (‘different’ in that they are represented by different Chinese ideo­
graphs, each of which may have more than one explanation as to its origins). Ambiguity is
avoided because Chinese is a tonal language: most apparently homophonous surnames in
Chinese are distinguished by different tones, which are lost in English transcriptions.

In many entries, FaNBI gives additional information about family names, over and above
the etymology, for example information about Scottish clans or historical information on
great and powerful families such as Cecil and Cavendish. In other cases, brief summaries
of obsolete occupations are given, as for the surname Reeve: (p. 227)

In medieval England a reeve was an administrative official responsible for the ad­
ministration of a manor, including organizing work done by the peasants on the
land for their lord, collecting rents, selling produce, and so on.

Elsewhere, explanations of relevant terms in the feudal system of land tenure are given,
for example at Ackerman. Reaney’s (1997: 2) explanation of this surname says tersely:

OE [i.e. Old English] æcermann ‘farmer’, a husbandman or ploughman.

The FaNBI entry, having the luxury of greater space, is able to explain:

An ackerman was a bond tenant of a manor holding half a virgate of arable land,
for which he paid by serving as a ploughman.

For further information about FaNBI methodology, see Hanks et al. (2012).

15.2.2 Multidisciplinary Surname Research

Redmonds et al. (2011) have clearly shown the benefits of a multidisciplinary approach to
surname research, coordinating philology, history, and genealogy with geographical and
biographical evidence, where (for example) they consider a wide range of historical
sources to determine the origin of the name Tordoff. The 1881 distribution shows that this
surname was concentrated in the West Riding of Yorkshire, encouraging the researcher to
search local records from this area. However, Redmonds et al. (2011) established that

the surname survived in Dumfriesshire into the late fifteenth century. The next ref­
erences place it in York between 1499 and 1524, where the family were pewter­
ers, and then in and around Leeds and Bradford by 1572, where it ramified suc­
cessfully in the village of Wibsey. More than 95 per cent of the 707 Tordoffs in
1881 lived in the West Riding, with Bradford (386) and Leeds (145) the major cen­
tres; the surname is still numerous in both places at the present day.

Page 14 of 24
Family Names

(Redmonds et al. 2011: 99)

The Dumfriesshire origin of the name led to the conclusion that the surname Tordoff
‘derives from a locality known as Tordoff Point on the Scottish side of the Solway Firth’.
Without the prosopographical evidence, this origin may not have been so easily found or
so confidently asserted.

Redmonds (1997) has also shown the advantages of considering a wide range of historical
sources in determining a surname’s etymology, particularly in his analysis of alias names.
With a purely philological approach, a surname’s origin can often be identified through
the comparison of similar name forms. However, where a name has been altered by scrib­
al influence to such an extent that its form is no longer etymologically representative, lin­
guistic comparison is of little help, and a different approach is required. Redmonds’ inves­
tigation of a large number of (p. 228) sources has allowed him to discover certain alias
names, where a person is recorded with two or more names, which suggest an etymologi­
cal connection between two surnames which might not appear to be related on form
alone. One such example is the case of ‘Simon Woodhouse alias Wydis’ from Thornton le
Moor in 1611 (Redmonds 1997: 125).

DNA evidence is also relevant (see Sykes and Irven 2000; Jobling 2001; Bowden, et al.
2008; King and Jobling 2009). Redmonds et al. (2011: 156) argue that ‘just as a father
passes on his surname to all his children, so he passes on his Y chromosome type to all
his male children,’ and they then pass the same Y chromosome type to their children, and
so on. By comparing the Y chromosomal DNA of different people with the same surname,
it is possible to demonstrate that the bearers share a common ancestor. In this way, Sykes
and Irven (2000) showed that the English surname Sykes is most probably monogenetic,
despite previous work that predicted it to be polygenetic.

The multidisciplinary approach has not only involved the application of wider historical
knowledge and DNA evidence to surname study, but also the use of surname data in other
historical studies. McClure (1979) used toponymic surname data to investigate rural and
urban patterns of medieval migration, and the value of this methodology led to its use in
further migration studies (see Penn 1983; Rosser 1989; Kowaleski 1995). Researchers in
demography and geographical information science have made use of surname evidence
(Schürer 2002, 2004; Longley et al. 2005), while lexicographical research using surname
evidence has also been carried out (see, for example, Mawer 1930; McClure 2010a,
2010b, and the Swedish works, predominantly by students at Lund University, which pro­
vided antedatings for a large number of words, such as Fransson 1935; Löfvenberg 1942;
Thuresson 1950; Jönsjö 1979).

15.2.3 Ireland

As a result of the complex development and anglicization of Irish Gaelic names, the con­
struction of an Irish surname dictionary is no simple task. The standard work was Woulfe
(1923), which took full account of this difficulty, being a dictionary in two parts, the first
of which lists Irish Gaelic surnames with their anglicized and English equivalents, while
Page 15 of 24
Family Names

the second contains etymological and historical discussion. This important work was fol­
lowed by MacLysaght’s (1957, 1985) The Surnames of Ireland.

Both Woulfe and MacLysaght were redoubtable scholars with a deep knowledge of Irish
family histories and an understanding of the linguistic vicissitudes that have affected fam­
ily names in Ireland over the centuries. As a result, Ireland is better served by its sur­
name dictionaries than other European countries including England. However, neither of
them includes evidence for early bearers, which makes it difficult for subsequent scholars
to evaluate their more controversial etymologies. By contrast, FaNBI includes early bear­
ers from several Irish sources, notably the Annals of Ulster, the Tudor Fiants, and a list of
nearly 60,000 individuals (Flaxgrowers) published by the Irish Linen Board in 1796. A
more recent work, providing etymological, historical, and (p. 229) distributional informa­
tion and based on the 1980s Irish telephone directory, is by de Bhulbh (1997).

15.2.4 Scotland

The standard work on Scottish surnames is Black (1946). This is a remarkable work of
scholarship, all the more remarkable because it was compiled in the New York Public Li­
brary. It contains over 8,000 surnames recorded in Scottish historical documents since
the medieval period. Wherever possible, entries in this dictionary include etymology, in­
formation about early bearers, and variant spellings. Entries for surnames derived from
Scottish place-names are particularly thorough and informative.

A more concise, though readable, work on Scottish surnames is Dorward (1978), which
contains entries for over 1,000 common Scottish surnames, explaining their etymological
origins and geographical distribution.

15.2.5 Wales

Dictionaries of Welsh surnames tend to be short in comparison to those from England,

Scotland, and Ireland. The dominant patronymic naming system of Wales (see section
15.1.2.4) means that there are a relatively small number of different family names in the
country. For this reason, Welsh surname dictionaries have sufficient space to give thor­
ough accounts of surname origins and development. The two main works are Morgan and
Morgan (1985) and Rowlands and Rowlands (1996, 2014), which between them give a
comprehensive account of surnames in Wales. Entries in the Morgan and Morgan volume
represent the medieval Welsh personal names that are the etymons of most surnames of
Welsh-language origin. Rowlands and Rowlands is a more user-friendly work, of particu­
lar usefulness for genealogists.

Page 16 of 24
Family Names

15.3 Studies of Family Names in Continental

Europe
Not every country in Europe has a reliable dictionary or other study of family names,
while even those that do exist are rarely comprehensive. Some local historical and region­
al studies are available, but much work remains to be done by way of investigation of the
family names in Europe.

Where national surname dictionaries are not available, DAFN provides at least a starting
point. American family names come from all over the world, so DAFN may be regarded as
roughly equivalent to an international comparative dictionary of world (p. 230) surnames.
In some cases, DAFN is all there is; in other cases, not even DAFN includes information
about family names in certain regions of the world.

15.3.1 The German-speaking Lands

The main dictionaries of German family names are by Gottschald (1932), Brechenmacher
(1936, 1957), and Bahlow (1967, 1993). Gottschald’s work has extensive lists of name
variants and etymological explanations for some of the names, but no examples of early
bearers. Bahlow includes an occasional mention of some early bearers under certain en­
tries, while Brechenmacher includes more extensive explanations, often supported by
early bearers.

A major research project currently in progress in Germany is Der Deutsche Familienna­

menatlas (DFA), a collaborative project based at the universities of Mainz and Freiburg,
under the direction of Damaris Nübling (Mainz), Konrad Kunze (Freiburg), and Peter
Auer (Freiburg). The research involves the systematic analysis of surnames, using tele­
phone directories, with geographical distribution maps of selected surnames and sur­
name features. This kind of distributional analysis represents a key development in family
name study, with a focus on the systematic computational analysis of large datasets.
Eventually, a new etymological dictionary of family names in Germany will be based on
the Atlas, superseding existing works. There are several local studies of surnames of par­
ticular German regions, while Zamora (1992) provides an account of Huguenot names in
the German states of the seventeenth and eighteenth centuries.

The standard work on Austrian family names is Hornung (1989), while Finsterwalder
(1978) provides a more closely focused account of family names in Tyrol. The standard
reference work for Swiss names is Meier (1989), which includes all family names borne
by Swiss citizens in 1962. Each entry contains a list of the Cantons in which bearers of
the family name are found, the year or period when the family name first appeared in the
country, and the cantons in which the name has occurred previously but has since died
out. For names that are not of Swiss origin, the bearer’s previous country of residence is
given.

Page 17 of 24
Family Names

15.3.2 Belgium and the Netherlands

A major scholarly and comprehensive dictionary of surnames in Belgium (including en­

tries for the majority of Dutch surnames that have any substantial frequency) is Debra­
bandere (1993), in which the entries contain etymologies, variant forms, and early bear­
ers. Not only surnames from Dutch- and French-speaking Belgium but also surnames
from northern France, where there was once Flemish influence, are included.

A database showing the geographical distribution of surnames is the Nederlandse Fami­

lienamenbank, hosted at The Hague’s Centraal Bureau voor Genealogie (2000–). Howev­
er, there is still no prospect of a comprehensive dictionary of Dutch family names.

(p. 231) 15.3.3 Scandinavia (Denmark, Norway, Sweden) and Finland

While the languages of Denmark, Norway, and Sweden are very closely related, patterns
of surname development in Scandinavia show distinctive national and regional differ­
ences. There are generalizable differences in the types of surnames used in the different
countries, with, for example, the majority of Danish and Swedish family names being
patronymic, while most Norwegian family names are of locative origin.

Scholarly works on Scandinavian family names and their origins include Modéer’s (1989)
survey of the history of Swedish personal naming and family naming, Veka’s (2000)
dictionary of Norwegian family names, and Knudsen et al.’s (1936–64) study of Old Dan­
ish forenames and nicknames.

Finland has its own history of family naming, with perhaps the most characteristic fea­
ture of its names being the common ending -nen, originally a diminutive and possessive
suffix, which was later simply added to patronyms as a way of creating surnames. Studies
of Finnish family names include Mikkonen and Sirkka (1992) and Pöyhönen (1998).

15.3.4 France

The standard reference works for the surnames of France are Dauzat (1945, 1951) and
Morlet (1991). These dictionaries are extensive collections of names, giving etymologies
and variant forms under each entry. However, neither dictionary provides information
about early bearers.

15.3.5 Italy

A comprehensive dictionary of Italian surnames is DeFelice (1978), in which most entries

include a list of variant forms, etymologies, and the geographical distribution of the
name, although early bearers are not provided. DeFelice (1980) is a more discursive
work, providing information on the history, typology, and geography of Italian family
names.

Page 18 of 24
Family Names

15.3.6 Spain and Portugal

The nearest to a comprehensive dictionary of Spanish family names is Tibón (1995

[1988]), a heroic one-man effort to provide etymological and other information on sur­
names throughout the Spanish-speaking world, without access to the necessary appara­
tus in support, such as databases of medieval records, distributional surveys, census da­
ta, and so on.

There are several surveys of family names in certain areas. Notably, Catalan is
(p. 232)

well served by Moll (1982) and Coromines (1989–97), while Basque names are described
by Michelena (1973).

Machado (1984) includes information about Portuguese family names as well as vocabu­
lary words.

15.3.7 Hungary

Kálmán’s (1978) work provides an account of the origins and history of Hungarian family
names, along with discussion on given names and place-names.

15.3.8 The Slavic and Baltic Countries

A selection of the numerous works on family names from Slavic countries are: Rymut
(1990–4, 1999–2002) on Polish names; Beneš (1998) and Moldanová (1983) on Czech sur­
names; Unbegaun (1972) on Russian surnames; Red’ko (1966) on Ukrainian surnames;
and Merku (1982) on Slovenian surnames in north-east Italy. Mention may also be made
here of Maciejauskienė (1991) on Lithuanian surname history and Siliņš (1990) on the vo­
cabulary of Latvian surnames.

15.4 Studies of Jewish Family Names

As Jewish family names belong to members of a large religious community, rather than
the people from a particular country, Jewish family name study requires the analysis of
records from many parts of Europe and the Near East. Jewish family name studies tend to
focus on particular countries or particular Jewish ethnic divisions (notably Ashkenazic vs.
Sephardic), which helps to keep them down to a manageable size. Major works on Jewish
surnames are Beider (1993, 1995, 1996) and Menk (2005). Jewish names are also well
represented in Hanks and Hodges (1988) and DAFN.

Page 19 of 24
Family Names

15.5 Family Names in Asia

15.5.1 China

Chinese surnames are much older than those from other countries, in some cases reput­
edly dating back to the third millennium BC, to the time of the legendary ‘Yellow (p. 233)
Emperor’ Huang Di, and before. The earliest known account of Chinese surname origins
is written in Shi Ben, from the Warring States period (475–221 BC), but it is not clear
whether the names in this, and other such early writings, were borne by people who lived
at the time or have simply been drawn from characters of Chinese myths and legends.
Even so, it seems clear that surnames emerged in China during the Western Zhou dy­
nasty (1046–771 BC) and the Spring and Autumn period (770–476 BC). Most of today’s
Chinese surnames have their origin in the Han people, an ethnic group originally from
North China, who migrated across much of the country and whose culture was adopted
by many other ethnic groups.

The most comprehensive reference work for Chinese surnames available today is by Yida
and Jiaru (2010). It includes a collection of 23,813 surnames from historic sources, also
containing names that do not have their origin among the Han Chinese people. Corre­
sponding English spellings are also provided alongside the ideographic Chinese sur­
names. It is worth noting, however, that the central core of conventional Chinese sur­
names consists of only a few hundred items. An extensive work on the genealogical ori­
gins of Chinese surnames is by Chao (2000), which also provides information on etymolo­
gy and the geographical distribution of surnames in China today.

15.5.2 Japan

Two scholarly and comprehensive works on the names of Japan are Niwa’s (1981)
etymological study and his (1985) dictionary of Japanese surnames.

15.5.3 Korea

While studies of the etymologies and histories of Korean family names are few, genealogi­
cal information has been published by clan organizations for the majority of the 260 or so
Korean surnames, and is available in collections such as Han’gukin ŭi Sŏngbo: Ch’oidae
Sŏngssi wa pon’gwan (Korean Genealogies: updated surnames and clan seats).

15.5.4 The Indian Subcontinent

The contributions of Professor R. V. Miranda to DAFN and FaNBI have given a tantalizing
glimpse of the rich variety of historical, cultural, religious, and linguistic facts that can be
gleaned from the study of family names in the countries of the Indian subcontinent. Re­
grettably, however, there does not seem to be any immediate prospect of either a scholar­
ly or even a popular study of family names in India, Bangladesh, or Sri Lanka. Names in
Pakistan are accounted for by Ahmed (1999) insofar as they are of Muslim religious affili­

Page 20 of 24
Family Names

ation and therefore Arabic or Persian etymology. Schimmel (1989) offers a richly informa­
tive discursive study of Muslim names and culture.

15.6 International and Comparative Sur­

(p. 234)

veys
A Dictionary of Surnames by Hanks and Hodges (1988) is a dictionary with different aims
and different scope from any of the works mentioned so far. Rather than taking medieval
records as the starting point for compiling a list of surnames, Hanks and Hodges used
modern data collected from selected 1980s telephone directories. They also attempted
systematic coverage of European surnames. Their target audience was the whole English-
speaking world and beyond, not just the UK. People migrate; they move around; so a
modern study of family names must not be insular or parochial. It must contrast local sta­
bility with national and international patterns of migration. Therefore, this dictionary was
the first to contain entries for surnames from all over the European continent, including
extensive comparative lists of cognates and derivative forms in different languages. This
work was well received in North America, where the majority of surnames are of non-
English origin. This led ultimately to a larger and better-focused project, the Dictionary of
American Family Names (DAFN), which is the standard reference work for family names
in the USA. Because the USA has a great mix of names from many different countries,
this dictionary included contributions by onomastic and linguistic scholars from all
around the world. As a result, it is not just a source of information for those interested in
the names of the USA, but also a reference work with worldwide relevance. It is pub­
lished in three volumes, and contains over 70,000 entries drawn from the computational
analysis of over 88 million names of US telephone subscribers.

15.7 Data Analysis—Documents and Distribu­

tion
15.7.1 Geographical Distribution of Surnames

An important new approach to the study of surnames was developed by Guppy (1890),
whose work showed that there was often an enduring connection between a surname’s
present-day distribution and its place of origin. Guppy’s approach was an important pre­
cursor to the present-day analysis of surname distribution, as in Archer’s (2011) 19th
Century Surname Atlas, which has necessitated revision of what is considered the most
likely origin of many UK surnames, as well as enhancing understanding of how migration
within Britain has affected surname distribution patterns.

Archer’s atlas is available as a CD. It shows the distribution and frequency of each sur­
name recorded in the 1881 British census. Distributions can be viewed both by county
and by poor law union (PLU), both in actual numbers and proportionally (per 100,000 of

Page 21 of 24
Family Names

population). An advanced search option allows the distribution of a selected group

(p. 235) of names to be shown, so that patterns of a certain surname feature, rather than

just an individual name, can be plotted. This approach can be used to further our under­
standing of the distribution and origins of regionally specific naming features and pat­
terns in the UK.

A comparable CD resource for the names of Ireland is Grenham (2003), which gives the
distribution of surnames drawn from a variety of nineteenth-century sources. Most of the
data are organized by household, and so distribution maps are not quite as detailed as
those in Archer’s (2011) British atlas, although Grenham’s Irish atlas is still a valuable
onomastic and genealogical resource.

The distribution of a family name is information that is not just of use to genealogists,
who can sometimes uncover the probable geographical origin of a surname through such
information, but can also inform linguistic study. Medieval dialect lexis and phonology is
preserved in many present-day surnames, so an analysis of their national distribution can
aid investigation of historical dialects. By comparing family name data from different pe­
riods, the continuity or change of dialect distribution can be studied. Barker et al. (2007)
show the value of this approach in their analysis and comparison of surname distribution
using records from the sixteenth century to the present day.

15.7.2 Computational Analysis of Large Databases

There is an ever-increasing availability of large digitized surname databases from differ­

ent periods, which are only just beginning to be systematically analysed. Resources such
as the International Genealogical Index (IGI) (FamilySearch 2012), which contains hun­
dreds of millions of UK parish register entries, could in principle be used for the statisti­
cal analysis of surname frequency, geographical distribution, and the frequency and dis­
tribution of particular surname features. Such analysis could, for example, lead to a more
accurate account of distributional contrasts, of which an already known example is that
between patronymic surnames ending in -son, which tend to be characteristic of North­
ern England and Scotland, and patronymic surnames ending in -s, which are more typical
of the South Midlands and Wales. Another known example concerns the distribution of
locative surnames ending in -er, such as Chalker and Streeter, which are characteristic of
Kent, Sussex, and Hampshire. Many similar linguistic features of surnames are no doubt
waiting to be discovered as more and more data become available for computational
analysis.

Certain historical records with representative national coverage have only recently be­
come available, such as the English fourteenth-century poll tax returns (see Fenwick
1998–2005) and the Irish Fiants (see Nicholls 1994). The large amount of data these
records provide can be analysed computationally, so that national family name distribu­
tion patterns and changes over a number of centuries can be discovered, which will fill a
large hole in our knowledge of surname development. Rogers (1995: 224) recognized the
importance of the fourteenth-century poll tax returns specifically for this purpose, stating

Page 22 of 24
Family Names

‘it is . . . clear that, the rarer the name, the less likely it is that the distribution of (p. 236)
its early examples will be visible in the fourteenth-century sources until the Poll Tax be­
comes widely available’.

Now that these records, and other large collections of family name data, are accessible,
and historical records are being continually digitized, computerized systematic analysis of
family name characteristics can be carried out on a much larger scale than has been pre­
viously possible, to give a more complete picture of surname development than is current­
ly available. To this end, medieval and early modern spellings of surnames will need to be
linked, drawing on the expertise of philologists, historians, and demographers.

Demographic studies of, for example, the rates of surname death over time and the ef­
fects of migration from region to region, as well as from country to country, on surname
development and change, will also become possible, but only when even larger quantities
of machine-readable data from many different periods are available for comparison.

Such approaches will require careful consideration of many different sources of varying
onomastic value, but this kind of research will greatly improve our understanding of fami­
ly name distribution and history, for example through tests for statistical significance in
the relationship between a surname’s geographical origin and its distribution at different
periods, in order to determine the extent to which its distribution can be taken as a reli­
able indicator of its geographical, historical, and linguistic origins. It is therefore hoped
that future work in the field of family name studies will systematically analyse very large
digitized datasets, using techniques that have been developed in corpus linguistics among
other subject areas, potentially leading to important new discoveries on many different
aspects of family names and naming.

Notes:

(1) The term surname, which used to mean no more than ‘additional name’, is now used
interchangeably with family name.

(2) Notably among the so-called ‘Strathclyde Britons’, who, up to at least the fourteenth
century, spoke a language closely related to Welsh and lived in an area around the lower
Clyde valley.

Patrick Hanks

Patrick Hanks is Lead Researcher on the Family Names Project at the University of
the West of England. He is Editor-in-chief of the Dictionary of American Family
Names and the forthcoming Dictionary of Family Names in Britain and Ireland (both
Oxford University Press). He is co-author of the Oxford Dictionary of First Names. In
addition, he holds a part-time position as Professor in Lexicography at the University
of Wolverhampton. From 1990 to 2000 he was Chief Editor of Current English Dictio­
naries at Oxford University Press.

Page 23 of 24
Family Names

Harry Parkin

Harry Parkin is Research Associate on the Family Names Project at the University of
the West of England, Bristol. He is a linguist with particular interests in the history of
English surnames, the methodology of surname research, and the use of historical
onomastic data in philology, demography, and Middle English dialectology.

Page 24 of 24