Patrick Hanks

What's in a Name?

DAFN, The Dictionary of American Family Names, is the first attempt to explain the history and origin of the 70,000 most frequent family names in the United States, together with some rarer names that are of historical or etymological importance.

Each of us has a family name, generally inherited via our parents from our forebears (traditionally through the male line); however, most of us know remarkably little about how, when, and where our family name came into existence, and what it originally meant.

The main purpose of this book is to fill this gap in a systematic way: to provide outline information about where, when, and how our family names originated. Popular interest in family names and their origins has never been greater. Over 85% of Americans will find an entry for their surname in DAFN.

Until a few years ago, such an undertaking would have been impossible. Now, thanks to online resources, worldwide scholarship, and computer technology, the task can be attempted. Many American family names, even quite frequent ones, have never been studied at all. In such cases, DAFN has, wherever possible, carried out at least enough research to explain the language of origin of the name, and where possible its etymology. Until very recently, names scholarship has been scarce, and sometimes well‐intentioned efforts by untrained amateurs, especially in the 19th and early 20th centuries, have resulted in the perpetuation of false or misleading conclusions. In DAFN, folk beliefs about name origins, which are pervasive, have been discounted or identified as such, though no doubt some may have slipped through. Where the origin of a name remains unidentified or the etymology unexplained, despite the compilers' best efforts, DAFN says so. If an explanation can only be given tentatively, DAFN uses hedges such as “probably,” “perhaps,” and “possibly.” In surnames research there are very few certainties, however, so explanations that are highly probable are not hedged.

Uncertainty is expressed only where an explanation is genuinely doubtful.

The Structure of DAFN Entries

The entries in DAFN are structured as follows:

Headword: the surname as found in present‐day American name lists, with normalized spacing and capital letters, and no accents or other diacritics.

Comparative Frequency: in parentheses immediately after the headword, giving the frequency of the headword in DAFN's 88.7 million name sample (one‐third of the 1997 U.S. population). See the following article by Ken Tucker for details of DAFN's sampling, frequency, and selection policies.

Explanation: the source language and origin of the surname. Within each explanation are included at least some of the following:

Language of Origin: the language or culture in which the surname originated.

Original Spelling: if the surname is written with accents or diacritics in the source language, the accented form is shown as part of the explanation.

Region: if the surname is known to be strongly associated with a particular region of the mother country, DAFN gives this information too.

Typology: a classification of the surname's origin, indicating for example whether it is derived from a place name, from a nickname, from the personal name of a forebear, or from a word denoting the occupation or social status of a forebear.

Etymology: the linguistic history of the surname, generally including the form and meaning of the word from which it is derived.

Cross‐references: references, printed in small capitals, to other entries, where more details may be found. There are many, many variant spellings of surnames in America. DAFN treats each major variant as a main entry in its own right. This means that the dictionary is “self‐indexing”; the explanations contain an intricate network of cross‐references to so‐called anchor names, where a fuller explanation of the linguistic, cultural, or historical background is generally given.

Given Names: a limited selection of the most significant “diagnostic” given names (if any) that are associated with the surname in DAFN's sample of 88.7 million American names, together with a statement of the language or languages for which these given names are diagnostic:

Diagnosis and Confidence Measure: a statement of the probable language of origin (if not English) on the basis of the given‐name evidence, together with a measure of confidence in the diagnosis, expressed as a percentage. “French 6%” means that, on the basis of given‐name evidence, we can have 6% confidence that the surname is French. It does notmean that 6% of all bearers are of French origin. See the following article by Ken Tucker for a fuller explanation of diagnostic given names and confidence measures.

Forebears: for a few selected entries, a paragraph on some particularly important aspect of family history or ancestry.

Typically, the earliest known bearer in North America of a particular surname may be mentioned. In other cases, forebears are mentioned because they shed light on changes in the linguistic form of the surname. In a very few cases, comments are added on the ancestry of important figures in American history: generally, only the ancestry of presidents and a few other figures of similar historical significance.

In a general work of reference such as DAFN, it is of course impossible to do more than mention a tiny number of selected forebears of particular linguistic or general historical significance for certain names. Systematic research into immigrant forebears is not one of DAFN's objectives. Research into immigrant ancestors starts with ships' passenger lists and immigration records at relevant ports of entry. The Immigrant Ships Transcribers Guild ( is a good starting point. People sometimes speak as if all immigrants came through Ellis Island, which was the immigration processing center for the port of New York. But of course Ellis Island was only one of dozens of ports of entry and its records cover only a 32‐year period (1892–1924). Nevertheless, more than 22 million people, most of them immigrants, came through Ellis Island. Millions of others came through Philadelphia, Boston, Baltimore, New Orleans, and many other ports from the 17th century onward, not to mention the West Coast ports of San Francisco and elsewhere and the border crossings from Canada and Mexico.

Spellings of Family Names

For family historians and genealogists, DAFN does not provide all the answers: far from it. Rather, it provides a starting point, a framework within which family history can be conducted. It provokes questions which can only be answered by dint of detailed research on individual names, research that would be impractical in a work of general scope such as DAFN. Even with this constraint, very large number of the names in this book have never been researched before, so the editors and contributors were forced to do far more primary research than was originally envisioned.

DAFN gives the basic facts about family names insofar as they are known, summarizing where the name came from, what it meant at the time when it first came into existence, and, where relevant, its history. This turned out to be a massive task.

Many names have more than one origin, so research could not stop when one satisfactory explanation has been achieved. All reasonable possibilities had to be considered and explored.

Furthermore, the form of some American family names has been altered almost beyond recognition in the course of transmission from generation to generation and from place to place, including adaptation to the English‐speaking world. In such cases, the work of individual family historians and names researchers, too numerous to mention individually, has sometimes been of great value. By painstaking primary historical research, family historians have sometimes succeeded in linking a modern American surname to an 18th‐century German, Dutch, or French original that may have been very different in spelling and pronunciation from the present‐day name. In such cases, DAFN reports the link, and either cross‐refers to an existing entry where an etymological explanation is given, or, if no cross‐reference is possible, an explanation of the etymology (if known) is given at the new American form.

DAFN has surveyed current surname scholarship, including indigenous scholarship in languages other than English, and summarized the results in a readable form. As far as possible, the present‐day form of each name in the book is traced back to its original meaning in the country or region that it came from.

In addition, information is give about the comparative frequency of names and, in selected cases, the earliest known bearers of the name in North America.

Sometimes the trail tracking a name to its linguistic origin is long and tortuous: for example, to explain the origins of Sicilian American names requires expertise not only in Italian, but also in Greek, Catalan, and Arabic, which are important source languages for surnames in Sicily alongside Italian. To explain names that come from the British Isles, expertise is needed, not only in Middle English, but also in Old Norman French, Old Norse, Welsh, Cornish, Irish and Scottish Gaelic, and Manx. For this reason, DAFN is very much a team effort.

No one individual could possibly explain the origins of all the surnames of America, deriving as they do from all the world's languages. DAFN is very much a cooperative effort among leading names scholars from many different parts of the world.

Further examples of the complex trails of name origins can be given. The Dutch name Fonda is of Spanish or Catalan origin, having been brought to the Netherlands in the 16th century via the Italian port of Genoa. Names from the Iberian peninsula are not only of Spanish and Portuguese linguistic origin, but also Catalan and Galician, while a surprisingly large number of Hispanic surnames, including some very frequent and historically important ones such as Mendoza, Ortiz, and Aguirre, are ultimately from Basque, the language of the western Pyrenees and adjacent regions, which is unrelated to any other language. Many of the oldest English surnames are not merely of English origin, but come from Old French, Old Norse, medieval Welsh or Cornish. Scottish names are not only English, French, and Old Norse, but also Gaelic. Many names that are Scottish in form came to North America via northern Ireland. Gaelic names were invariably Anglicized before being brought to North America, often in ways that involved radical alterations in form.

People sometimes claim that one spelling of a name is “more correct” than another. There is no historical or linguistic argument to support such claims. The correct spelling of a name is whatever the family members say it is. If a family prefers to spell its name in a way that reflects a linguistic origin or etymology, or the spelling as found in the country of origin, that is their choice. An equally valid choice is to spell the family name in a way that reflects the development and/or pronunciation of the name in English‐speaking America. DAFN explains origins, but origins must not be confused with correctness.

Changing Society: Feminism and Family Names

In Europe, North America, and elsewhere, from the 12th century up to the 1960s, marriage almost invariably involved the adoption of the husband's family name by the wife, while a child's family name was inherited from the father not the mother.

In recent years this has begun to change, and family historians must take cognizance of the fact that the traditional notion of the family, which (at least as an ideal) has been comparatively stable for almost a thousand years, is now changing substantially.

It is by no means clear how many of the changes to social institutions such as marriage that have occurred in the past forty years and are going on even now will become permanent, how many will be abandoned, and how many will be carried further forward. Certainly, current conventions of childbearing expose the ludicrous irrelevance to the modern world of the literal meaning of the old word bastard. In the past, birth out of wedlock was regarded as a sufficient reason for ostracization of both mother and child (though, curiously enough, not the father), but now it is no longer stigmatized, and the term survives only as a meaningless term of abuse or, in Australia, as a cheerful term of reference meaning “regular guy.” Childbirth is no longer an unavoidable concomitant of sexual pleasure; marriage is no longer a commitment entered into early in life, “for better for worse,” a sort of desperate lottery in the face of the ever‐present possibility of an early death.

Instead, marriage is now often entered into by adults who may already have had children, as a signal that they have established what they hope may be a stable relationship. These changes are particularly noticeable among people whose priorities include development of a career—for among conservative families, especially rich ones, marriage as a form of social and political alliance continues to thrive much as it did in the past.

Due in part to the fact that many women now aspire to a career as well as a family, and in part to birth control, enabling families to be planned rather than accidents of nature, the role of childbearing in the marriage contract has diminished. The grinding cycle of annually renewed pregnancies, interspersed with household management, infant mortality, puerperal fever, and concluded by an early death, is no longer a married woman's normal lot, as it all too often was in the past.

Marriage is no longer an essential preliminary to childbearing, and even when people do marry, among many people it is now considered distinctly old‐fashioned or unchic for a woman to take the husband's surname. This results in a dilemma for the naming of children. No longer does a child automatically inherit the father's surname. Deciding on a surname has become a matter of choice, not fate. Many children still do receive their father's surname, but others are named with the mother's surname or receive both, joined by a hyphen. Whatever the social implications of these changes, they will certainly add a research dimension for future family historians. Fortunately, however, modern vital records are generally more detailed, more reliable, and more publicly available, than those of past centuries, so the problem may not be as acute as it may seem at first sight.

When Did Family Names Originate?

The obvious answer to the question, “When did family names originate?” is that they originated at many different times and in many different places. Nevertheless, some broad trends may be distinguished. The phenomenon of the modern family name is, broadly, of medieval European origin and is associated with the rise of bureaucracies. Tax collectors, beadles and bailiffs, sheriffs and reeves (officers of law enforcement and the courts) had, like their modern counterparts, a particular need for the precise identification of individuals. In a period when the eldest son inherited property and titles, it became normal to inherit the father's surname along with other property.

In Europe family names were unknown before the 11th century (although some clan names existed, which later became family names, for example in Ireland). Within about 250 years, i.e. by the beginning of the 14th century, surnames were established in most of the countries of central and western Europe.

By 1300, hereditary surnames were the norm and patronymics were the exception. However, they were less stable than they are today. A person might adopt or acquire a new and different surname at any time.

Older naming traditions, in particular the patronymic tradition, have either been assimilated, in Europe, America, and elsewhere, into the system created by medieval European feudal bureaucracies, or else have been replaced by it.

The Ancient Names of East Asia

Undoubtedly, the oldest surnames in this book are those of Chinese origin. The majority of Chinese surnames are around three thousand years old, while some are believed to date back to the “model emperor” period that began almost fifty centuries ago. Although modern Chinese Americans bear modern American surnames that are derived, linguistically, from these extraordinarily ancient Chinese names, we should not assume that a surname in China had the same function and social status as a modern American surname. For further information about Chinese names and naming, see the article by Mark Lewellen in this volume.

Japanese and Korean surnames are much less ancient than Chinese names, but still older than European names. They originated in the 5th century AD and the 1st century BC respectively. Details are explained by Fred Brady and Gary Mackelprang in their articles below.

At the time when surnames were being formed in East Asia, Europeans surnames as we know them had not yet been formed.

Names and Naming in Ancient Rome

It is hard to know whether to be more impressed by the similarities or the differences between Classical Latin names and modern personal names. Ancient Roman names typically had three parts: a praenomen(equivalent to a modern forename), a nomen(a clan name, somewhat similar to a modern family name), and thirdly a cognomen, a nickname or distinguishing name that was attached to a particular individual, but which could be inherited. When a cognomen was first coined, it was known as an agnomen; the term cognomenoften implies some sort of hereditary use within a family. Thus, the term crassusmeans ‘thick‐set’, and was used as a cognomen by descendants and family members (who may themselves have been tall and slim) of a thick‐set individual who acquired the name as an agnomen with reference to the shape of his person.

In Republican Rome, only a few dozen praenomina were in regular use: Marcus, Lucius, Gaius, Publius, etc. The number of family names or clan names was not very much larger: Julius, Antonius, etc. This made the adoption of a distinguishing cognomen all more important as the size of the population of ancient Rome grew. Thus, contrary to popular belief, the forename of the founder of the Roman empire was not Julius: that was his clan name. His praenomen or forename was Gaius. And the name by which he is universally known, Caesar(which has given rise, indirectly, to numerous European family names and vocabulary words meaning ‘ruler’ or ‘emperor’ e.g. Russian tsar, German Kaiser), was in origin a cognomen or inherited family nickname meaning ‘fine head of hair’. The famous Roman military commander Publius Cornelius Scipio (236–184 BC) was a member of the Cornelii clan, with the hereditary cognomen Scipiomeaning ‘rod or staff’. Later, he acquired the agnomen Africanus, with reference to his victories over the Carthaginians in North Africa, in a way not dissimilar to the way in which, over two thousand years later, the British generals Montgomery and Alexander were granted the titles “Viscount Montgomery of Alamein” and “Earl Alexander of Tunis,” in recognition of theirvictories in North Africa.

Other well‐known Roman cognomina were nicknames such as Ahenobarbus‘red beard’, Flaccus‘big ears’, and Posthumus‘born after the death of his father’. There have been attempts to relate modern Italian surnames to Classical Roman names, but in reality, the naming system of ancient Rome was swept away by the new naming system introduced by the early Christian Church. Ancient Roman names survived only if, as in the case of Marcusand Antonius, they were also borne by early Christian saints.

Patronymic and Genealogical Naming Systems

Before hereditary surnames became established, a patronymic or genealogical system was the norm throughout central, northern, and western Europe. People were known by their parentage, sometimes supplemented by a nickname. These names were not hereditary. Patronymic naming survives to this day in Iceland, where an individual called Hermann Pálsson is the son of someone called Pál. Alternation between generations is common, so this Pál may well be Pál Hermannsson. Elsewhere in Scandinavia, the patronymic system survived into the 19th century, and there are records of the conservative mirth that greeted the notion that a woman might be known as Anna Andersson (as opposed to Anna Andersdottir—‘Anna, daughter of Andrew’).

In the Old Icelandic sagas, a rich mixture of patronymics and nicknames is found. Thus, Burnt Njal's Saga begins with a man whose name is Mord Fiddle, the son of Sighvat the Red. Before long, the saga mentions a man called Hoskuld, who is given a much fuller genealogical name: Hoskuld Dala‐Kollsson, son of Thorgerd, daughter of Thorstein the Red, son of Olaf the White, son of Ingjald, son of Helgi and of Thora, daughter of Sigurd Snake‐in‐the‐Eye, son of Ragnar Hairy‐Breeches; Thorstein the Red's mother was Aud the Deep‐Minded, daughter of Ketil Flat‐Nose, son of Bjorn Buna.

In this fine genealogical name, no fewer than seven generations are remembered, with two matrilineal branches and six memorable nicknames.

As in Scandinavia, so also in Wales the patronymic system was in regular use up to the 19th century, long after hereditary surnames had been adopted elsewhere in Britain and Europe. John and Sheila Rowlands mention the example of a certain 16th century gentleman who was David ap William ap David Lloyd ap Tomas ap Dafydd ap Gwilym ap Dafydd Ieuan ap Howel, ap Cynfrig ap Iorwerth Fychan ap Iorwerth ap Grono ap Tegerin—12 generations of forebears in a single name! Knowing one's ancestry was important for establishing one's right to the inheritance of land and for other legal purposes.

Examples such as those just given may be colorful, but they are also cumbersome, and it is easy to see why they came to be replaced in most places by the streamlined efficiency of a single hereditary surname, passed down from one generation to the next.

The Rise of Hereditary Surnames

The most ancient hereditary surnames in Europe are probably those that developed from ancient patronymics or clan names. In Ireland surnames such as Ó Conaill and Ó Néill date back at least to the 10th century and probably beyond. It is only a short step from a name that means ‘descendant of Niall’ to a hereditary surname.

Elsewhere in Europe during the 11th century, rich and powerful families began to take surnames based on their lands and estates. More humble folk, too, could be known by the name of the place from which they came, if (as was the exception rather than the rule) they ever moved away from that place.

Other hereditary surnames were derived from nicknames, from occupations, and from features of the landscape, as explained in the section on Typology below. It is interesting (and perhaps a little surprising) that, with few exceptions, all the main types of surname are represented in all the languages of Europe. An example of an exception is that there are no Irish Gaelic names of habitational origin; this is probably due to the fact that hereditary Irish Gaelic names are always patronymic in form. The absurdity of a name appearing to state that someone is the son of a place name would be patent.

There are, of course, local differences in proportion among the different types of surname in different countries. Thus, ornamental names (names made up arbitrarily from vocabulary words with more or less pleasant associations) are characteristic of Swedish and Jewish surnames. In Norway habitational names, often deriving from a single farm, are very common. In Italian and Czech, derivatives of pet forms of personal names are especially common.

From an American perspective, it is important to note that most European surnames were well established two or three hundred years before the first European settlements in North America, and had ample opportunity to undergo changes long before they came to this country. The serious student of the origin of a surname must not only know the language of the country of its origin, but must be acquainted with the medieval forms of that language, the medieval meanings of its words, the nature of medieval professions (for many surnames denote occupations that have long since vanished), and the structure of medieval society, for some surnames can only be understood in the context of the medieval feudal system.

Hereditary surnames did not become fixed overnight. In fact, in many records from the 11th to the 14th century it is impossible to be certain whether a surname is hereditary, or merely a distinguishing epithet. Did a medieval Richard Skinner inherit the name Skinner from his father, or does it merely indicate that his occupation was skinning animals and making leather goods? If we find that Richard Skinner had a son called John Richardson, it is clear that in this particular case the appellation was not established as a hereditary surname. On the other hand, if we find the surname used alongside an appellation denoting an occupation, e.g. ‘Richard Skinner the Baker’ or even ‘Richard Skinner the Skinner’, it is clear that it has become a hereditary surname.

Often, there is variation, and the same individual may be known by more than one name. To take an actual example, Reaney notes in Cornwall in 1297 a certain Johannes Gyffard dictus le Boeuf(‘John Gifford, called the Bull’). His descendants, if any, could equally well have come to be known by the surname Gifford or Bull. Unfortunately, records that would enable us to establish hereditary status of these early names in individual cases are few and far between. The process is clear, but the details are lacking. The surname historian must therefore rely to a large extent on probabilities and inferencing from early records.

By the 14th century, hereditary surnames were well established in most parts of continental Europe and the British Isles. And in the 17th century they began to be brought to North America, ready for a new chapter in their development.

Adoptions from Other Naming Systems

In 20th‐century America, a hereditary surname is more or less obligatory, for administrative reasons. Immigrants from places where there is no indigenous tradition of hereditary surnames, such as Islamic countries or parts of the Indian subcontinent, have very often adapted their personal names to the American system. Thus, Muhammad and Abdullah are found as American surnames, in addition to being Arabic personal names found throughout the Muslim world.

By a similar process, among people from southern India personal names such as Ganesh, Hari, and Murthy have been pressed into service as surnames in an American context, although they are found only as personal names in southern India.

For a few cultures, it was possible to find neither reliable reference works nor a scholar able or willing to explain the history and meaning of the names. All attempts to find a means of explaining Vietnamese, Laotian, Cambodian, and Thai names and naming practices were unsuccessful. Where the evidence points unmistakably to a Vietnamese or other unresearched East Asian origin for an American surname, this fact is noted in DAFN, even though an etymological explanation cannot be given. The total number of names affected is small: under 300 entries altogether.

In other cases, the reader will look in vain for surnames from a particular country or culture. For example, very few American family names of Turkish origin are recorded in DAFN. This is no doubt partly because Turkish family names as such did not exist, when they were introduced by decree of Mustafa Kemal Atatürk. Being of comparatively recent origin, they are very diverse, so that each one has few bearers, and fewer still have come to America. Fewer than fifty Turkish American family names have reached the selection criterion of 100 bearers in a population sample of 88.7 million. Traditional naming practices in Turkey involved the use of a personal name with an honorific such as bey, rather than a surname, and these traditions survive in everyday use in Turkey today. Thus a man whose name (with surname) is officially Önder Renkliyıldırımis normally known in everyday use as Önder Bey.

African American Family Names

In addition, it must be noted that DAFN contains almost no entries of African linguistic origin. With few exceptions (Ethiopian names being prominent examples), names from the languages of Africa have not reached the U.S. in sufficient numbers to have become established as American family names.

There is no established tradition of names from African languages being used as American surnames. Of course, the vast majority of African Americans bear names that are of English, Scottish, or Irish linguistic origin, these mostly coming from the names of the owners of plantations on which their forebears worked. In other cases, African Americans forebears have adopted the surnames of illustrious historical figures such as Washington and Lincoln. In a few cases, it has been possible to say that a particular surname, or a particular spelling of a surname, is borne mainly by African Americans, but for the most part African American surnames are inextricably mixed with surnames of British origin. This fact serves to emphasize that the history of surnames, a branch of philology or historical linguistics, is independent of genealogy, the history of families.

The difficulties of disentangling African American surnames from Scottish American and English American ones in no way inhibits African American genealogical research. In recent years, several African American genealogical societies and some excellent web sites have been established, for example Afrigeneas ( and the African‐American Genealogy Ring (

American Indian Names and Naming Systems

It is a matter for regret that it has not been possible to record many details of American Indian names in DAFN, names from native American cultures. This is partly due to the enormous diversity of these cultures, partly to the decimation of the American Indian population from the 17th and early 20th centuries, so that numbers are now small, and partly to the fact that many if not most American Indians have adopted English or Spanish surnames.

DAFN contains a few entries for Navajo names, but very little else. For the most part, names of American Indian linguistic origin are rare. Almost none of them reach the selection criterion of 100 bearers in a population sample of 88.7 million that qualifies a name for entry in DAFN. Here, too, the interested researcher must be directed to native American genealogical societies and relevant websites.

Typology of Family Names

Family names may be classified into a small number of types according to their origin. DAFN uses the classification established in Hanks and Hodges (1988), which is explained below.

Patronymic Names

The oldest and most pervasive type of family name, with many of the highest frequencies, consists of those derived from the personal name of an ancestor. Two main strands in the origins of personal names can be identified: vernacular naming traditions and religious naming traditions. Vernacular names were originally composed, thousands of years ago, of vocabulary elements in a prehistoric form of a language. No doubt, while their meaning was still clear, they were bestowed for their auspicious connotations (e.g. Raymondis derived from ancient Germanic elements meaning ‘counsel’ and ‘protection’). In religious naming traditions, on the other hand, names were bestowed in honor of a cult figure. Aside from Jewish names, the most powerful religious influence on naming in Europe has been the Christian Church. There is not a country in Europe that does not have surnames derived from forms of John, Matthew, Mark, Luke, Peter, Paul, and other saints, apostles, and missionaries of the Christian Church. It comes as something of a surprise, therefore, to note that in many countries, especially in northern Europe, baptismal names honoring Christian saints and biblical figures were a fairly recent introduction, at around the time when the bulk of family names were taking shape. These Christian names were in competition with the older and better‐established vernacular naming traditions.

Surnames derived from ancient Germanic personal names are found not only in German, but also in Dutch, English, French, and many other languages. The court of Charlemagne (742–814) played an important role: it was Christian and Latinspeaking, but the vernacular language at that time was the Frankish dialect of Old High German, and the personal names in use were Germanic rather than Latinate. These personal names were adopted in many parts of northwest Europe, especially among the ruling classes. They were in use among the Normans; hence, many common English and French names such as Richard, Robert, and William (Guillaume)are of Germanic origin and have cognates in other European languages.

Some Germanic personal names such as Siegfriedalso have Slavic derivatives, but on the whole the Slavs had their own inventory of personal names. In western Slavic‐speaking areas (in particular, in Poland, the Czech lands, and Slovakia), these native Slavic names have also given rise to family names. In Russia, on the other hand, vernacular Slavic names were proscribed as baptismal names by the Orthodox Church in favor of names honoring Christian saints. For this reason, Russian patronymic surnames are mostly derived from saints' names rather than from vernacular Slavic names.

The most basic type of surname derived from a patronymic— that is, from a person's father's given name—simply presents the father's name as a distinguishing epithet alongside the bearer's own given name. Surnames of this type are found in almost all European languages, but in most of them they are considerably less common than names formed with explicitly patronymic endings.

The range of affixes which have been utilized with a patronymic function is very wide. Some are prefixes (Gaelic mac, Welsh ap, ab, Norman French fitz, Italian fi‐); others (the majority) are suffixes. These were for the most part originally adjectival or possessive in function (English ‐s, North German ‐ingand ‐er, Romanian ‐esco, Russian and Bulgarian ‐ov), or else result from a more or less reduced form of a term meaning ‘son of’ (English ‐son, Danish/Norwegian ‐sen, Swedish ‐son).

In such cases the surname was almost always originally patronymic in function, although the reference seems occasionally to have been to a grandfather or more distant relative, and in some early examples women are known to have acquired the given name of their husbands as a distinguishing epithet. It is likely that some hereditary surnames are derived from this use.

In this category also belong surnames that are derived from shortened or familiar forms of given names, pet forms, and forms with diminutive suffixes. In the Middle Ages such forms were in common use, often almost to the exclusion of the official baptismal form, hence the frequency of such common English surnames as Hobsonand Dobson, based on vernacular forms of the baptismal name Robert, or the equally common northern and central European derivatives of Johannor Hans, German forms of Latin Johannes (John), or the great profusion of Italian and Czech surnames derived from diminutive forms of given names.

Metronymic Surnames

Much less common than patronymics, with no more than a handful of surviving examples in the majority of European languages, are metronymics, derived from the name of the first bearer's mother. Since European society has been patriarchal throughout the historical period, it was the given name of the male head of the household that was normally handed on as a distinguishing name to successive generations. The few exceptions (e.g. Catling, Marguerite, Dyott) seem to be derived from the names of women who were either widows for the greater part of their adult lives, or else heiresses in their own right.

In this respect Jewish naming practice differs from that of the rest of Europe, since metronymics are common among Ashkenazim (see, for example, Chaikin, Dworkin, Sorkin, and Rifkind). There are several probable reasons for this: (a) both before and after surnames came into use, Ashkenazic Jews often used nicknames, many of them consisting of a parent's given name plus Yiddish possessive ‐s; many of the nicknames containing the mother's given name presumably gave rise to metronymic family names; (b) in other cases, these nicknames consist of the spouse's given name plus Yiddish possessive ‐s, hence men could have taken these nicknames as family names and passed them to their descendants; (c) it is probable that children of deserted mothers (or widows) took family names based on the mother's given name. In connection with (b), there is a class of surnames which seems to exist only among Ashkenazic Jews, indicating explicitly the husband of the woman named, for example Esterman ‘Esther's husband’. In other cases, we cannot tell whether Ashkenazic family names belong in this category or not: Roseman, for instance, might be one of these names (cf. the Yiddish female given name Royze, Rose) or it might be merely an ornamental name. Perlman is even more complicated: it could be one of these names (cf. the Yiddish female given name Perl, Pearl); it might be an ornamental name; or it might indicate someone who dealt in pearls (though this last possibility is the least likely because the relative high frequency of the name is at odds with the small number of Ashkenazic Jews who dealt in pearls).

A fuller account of the origin and meaning of conventional given names may be found in Hanks and Hodges (1990, 2001) and in Pickering (1999).

Kinship and Other Connections

A few surnames derive from some other relationship between the bearer of the surname and the bearer of a given name that forms part of it: for example employment (e.g. Bateman ‘servant of Bartholomew’), connection by marriage (e.g. Hickmott ‘Richard's in‐law’), or residence in the same dwelling (e.g.Anttila ‘person from Antti's farm’).

A small group of surnames, with representatives in most European languages, identifies the bearer simply by a word denoting a family relationship (e.g. Neve, Neave, Neff ‘nephew’ or ‘cousin’, Eames ‘uncle’, Ayer ‘heir’), presumably to some well‐known local figure.

Surnames from Lack of Kin

A small but interesting group of surnames are those borne by foundlings, children abandoned by their mothers due to the social stigma of bastardy or simply inability to cope with another mouth to feed. Some surnames based on Christian saints' names are undoubtedly of this origin, being taken from the name of the patron saint of the local church where the baby was abandoned or which ran the local orphanage; however, these are indistinguishable from patronymics.

Other names that have this origin include Dutch Weese, Polish Serota, French Jetté (literally, ‘thrown out’), Italian Innocenti, Comunale (like English Parrish, a name for a child reared at the expense of the community), D’Amore (literally, ‘of love’), Di Dio (literally, ‘of god’), and Esposito (literally, ‘exposed’).

Local Names

Surnames derived from words denoting places may be divided into two broad categories: topographic names and habitational names, explained in the next two sections.

Other kinds of local surnames may refer to counties, regions, the names of islands, and indeed whole countries. As a general rule, the further someone had traveled from his place of origin, the broader the designation. Someone who stayed at home might be known by the name of the farm or house where he lived; someone who moved to another town might be known by the name of village that he came from; while someone who moved to a distant city or another county could acquire a surname denoting the region or country from which he originated.

Habitational Names

Habitational names are taken from the names of towns, villages, farmsteads, or other habitations, most of which existed long before surnames came into being. Habitational names include names derived from the names of individual houses with signs on them (where the surname is also the word for the sign, e.g. Swan, Bell).

It is sometimes difficult, especially in the case of multipleelement names (in England usually a defining adjective plus a generic noun), to be precise about whether a surname is derived from an identifying topographic phrase such as ‘(at) the broad ford’ or ‘(by) the red hill’ or from an established place name such as Bradford or Redhill. It is also possible that in some cases what has been thought of as a topographic name is in fact a habitational name from some unidentified place now lost. In Europe, the geographical distribution and linguistic form of surnames is sometimes a source of evidence for local historians of lost places.

Topographic Names

Topographic names are derived from general descriptive references to a feature of the landscape such as a stream, a ford, a tree, or a hill. Topographic surnames can also refer to a river by its name, or to a man‐made feature such as a castle, a city wall, an abbey, or a church.

Some surnames that are ostensibly topographic, such as Hall and Monkhouse, are in fact occupational, for they originally denoted someone who was employed at such a place, for example at a great house or monastery.

Regional and Ethnic Names

Another category of local surnames comprises those denoting origin in a particular region or country. These tended to be acquired when someone migrated a considerable distance from his original home, so that a specific habitational name would have been meaningless to his new neighbors; he would be known simply as coming ‘from the East’, or ‘from Devon’, or ‘from France’. Many of these names derive from adjectival forms (e.g. French, Dench, Walsh); others are in the form of nouns denoting a person's nationality (e.g. Fleming, Langlois, Moravec).

In some cases these may have been originally nicknames bestowed with reference to the imagined character traits associated with the inhabitants of the region or country concerned, rather than denoting actual nationality. So, for example, someone in England called French may actually have been French, or he may have adopted sophisticated or affected mannerisms and tastes popularly associated with French people and culture.

In other cases, such names could denote a trading connection with the place named, especially in the case of a major trading port such as Danzig.

House Names

Another type of local appellative is the house name, referring to a distinctive sign of a kind commonly attached to houses before the days of numbered street addresses and general literacy. A number of surnames are documented as having this origin. Several Jewish surnames are derived from the names of houses in the Jewish quarter of Frankfurt‐am‐Main, for example Rothschild, (the house of the) ‘red shield’. However, the importance of this category of surname has sometimes been exaggerated; some names that have been so explained are in fact more likely to be nicknames of uncertain significance, or, in the case of Jewish surnames, ornamental names.

Occupational Names

There are many types of surname that are explicitly occupational, in that they refer directly to the particular trade or occupation followed by the first bearer. Buried within this dictionary lies an inventory of the common trades of medieval Europe. These occupations can be divided into classes such as agricultural (e.g. Sheppard, Bouvier), manufacturing (e.g. Smith, Wright, Glover), retail (e.g. Monger, Chandler, Draper), and so on.

They can also be classified according to linguistic criteria. The most basic type of occupational name is represented by words straightforwardly denoting the activity involved, whether as a primary derivative of a verbal roots (e.g. Webb, Hunt) or formed by means of an agent suffix attached to a verb (e.g. Weaver, Baker, Tissier) or to a noun (e.g. Webber, Weber, Weaver, Webster, Potter, Stolarz). Some occupational names are derived from a noun plus an agent noun from a verb (e.g. Ledbetter, Schuster, Stellmacher).

Another very common type of surname refers to a calling by metonymy, naming the principal object associated with that activity, whether tool (e.g. Axe, Pick, Nadel, Swingle, Szydlo) or product (e.g. Ballestra, Brott, Maslow). In other cases, the connection is more indirect (e.g. Daino ‘fallow deer’, denoting a deer hunter).

Particularly in the case of Ashkenazic Jewish surnames, occupational names may have attached to them the explicit suffix ‐man(n)(e.g. Federman, Hirshman). This is also occasionally the sense of German ‐mannand English ‐manoccupational names (e.g. Habermann, Zimmermann, Millman), but this is a sematically complex suffix, with a variety of different meanings.

Another group is similar in form to one type of surname derived from nicknames (see below), but semantically it clearly belongs in the category of occupational names. Members of this group consist of a verb‐stem plus a noun, describing the typical action and object involved in the trade of the person concerned, sometimes in a humorous way (e.g. Catchpole for a bailiff).

A class of names who importance has been understated in the past are those that denoted a servant of member of the household of some person of higher social status, ranging from a master craftsman or a humble tenant farmer, to a member of a religious order, a prince of the church, an aristocrat, or royalty. This is the source, for example, of the English surnames Maidment and Parsons; also, usually, of Prior, Monk, King, and Earl. Many surnames derived from a term denoting a person of high rank probably have this origin. Occasionally, the servant relationship is made explicit, for example by use of the genitive case or by the German and English suffix ‐man(n)in another of its senses. More often, however, the servant relationship is implied, not stated in the form of a surname.

Status Names

One group of surnames, sometimes classified as occupational names and by others as nicknames, are names that originated with reference to social status. These for the most part denote a particular role in medieval society (e.g. Bachelor, Franklin, Knight, Squire). It must be remembered, however, that many names that are ostensibly status names (e.g. King, Prince, Duke, Earl, Bishop) are most unlikely to have denoted a holder of the rank in question. In such cases the name was probably originally borne by a servant of the dignitary mentioned; in other cases it may have been given as an “incident name” to someone who had acted such a role in a pageant or other festivities, or else mockingly to someone who behaved in a lordly manner. Jewish names of this type (e.g. Kaiser, Graf, Herzog) are probably all ornamental only.

Other status names (e.g. Alderman, Beadle, Sherriff, Reeve) denoted social status with a particular administrative function. For such names, it is not possible to make a clear‐cut distinction between status and occupation.

A particular group of status names cannot be understood without reference to the feudal system of land tenure. See, for example, the entries for Ackerman and Huber. In Czech Svoboda literally means ‘free’ but in particular denoted a free peasant farmer in contradistinction to a serf. Dvořák denoted a superior class of farmer, a ‘lord of the manor’. A Sedlák was a slightly lower class of farmer, but with more land than a Zahradník, a smallholder, or a Chalupník, a cottager. Similar traces of an older social hierarchy are preserve in the surnames of several other European languages.

Surnames from Nicknames

Surnames derived from nicknames form the broadest and most miscellaneous class of surnames. To some extent this is a catchall category, encompassing a number of different types of origin. The most typical classes refer adjectivally to the general physical appearance of the person concerned (e.g. Black, Blake, Schwarz, Russell), or personality trait (e.g. Cortes, Hendy, Karg, Kluge). Others point, with an adjective and noun, to some particular physical feature (e.g. Whitehead). Many nicknames refer unambiguously to some physical deficiency (e.g. Mank, Balfe, Bobo), while others may be presumed to allude to it (e.g. Hand, Daum). Others probably make reference to a favored article or style of clothing (e.g. Boot, Cape).

People in past ages were less squeamish and certainly more forthright than we are today, and it will come as a surprise to many Americans that the origin of their surname may draw attention to a physical deformity or indeed may be obscene. It seems at least probable that the surname of England's greatest poet, like Wagstaff, is of bawdy origin.

Many surnames derived from the names of animals and birds were originally nicknames, referring to appearance or character, from the attributes traditionally assigned to animals. In the Middle Ages anthropomorphic ideas were held about the characters of other living creatures, based more or less closely on their observed habits, and these associations were reflected and reinforced by large bodies of folk tales featuring animals behaving as humans. The nickname Fox (Foss, Fuchs, Goupil, Lis) would thus be given to a cunning person, Lamb to a gentle and inoffensive one. In other cases, however, Lamb might be a metonymic occupational name for a shepherd. In other cases, surnames derived from words denoting animals may be of anecdotal origin (see next paragraph). DAFN normally attempts to explain the likely application of a word as a name, but inevitably there must always be some uncertainty as to the precise application, nowhere more so than in names derived from words denoting animals.

Anecdotal Surnames

Another group of surnames consists of nicknames that arose as the result of some now irrecoverable incident or exploit that involved the bearer. Probable examples include Followell, Tipper, Mezzanotte, Musil. In modern nicknames borne by individuals within a community, this type of name is common, but it is also apparent that the reason for the nickname, which may only ever have been known to a few people, is quickly forgotten, whereas the surname may continue to have enduring currency. It is fruitless to try to guess now at the events that lay behind the acquisition of nicknames such as Death and Leggatt in past centuries.

An unusual class of anecdotal surnames occurs in Czech, with name derived from the past participle of a verb. These are often difficult to gloss in English. Examples include Doležal, a nickname for a lazy man meaning something like ‘laid back’, Doskoˇil, an agile man, literally ‘leapt about’, Kratochvíl ‘had a good time’, Kasal ‘bullied’, and Kvapil ‘rushed’.

Seasonal Surnames

Related to such “incident” names are names that refer to a season (e.g. Winter, Lenz), month (e.g. May, Davout), or day of the week (e.g. Freitag). It has been suggested that these names refer to the time of birth, baptism, or conversion. In the cases of more recently acquired surnames, in particular Jewish names, reference is sometimes to the time of official registration of the name. Certainly surnames derived from the names of various Christian festivals (e.g. Christmas, Toussaint, Ognisant) seem to have been acquired in this way. But the seasonal names may also have referred to a ‘frosty’ or ‘sunny’ character, while the medieval day names may have referred to feudal service owed on a particular day of the week. No explanation offered for either Christian or Jewish names in this group has been proven conclusively.

Humanistic Names

A small group of names, mostly of Dutch and German origin, are here dubbed “humanistic names.” These are Latin forms of vernacular originals, coined during the heyday of Renaissance humanism in the 15th and 16th centuries. In some cases, the alteration consisted of nothing more than adding the Latin ‐(i)usnoun ending to an existing name, as in Bogardus, based on Dutch Bogard‘orchard’, or Goetschius, from German Goetsch, a pet form of Gottfried.In other cases, the whole surname was translated into Latin: Agricola is a translation of Dutch Boerand German Bauer‘farmer’; Faber is a translation of German Schmidtand Dutch Smit‘smith’; Silvius ‘of the woods’ represents vernacular names such as Dutch Van den Bosch and German Forster.

The pattern of forming surnames from Latinate humanistic elements was copied in Sweden in the 18th and 19th centuries, when surnames became generally established there.

Ornamental and Arbitrary Coinages

A category of surname not found in most European language and apparently confined to communities where the adoption of surnames was late and enforced rather than organic, is the ornamental or arbitrary coinage. For further remarks on this class, see the sections on Swedish, Finnish, and Jewish surnames in the area‐by‐area survey below.

Variants, Diminutives, Augmentatives, and Pejoratives

Certain classes of surnames are derived from base forms of personal names and nicknames, and occasionally from vocabulary words too.

Diminutives include surnames formed from vocabulary words with an affectionate suffix (e.g. Czech Bajorek ‘little marsh dweller’, Italian Abello ‘little bee’, Castello ‘little castle’), as well as the much more widespread types derived from pet forms of personal names and nicknames (e.g. Polish Bolek, a diminutive of Bolesław, English Jess and Jessel from Joseph, Russell from Old French rouse‘red‐head’). In practice, it is not always possible to differentiate between a diminutive and a simple variant. More often than not, however, diminutives are distinguished by specifically diminutive suffixes, of which Italian has a particularly rich and productive set. Polish and Czech are not far behind, while diminutives of one kind and another are also found in most other European languages. They are rarer in Spanish, however, which does not boast the wide variety of derivative forms in its surnames that are found in other European languages.

Diminutives and variants are common in all languages. Augmentatives and pejoratives are much rarer. Whereas diminutives mean ‘little’ and are often affectionate, augmentatives mean ‘big’. A typical augmentative ending is Italian ‐one, as in Italian Iacovone, ‘big Jim’, or Colone, ‘big Nick’. Finally, mention must be made of pejoratives, where an ending that originally had an insulting or derogatory force has been added. A typical pejorative ending is French ‐ard, as in Bechard, a nickname for a gossip, from a pejorative derivative of bec‘beak’.

Surnames, Genealogy, and Genetics

It has been emphasized that DAFN is a book about the history of names, not the history of families. To some extent, of course, the two go hand in hand: the origins of the surname Aaberg are, unsurprisingly, associated with the origins of the Aaberg family. In some cases, it is possible to identify the first person (or one of the first) who bore a particular surname in North America. For example, modern bearers of the surname Zollicoffer are descended from Jacob Christoph Zollicoffer, who came to the U.S. from Switzerland in the early 18th century and helped to found the settlement of New Bern, North Carolina.

In some cases, the form of a European surname has been altered to a uniquely American form. Thus, bearers of the American surname Styron are almost certainly all descended from a certain George Styring, who came from Yorkshire, England, to North Carolina in about 1720. Americanization is a process particularly characteristic of French, Dutch, and German surnames during the 18th century, but the process continues to this day. Names of late 19th‐ and early 20th‐century immigrants tend to keep the form that they had in the language of origin, but even so there are exceptions. The comparatively rare Polish name Bładaśewicz, for example, was simplified to Bladey as an American name in the mid 20th century. On the other hand, at least 199 telephone subscribers have opted not to simplify the distinctively Polish name Krzywicki, even though that spelling can be baffling to English speakers unfamiliar with the Polish language. In an American context, the name is often given an American pronunciation (“Krizwikky”) rather than a Polish one (“Krzheeveetsky”). This special relationship between pronunciation and spelling is characteristic of Polish American names.

The Italian name Abate is now sometimes heard in America as two syllables (rhyming with “fate”), rather than as three, as in the original Italian. The surname of Madonna Louise Veronica Ciccone similarly varies between two syllables with an initial S‐(the American way) and three syllables with an initial Tsh‐(the Italian way), depending on what sort of folk hero the speaker wants her to be.

The words of the English language do not normally carry accents, and this has posed particular problems for immigrants whose family name contained an accent in the language of origin. The Czech surname Kuˇera, for example, is found in two spellings in America. In the one case, Kucera, the accent has simply been dropped; in this case the pronunciation is found both as “Kootchera,” resembling the Czech original, and as “Koosera,” under the influence of the accent‐less American spelling. In the other case, Kuchera, the spelling has been Americanized to reflect the pronunciation.

In DAFN, surnames as headwords are not accented. Instead, the accented form in the language of origin is given as part of the explanation.

Names with a comparatively low frequency are typically derived from a single ancestor, though it is a matter of chance whether sufficient records still exist to enable lineage back to that ancestor to be traced. On the other hand, very common names such as Brown, Johnson, Jones, Miller, Smith, and Williams are not only extremely frequent in Britain, the country of their linguistic origin, but also have typically absorbed cognates and translation equivalents from many other languages and indeed have been adopted by people with no previous connection at all with the name, for example on grounds of euphony or simplicity.

If your name is one of these very common names, then tracing your ancestry is virtually impossible, unless reliable family records have been kept by family members in previous centuries (for example in the front or back pages of a family Bible).

Obviously, it would not be practical for a work of general reference such as DAFN to identify and trace the history and descendants of each immigrant. Ideally, every single name in this dictionary merits a whole book to itself. In DAFN, therefore, only the most salient points are mentioned. In particular, the early immigrants singled out for mention in DAFN are cited as evidence for the linguistic and cultural origins of a name or for changes in its form. The fact that a particular 17th‐ or 18thcentury immigrant mentioned in DAFN does not necessarily imply that all modern American bearers of a given surname are descended from him.

The tremendous advances in the study of genetics in recent years is beginning to have an impact on family history. The work of the geneticist Bryan Sykes, for example, has shown how DNA can be used to establish whether all (or most) bearers of a particular surname are in fact related. In years to come, no doubt DNA evidence will help to establish systematically, on a basis of probabilities, whether two differently spelled names are variants of the same original name, or whether they are in fact separate and independent. In the next section, some examples are given, drawing on surnames of English origin. There can be little doubt that the history, geographical distribution, and genetics of surnames in most other parts of Europe follow similar patterns. The potential of this work is therefore of relevance to a large proportion of American family names.

Monogenetic and Polygenetic

A distinction is made between monogenetic surnames and polygenetic surnames. Monogenetic surnames are those which are derived from just one original bearer at one particular place and time, whereas polygenetic surnames were coined independently in many different places. Smith, Brown, Milton, and Newton are good examples of English polygenetic surnames, and one would expect this to be confirmed by DNA studies. On the other hand, Sykes, Hanks, and Hardcastle are examples of names that are most probably monogenetic: they originated at one place and at one time, i.e. they come from one single ancestor.

It is not normally possible to identify the actual original bearer of a monogenetic surname, for all records of him— rarely her—may have vanished (if they ever existed). However, it is often possible to establish that a name is most probably monogenetic on the basis of its distribution. As long ago as 1890, H. P. Guppy noticed, on the basis a study of Kelly's directories, that many family names are significantly associated with a particular place. On the basis of computational study of the current geographical location of 15,000 British family names, Hanks (1992) concluded that many surnames in Britain are monogenetic, while very many others have “become monogenetic”— i.e. the name may have been coined in more than one place and in more than one time in the Middle Ages, but subsequently the family lines descended from all except one of the original medieval bearers have died out. Meanwhile, the descendants of just one medieval bearer have prospered and multiplied, resulting in the modern distribution of the name.

This was empirical confirmation of a prediction made by statisticians (e.g. Sturges and Hackett 1987; see also Feller 1957), which (broadly) says that frequent names tend to become more frequent, while infrequent names tend to become less frequent and, eventually, die out. Thus Asquith and Auty, the one a habitational name and the other from a Norse personal name, are both so strongly identified with West Yorkshire that the chances of their being monogenetic must be very high. Possibly, DNA research will establish whether present‐day bearers of the surname Asquith are all descended from a single individual, or whether there are several lines stretching back to different individuals all from the village of Askwith. The former is more likely than the latter, as is the case with the surname Sykes. Even though this name is now widely dispersed in the modern world, and even though there are several places in northern England called Sykes, any one of which might be the source of the surname, the survey of contemporary British surnames described in Hanks 1992 shows a statistically significant association with West Yorkshire. Sykes and Irven (2000) argue that, since male children have consistently inherited both their surname and the Y‐chromosome from their father, DNA evidence can be used to confirm or disconfirm the monogenetic hypothesis for any given surname. Combining DNA evidence with geographical distribution and evidence from local history, they show that the majority of present‐day bearers of the name are not only related but also can trace their origins back with confidence to a family living in Slaithwaite near Huddersfield in Yorkshire, England, in the 15th century. The name can probably be traced further back, to the 13th century, when William del Sykes held land in Flockton, nine miles to the east of Huddersfield (Redmonds, 1992).

The bringing together in this way of historical and linguistic evidence, statistical studies of geographical and demographic distribution (both modern and historical), and genetic research is in its infancy, but it promises exciting developments in years to come. The examples cited are from England, but there is every reason to hope that these principles can be applied successfully to American family names and indeed on a worldwide basis, not only for monogenetic surnames, but also for different strains of polygenetic surnames.

Genealogical Resources

In 1999–2000, the world of family names research was transformed in a different way, by developments on the Internet. Probably the most important of these was the release of the International Genealogical Index (IGI) as a freely available on‐line resource via At the same time, genealogical web sites and in particular discussion forums began to flourish and 19th‐century censuses were published in electronic form. With resources like these, research that had previously taken days or was literally impossible can now be accomplished in hours.

The IGI is a vast record of births and marriages (not deaths) from civil and church registrations, and other sources from past centuries, transcribed and compiled by volunteer members of the Church of Latter‐day Saints (Mormons). Of course, no conclusions can be drawn from frequency (or lack of it) in IGI, for IGI is not a systematic survey such as a census. Also, IGI has a bias in favor of English, German, and Scandinavian records, and in particular it has a policy of avoiding Jewish records, following an agreement between Jewish and Mormon religious leaders.

The entries in IGI have never been subjected systematically to the scrutiny of modern scholarship (indeed, to do this on such a vast scale is almost unimaginable), and there are undeniable blemishes (it has often been pointed out that some entries are clearly errors; a few are merely fanciful; there are variations in details, so that a single individual may occur several times with slight differences in date of birth or spelling of a place name; other entries betray an ignorance by the transcriber of such matters as the geography of language and political boundaries in 18th‐ and 19th‐century Europe). Researchers using IGI should particularly bear in mind the “failure to find” fallacy. For example, failure to find IGI records for a particular surname in Spain or Turkey does not mean that the name is not Spanish or Turkish; it may only mean that the relevant records have not yet been transcribed.

But such blemishes are unimportant to a researcher seeking a broad overview as a starting point for more detailed research. The greatest benefit of IGI to family‐name researchers is that, in its electronic form, it gives an immediate overview, showing immediately when and where a particular surname was most frequent in years gone by. It contains hundreds of millions of entries (it is growing all the time), the vast majority of them accurate and reliable.

Other genealogical resources must be treated with equal caution. Studies of particular surnames from the 19th and 20th centuries, often printed privately, are of very variable quality. Some contain priceless information, meticulously researched, with detailed citation of sources and evidence, informed by an understanding of the linguistic and historical processes that conditioned the development of family names. Others are little more than a garbled mishmash of half truths and folk beliefs.

The same wide variability can be observed in contributions to on‐line forums and discussion groups. Some contributions are well‐informed and reliable; others are of very doubtful reliability. One particularly common characteristic, both of older studies of individual surnames and of modern contributions to genealogical forums, is that researchers of American family names often seem to have a very hazy grasp of the facts of life and language in feudal medieval Europe and the processes of surname formation. It is hoped that DAFN will go at least some way toward remedying this defect.

Despite the undeniable benefits of on‐line resources, sooner or later the serious genealogist or family historian will feel a need to visit the local library and historical record offices, both in the places where the surname first became established in North America, and in the locations in Europe or Asia from which it was first brought to America. For the serious one‐name researcher, there simply is no substitute for getting to grips with original documents.

Rarer Surnames

If you are among the 15% of Americans for whose name there is no entry in DAFN, you can be sure that your surname is a rare one—specifically, you can be sure that there are fewer than 100 listed telephone subscribers. This is because every surname with a frequency greater than 100 in 88.7 million (the size of Dr. Tucker's database, based on telephone directory sampling) gets an entry. Even if your name is not an entry, there are good chances that it may be a spelling variant of another name that is, for of course the spelling of surnames is notoriously unstable. Therefore, it is worth searching DAFN for names similarly spelled. Vowels, in particular, are unstable. In a few cases (for example, Wege, Waggy, and Wagy—see the entry for Wege), it has been possible to pinpoint precisely when and where certain variants originated; more often, however, precise details about the origin of variants are unknown. In DAFN only the more common variants are recorded: rarer variants, with a frequency below 100 in 88.7 million, do not receive an entry unless they are of particular historic or etymological importance.

So, for example, if your name is Goodykoontz, a surname with a frequency of only 56 in 88.7 million (0.63 per million), you will not find your name in DAFN. But with perseverance and a little imagination, you may find your way to the DAFN entry for Gutekunst and discover that the latter is a name of German origin, denoting a particularly skilled workman.You will then want to consider the possibility that Goodykoontzmay be an American variant of Gutekunst. If the two names are both found in a similar location at an early date (i.e. in 17th or 18th century America), the likelihood that the former is a variant of the latter is greatly increased. Now, you may decide to consult other sources, starting perhaps with the 1880 U.S. census, which is now available in electronic form, or the International Genealogical Index (IGI) at (see the section on “Genealogical Resources” above). How frequent was your name in 1880? Well, Goodykoontz has 81 occurrences in the 1880 census, when the U.S. population was approximately 50 million (1.62 per million). So the name today seems to be less half as frequent as it was 120 years ago.

If by this time you have got the bit between your teeth, you may also be inspired to check out genealogical discussion forums and websites (“Cyndi's list” of genealogical sites on the Internet is a good guide:, to see whether any research on the history of your family has already been done. You may also decide to ransack second‐hand bookstores, or consult the Library of Congress in Washington or the Family History Library in Salt Lake City, in search of a family history. (As it happens, a 16‐page booklet on the Goodykoontz family was published in 1914.) You will notice that, although your surname was already widespread in many different states in the 1881 census and is even more diffuse today, 18th‐century records of it are centered on Floyd, Virginia, and Frederick, Virginia.

In other cases, a rare family name may be unrelated to any existing DAFN entry. For family historians and genealogists, rare family names are generally easier to research than frequent ones. If your name is Smith or Jones, it is going to be hard, if not impossible, to distinguish your particular ancestors from literally millions of other Smiths and Joneses who have lived in the past. But if your name is Aspiras, Babigian, Caraccia, or Gopinath (some of the other surnames with a frequency of only 56 in 88.7 million), it may well be that you are related to all other bearers of that name, and you may decide to start your family history research on this basis.

If your name is not a variant of an existing entry in DAFN, your first step will probably be to establish the country of origin. Here again, comparative frequency is important. A surname that is rare in the U.S. may be very common in the country of origin. A case in point is Gopinath, which is not found in the IGI but, it turns out, is frequent in Kerala, India. So in this case, the hypothesis that you are related to all recorded bearers of the name begins to look much less plausible.

Because people are so mobile (a bearer of an Irish Gaelic name may have reached America via Australia; a Chinese via Singapore or Malaysia; a Scot via Canada or New Zealand, and so on and so forth), serious research into American family names must inevitably be conducted on a worldwide basis. Thanks to the Internet, this is now possible.


Many individuals and organizations have supported DAFN in various ways during its long gestation period. In 1991 Donnelley Marketing kindly gave permission for use of data on surnames from their files to get the project started, and Kenneth W. Church, of AT&T Bell Laboratories, provided this data in the form of a headword list. James Howes, head of Reference Computing at Oxford University Press in Oxford in the early 1990s, established the text structure, created routines for data conversion, and gave valuable computing support. Kathleen Much helped with the start‐up of the editing process. Marion Harris pioneered work on the genealogical notes and helped to recruit and organize the team of consultants and contributors. Flavia Hodges gave support and encouragement, including permission to make use of data from the Dictionary of Surnames(1988), coauthored with Patrick Hanks.

In addition to the official project contributors, scholars and genealogists too numerous to mention individually have kindly answered questions about particular names and generously volunteered valuable information. Pirjo Mikkonen vetted the contribution for Finnish names; Xulio Viejo Fernández was consulted with reference to all Asturian names. Małgorzata Calhoun contributed to the editing of Slavic names. Marie Lukášová proofread all the Czech entries, and Vladimir Benko gave valuable computational support in processing Czech and Slovak entries. Anthony Wood provided a systematic identifi‐ cation of Muslim names. Michael Morrell helped to edit the introductory essays.

Thanks are also due to the publishing team at Oxford University Press, both in New York and England, over a twelveyear period. Linda Halvorson helped shape the original project design. Claude Conyers oversaw the long, slow, and (no doubt) frustrating period of project development from 1991 until his retirement in 1998; his support, in keeping the project on course and encouraging the maintenance of editorial standards of coverage and accuracy, overcame many obstacles. Without the support of Edward W. Barry, President of Oxford Univesity Press USA until 1998, the project would never have been possible at all. Tim Benbow, Director of the Dictionaries Department in Oxford until 1997, gave valuable advice and moral support. Frank Abate, formerly head of U.S. dictionaries at Oxford University Press USA, gave valuable advice and encouragement, as did his successors Elizabeth Jewell and Christine Lindberg.

Finally, special mention must be made of Timothy J. DeWerff, who, following his appointment as Director of Editorial Development at Oxford University Press in New York in mid‐2002, galvanized the finalization process. His exceptional energy, tact, wisdom, and enthusiasm made it possible to complete the final stages of this huge, complex project in the incredibly short time of six months.