How Migration and Language Changes Shape Surnames

Every surname is a fossil of two journeys — the journey its bearers made through space, and the journey the name made through sound. Here is how to read both, and how they are braided together in your own family’s records

The simplest way to misread surname history is to imagine names standing still. They never did. A surname froze at one moment in one locality — and then travelled: with apprentices to towns, with clerks and knights in retinues, with wool and coal and cotton, with emigrant ships, and (a beat later) with the paperwork of new countries. Meanwhile the name itself kept moving — through dialect boundaries and sound changes, through bilingual clerks and alien alphabets, through translation and folk etymology, until universal literacy finally pinned most spellings to the page. Migration and language change are the twin engines of every surname’s biography, and this guide — the eighth in our series — draws them together, mining the same traceable scholarship used in our earlier pillars (Welsh Surnames; Spelling Variations; Occupational Surnames; Patronymic and Matronymic Surnames; Locational and Topographic Surnames; Frequency and Distribution; Surname History vs. Family History; US Census Data).

The standing editorial rule, as ever: the surname’s journey (origin, spread, transformations) is evidence about a name; your family’s journey must be proven document by document. Migration logic and language-change logic tell you where to look and what spellings to expect — never what you are descended from.

Key takeaways

• Movement created the name-stock. In late medieval towns, “up to half or more” of persons listed bore surnames from other towns or villages (McClure 1979) — place-name surnames are, first of all, migration records.
• Names travelled with households and offices — Wrotham → Rootham across archiepiscopal retinues; English midland place-names → Scotland under David I (Laidlaw from Ludlow).
• Distance garbles; dialect reroutes. The further a family moved from origin, the more speech won over etymology (Greenwich → Greenidge; Leicester → Lester; Bromwich → Bromage; Stopford = Stockport), and dialect words fossilised into rival surnames for one trade (Fuller/SE, Tucker/SW, Walker/Midlands-and-North).
• Sound change fossilised too. Middle English West-Midland rounding of man → mon, and -us for -es, survive in name distributions; modern phone-book maps of Pett/Putt still trace medieval phonology (Viereck 2009).
• Renaming under migration follows five known routes (Woulfe, 1923): phonetic spelling, translation (Mac an tSaoir → Carpenter; Mac Seáin → Johnson!; Ó Bruic → Badger), attraction to commoner local names, assimilation to “respectable” names, and outright substitution.
• Prefixes were politics as well as grammar. Irish O’ and Mac were dropped wholesale in the eighteenth century and readopted under the Gaelic League — with measurable speed (Grenham’s civil-registration figures; MacLysaght’s Irish Families, 1957).
• Surnames are economic seismographs, too: in the Cotswolds, Shepherd collapsed from 3.35% of recorded names in 1381 to 0.10% by c1600 as the economy shifted from raw wool to cloth — while Webb‘s local index rose (Parkin 2014).
• Read your surname’s layers from the newest backwards, and prove every hop with a document. The chain is testable end-to-end: Heinrich Focht (Hamburg) → “Emerik Focht” (manifest) → Henry Focht (Declaration) → Henry John Ford (petition, 1928).

1. Names as fossils of two journeys

The pre-history of this pillar is the medieval byname — a description, not yet an inheritance. In the century and a half before hereditary surnames hardened, documents listing tenants reveal a society in motion: Peter McClure’s foundational study of migration evidence (Patterns of Migration in the Late Middle Ages: The Evidence of English Place-Name Surnames, Economic History Review, 1979) opens on the crucial observation that “up to half or more of the persons listed” in thirteenth- and early fourteenth-century town documents “bear surnames derived from the names of other towns or villages” — a whole class of names coined because people had moved. A toponymic byname was a migration record with the person still attached to it; only later did it ossify, and — as Emily Steiner’s work on naming stresses — such names functioned through local context, “identifying people not because they belonged to those individuals, but [because they] happened to describe… those individuals in certain locations.”

From this starting point, surname history reads as a two-axis problem. On the axis of space: who moved, how far, by what routes. On the axis of sound: how speech and scripts reshaped the name at every waypoint. The axes interact — movement through dialect zones multiplies phonetic drift, exactly the pattern behind our spelling pillar’s variant fans. One methodological caution modern research insists on: name spellings may reflect the scribe’s dialect rather than the population’s (the McClure–Kristensson debate of the 1970s on lay-subsidy rolls), so phonological readings of name forms need local-context checks.

2. How migration shaped surnames

(a) Movement coins names — and misleads the mapmaker

Roughly 60% of medieval English settlements shared their name with somewhere else (McClure’s figure, as used in recent work, which counts sixteen Ashbys in England alone); typical migration to town was short — in the majority of cases “within 14 or 20 miles” — with county bynames taking over beyond about the fifty-mile mark, where precise knowledge of villages faded. That ambiguity is why modern surname-migration studies (a 2026 Bilkent thesis using Medieval Coroners’ Rolls, gravity models and the fourteenth-century Gough Map’s road network) can now estimate rather than assume which like-named place a byname meant — and why “the place-name surname evidence gives the crudest information about migration” even as it remains indispensable. The traps are legendary, and traceable to Reaney and McKinley: a surname Wales can come from a village near Sheffield; Holland from places in Essex, Lancashire and Lincolnshire; Spain (de Hispania, 1086) from Epaignes in Normandy and probably Espinay in Brittany; Bedford from a Lancashire village and Warwickshire’s Bidford. The same scholarship also showed what localization survives: in an analysis of the 1975 marriage indexes for England and Wales, Gabriel Lasker and Bernice Kaplan found 272 place-derived surnames still occurring at or near their eponymous districts about 74% more often than chance — migration over seven centuries has not erased the geography (even though only 0.6% married in the eponymous district itself).

(b) Retinues and administration carried names far

Elite movement wrote some of the best-documented surname journeys: the Wrotham → Rootham line, riding with the archbishops of Canterbury (a son Sheriff of Devon by 1198, a Bedfordshire priory link from 1206); and King David I’s Normanisation of Scotland, which carried midland-English place-names north as founding Scottish surnames — Ramsey, Lindsay, Coventry, Barclay, Hamilton, and Laidlaw from Ludlow, Shropshire. A knight was rarely alone; households reproduced their own name-geography.

(c) Trade and industry moved names — and moved them specifically

The wool economy is the classic laboratory. Medieval English cloth-making vocabulary was regionally divided — fuller (south-east), tucker (south-west), walker (Midlands and north) — and those regional words fossilised into regionally distinct surnames, a distribution still legible in 1881 census maps and in later phone-book data. In the Cotswolds, the great fourteenth-century wool region, Harry Parkin’s diachronic study (2014) found Shepherd the single most frequent recorded name in 1381 — 3.35% of all names, with Rutland (2.08%) the next-highest county — a Banwell index of 3.47 against a national spread; and the region’s poll-tax records (covering roughly 60% of the kingdom’s population in 1381) also show immigrant weavers by trade-name — Brabant bearers, and a weaver of Flemish origin in Campden — plus, by the 1580–1620 registers, toponymic surnames from the textile NW (Kendal-cloth links; Gaskell from Gatesgill, Cumberland; Brownrigg from five Cumberland places). Migration followed the loom: -by surnames — the Danelaw element rare in the Cotswolds’place-name stock — arrive in the Cotswold registers along the Fosse Way corridor (one Busby from the North Riding, 178 miles; a Battersby, 163 miles), and Welsh patronymic-frozen names appear more commonly than their English equivalents in the same registers (15 Pritchards against 6 Richardsons, with a live John ap Thomas still bearing an unfrozen Welsh patronymic in the 1608 Gloucestershire muster roll). Names, in short, track the economy and the roads; McKinley’s county comparisons add the class dimension (locative names concentrated among landholders; personal-name surnames — 30% in Suffolk in 1327 but 3% in Lancashire — differing regionally by class structure).

(d) Emigration carried the freeze; renaming followed arrival

The diaspora studies in our frequency pillar show the pattern at national scale: Welsh-origin names at 3.8% of the US population with an Appalachian concentration; Davies respelled Davis across the Atlantic within a generation; the adoption of plain Smith by unrelated Americans “to blend in”; and, under slavery, the adoption of masters’ Welsh and Scottish surnames — migration history written into frequency tables. Each is a reminder that an American count of a name measures inheritance and adoption patterns in the migration process, not a single founding family.

(e) Immigration and the five routes of renaming

Patrick Woulfe’s classic classification (1923) remains the field’s map for how names change at a language border — phonetic rendering (O’Brien for Ó Briain; O’Neill for Ó Néill); translation — where an Irish name was replaced by an English meaning-equivalent: Ó Bruic → Badger, Ó Bruacháin → Banks, Mac an tSaoir (“son of the wright”) → Carpenter or Freeman, Mac Seáin → Johnson, Ó Draighneáin → Thornton, Ó Gaoithín → Wyndham; attraction — rare names drifting onto commoner near-homophones (Ó Braoin → O’Brien; Ó Maoil Sheachlainn → MacLoughlin); assimilation to a more genteel English name (Broderick for Ó Bruadair); and substitution with no link at all (Loftus for Ó Lachtnáin). Most Irish anglicisation happened 1550–1600 in the State papers; and the English forms then drifted under English sound change — O’Brien and O’Neill were first sounded O’Breen and O’Nail. Modern Britain shows the same engines running: FaNUK’s files record the Walloon-Huguenot Lefèvre arriving in Kent and clipping to Feaver; German mining engineers and Ashkenazic Jewish families establishing Jacobi; Jewish Londoners’ Latinized Faber; a Chinese Li/Lee convergence (over 9,000 Li bearers by 2011, nearly half of Lee’s 84,000 bearers of Chinese ancestry today); Patel at 100,000+; Muhammad at 30,000+ in several spellings. And migration-era name changes may be the emigrant’s own work — the US immigration service’s historian lists the stock reasons (fictitious names, a stepfather’s surname, a maiden name, a farm-name custom) and corrects the “Ellis Island did it” legend our second pillar examined — with official documentation of changes only from 1906.

(f) Dropping and readoption: prefixes as linguistic politics

nowhere is the migration–language–identity knot more visible than in the Gaelic prefixes. Edward MacLysaght’s classic account: “the degradation of the Gael… [produced] the wholesale discarding [from the eighteenth century onwards] of the distinctive prefixes O and Mac” — down to Daniel O’Connell’s uncle, the celebrated “Hunting Cap,” signing himself Maurice Connell as late as 1803. John Grenham’s tabulation of Irish civil-registration births turns the readoption into data: M’ / Mac surnames were 10.14% of all births in 1865 and 10.48% in 1913 — barely moving — while O’ rose from 1.67% to 3.2% over the same half-century, and then leapt in the revivalist 1920s (with Munster, where O’s live, reasserting hardest; one name’s arithmetic — Brien vs O’Brien — went from about 50/50 to roughly 95/5 in favour of the prefix by 1922). The generational sweep of one headline surname: under 2% of Sullivans carried the O’ in the mid-nineteenth century; 14% by 1890; 21% in 1914; 60% by 1944; over 80% today. Errors travelled the same roads — the Mac-Gormáin family now O’Gorman, courtesy of the Chevalier Thomas O’Gorman’s pedigree-making in France. For genealogy the lesson is double-edged: in older records expect prefix-less forms (and treat Mc/Mac/M’ as one), in post-1900 records expect resumption — and never treat the prefix’s presence as proof of anything except an era’s language politics.

3. How language change shaped surnames

Sound change, frozen. Middle English regional sound developments survive in surname geography. Parkin’s Cotswold data pushed the West-Midland rounding of /a/ to /o/ before nasals (man → mon) further south than dialect surveys had found it — visible in poll-tax bynames — and documented the odd West-Midland -us for -es. Wolfgang Viereck’s modern study maps the still-fossilized descendants of Old English /y/: Putt in the south-west vs Pett/Petts in the south-east, in phone-book distributions seven centuries after the sound shifted. And the trade-word fossil beds — Fuller/Tucker/Walker, Webb/Webster/Weaver — show vocabulary change as surname geography. George Redmonds’ catalogue of the transformations (vowel shifts, consonant losses, metanalysis, metathesis, suffix confusion, abbreviation) is the working taxonomy for tracing seemingly unrelated names to one origin — proved, where the documents allow, by aliases: Simon Woodhouse alias Wydis (1611) and, as late as 1739, Thomas Dickson alias Dickinson; and the metanalysis chain Wolstenholme → Wolstenholm, Woolsman, Ousman → Houseman, read off matching distributions.

Standardisation, the ratchet. The FaNUK team’s summary is the compact history: many surnames “became garbled in speech, mis-heard and mis-written in the records, and re-interpreted to fit known names or name-patterns,” with precise forms “only… permanently fixed in the twentieth century” — after which some families re-adopted the etymological spelling (Brown/Browne, Wooster/Worcester). Distance drives the garbling: Greenwich→Greenidge, Leicester→Lester, Bromwich→Bromage; Stopford survives as an old spelling of Stockport.

Bilingual transcription. Welsh clerking produced the ap-fusions of our first pillar (and Shone for Jones in the north-east); US census-era data-processing turned PEÑA into PENA, deleted O’HARA‘s apostrophe, and (by policy) merged SMITH JR into SMITH; transliteration between scripts — Cyrillic, Hebrew, Chinese — stacked variant upon variant down the emigrant’s paper trail.

Translation and folk etymology — the second-order changes. Names can change meaning rather than spelling. The Welsh Ddu → Dee, Caradog → Craddock; the Irish translation-table above (with its famous mistranslations: Woods for Ó Caoilte/Kielty, whose sounds overlapped coilte; King for names that merely lived by Lough Ree — rí, “king,” the English clerks’ false etymology); and the Cavan story — Mac Gabhann (“son of the smith”) translated wholesale to Smith in one county while neighbouring families simply respelled it (Mc)Gowan. In America, translation flows the other way and multiplies: Müller arrivals became Millers — which is why Miller is far commoner in the US than in Britain. Folk etymology completes the process, re-binding opaque names to familiar words and patterns — the same force the FaNUK editors must undo, name by name, in their corrections.

Death. Language change also ends names. Sturges and Haggett’s model, run on 1,000 surnames from 1350 (each borne by about 2.2 men), leaves 430 surviving 23 generations — and FaNUK’s extinct-name lists (Jubber, “maker of liquor-flasks”; Upholder, “second-hand dealer”; the medieval Dryver, Holdare, Tasker, once common in the Cotswolds) show whole trades’ worth of names departing with their economies.

4. Layers, and the reading order

A surname arriving in the records of your search is a stack: origin (a place, a trade, a father) → early byname spellings → a migration that may transform the form en route (Lightowler → Lightoller/Lightholder) → a standard-spelling freeze at some date → possibly a travel respelling (Davies → Davis; Woulfe’s five routes) → possibly a second-language overlay (translation, attraction, substitution) → possibly a modern re-assertion (Browne; O’Sullivan; Welsh ap revivals). The practical rule of this pillar: read the newest layer first and work backwards, because each document only ever shows the layer nearest its own date — and because, as the immigration-service historian notes, “the more direct the route, the less likely the change”: every transfer station left another fingerprint.

5. Three case studies

1. Lightowler → Lightoller / Lightholder (one name, three spellings, one documented route). Redmonds traced the family from Rochdale parish in 1246 into Yorkshire and the Colne Valley, surfacing in Bradford records as Lightoller and Lightholder — migration and phonological reshaping compressed into one provable chain. It is the model for reading an ancestral surname’s journey: find the earliest form, follow the places, and collect the alias/variant bridges.

2. Heinrich Focht → Henry John Ford (an American layer-stack, document by document). JewishGen’s reconstruction of one 1922 emigrant shows the full modern machinery: the child is Heinrich Focht on the Hamburg out-passenger list and school records; the ship manifest and his medical inspection card read “Emerik Focht” (a copyist slip that persisted, uncorrected, through Ellis Island’s registry); his 1924 Declaration of Intention reads “I, Henry Focht” signed Heinrich Focht; and his 1928 petition records an approved legal change to Henry John Ford — which is why no Ford-family descendant finds him under Focht. Every stage is independently citable — and none of it happened at Ellis Island, which merely checked the manifest already written abroad (see our spelling pillar for the full myth-correction).

3. The Cotswold seismograph — Shepherd‘s collapse (Parkin, 2014). In the 1381 Gloucestershire poll-tax returns for the Cotswolds, Shepherd and variants are the commonest recorded name of all: 169 bearers, 3.35% (Rutland, next nationally, 2.08%; the Cotswolds’ index value 3.47). By the 1580–1620 parish registers, Shepherd has fallen to 0.10% of the region’s names — an index of about 0.76 — while the Webb line’s index rises from 1.96 to 5.02, and the Fuller/Tucker/Walker vocabulary continues its regional division of labour. The explanation is economic: as Cotswold prosperity moved from exporting raw wool (the shepherd’s stage) to manufacturing cloth (the weaver’s and fuller’s stage), the frequency landscape of surnames moved with it — a regional surname history written by the market. For family historians the moral is powerful: a surname’s frequency is not a constant; it is an economic time-series — and your line’s apparent rarity in one century is not evidence of anything about the next.

6. Reading your surname’s journey — a working sequence

1. Fix the origin type and epicentre (our locational, occupational and patronymic pillars; FaNUK; Surname Atlas; gbnames).
2. Build the variant log with distance in mind. Expect phonetic drift away from origin (the Bromage rule); treat each spelling as a dated layer.
3. Identify the migration hops. For trade surnames: apprenticeship indentures, city freemen rolls, livery admissions. For locative names: retinues and service (the Wrotham model); for industrial regions: the economic flows of the period (wool, coal, cotton, slate).
4. Cross the water in both directions. Woulfe’s five routes and the USCIS/manifests material for immigration; the diaspora index-maps and passenger-list methods for emigration; remember that both ends of the journey have their own paperwork.
5. Look for the language layers in order: phonetic drift → alias bridges (the Dickson/Dickinson class) → standard-spelling freeze (date it: civil registration, census, school) → translation or attraction or substitution → modern re-assertion.
6. Test with documents and DNA. The cluster method (spelling pillar) through the missing hops; Y-DNA and one-name studies across the branches (capstone pillar); Sykes-style studies where the origin question is monogenetic vs polygenetic.
7. Write the bridge note honestly. Surname journey proposed; family path proven; where the evidence ends, say so — the gap is itself a research instruction.

FAQ

My surname differs between records in Ireland/Tyrone and New York — did an official change it? Almost never (see the Ellis Island corrections in our spelling pillar). What you are seeing is a stack of legitimate layers: anglicisation at the source (Woulfe’s routes), prefix dropping or readoption over time, and the emigrant’s own informal change after arrival. Prove each layer with the document that created it.

Why do my surname’s spellings change as the family moves counties? Because spellings followed speech, and speech follows dialect — and because distance from the origin weakens the hold of the etymological spelling (Greenwich→Greenidge). A spelling change that coincides with a place change is often the trail-head to the family’s route.

Can a surname’s geography tell me where my family “originated”? It can generate a genuinely strong hypothesis — localization persists (the 74% excess found in 1975 marriage data) — but the surname’s epicentre describes the name’s founding population, not your line. The bridge across the “gap” requires documents (and, ideally, DNA) per our capstone pillar.

Our name was “translated” — how do I search the original? Work from meaning: assemble the original-language form (Woulfe’s tables are the Irish model; Mac an tSaoir → Carpenter/Freeman), then search sound-alikes and translations together; expect mistranslations (Woods for Ó Caoilte) and remember some translations were chosen, not imposed.

Where does language change stop? It largely stops with mass literacy and fixed records — early twentieth century in Britain, earlier or later elsewhere — with revival movements as the notable exception (O’Sullivan’s 2%→80%; the Welsh ap revivals in our first pillar). Even now, spelling remains negotiable at marriage, naturalisation and by deed poll: the stack keeps growing.

Sources and further reading

• Peter McClure, “Patterns of Migration in the Late Middle Ages: The Evidence of English Place-Name Surnames,” Economic History Review 32/2 (1979) — up-to-half town bynames; ambiguity estimates; distance patterns; “crudest information” verdict; McClure, “Surnames from English Place-Names as Evidence for Mobility,” The Local Historian 13/2 (1978); McClure, “Personal Names as Evidence for Migrants and Migration in Medieval England,” in Migrants in Medieval England c. 500–c. 1500 (OUP, 2020); “The English Surname Detective,” in Names, Texts and Landscapes in the Middle Ages (2022).
• Gabriel W. Lasker & Bernice A. Kaplan, “English Place-Name Surnames Tend to Cluster Near the Place Named,” Names 31/3 (1983) — 1975 marriage-index analysis; 74% excess localization; the Wales/Holland/Spain/Bedford traps (from Reaney and McKinley); Pickering/Puckering (McKinley, telephone directories).
• Batuhan Baltacı, Ambiguous Names Among Nefarious Deeds: Place-Name Surnames and Migration in Coroners’ Rolls, 1276–1348 (Bilkent University M.A. thesis, July 2026) — Coroners’ Rolls as an inclusive migration source; gravity models on the Gough Map’s road data; the nearest-place critique.
• Harry Parkin, Change in the By-Names and Surnames of the Cotswolds, 1381 to c. 1600 (UWE Ph.D. thesis, 2014) — Shepherd 3.35%→0.10%; Webb 1.96→5.02; Banwell ratios; Fuller/Tucker/Walker word-geography; West-Midland /a/→/o/ before nasals and -us/-es; wool-trade migration (Kendal, Gaskell, Brownrigg, -by names, Flemings: Brabant, Dyper/de Ypres, Clutterbuck); Welsh patronymics migrating (John ap Thomas, 1608 muster roll; Pritchard vs Richardson).
• Patrick Woulfe, Irish Names and Surnames (1923) — the five methods of anglicisation, with translation and attraction tables; Tomás Ó Flannghaile, “Irish Names in England” (1896); Edward MacLysaght, Irish Families (1957) — prefix discarding and Maurice Connell (1803); John Grenham, “How the Os and the Macs Came Back” (Irish Roots blog, 2017) — birth-registration percentages (Mc 10.14%→10.48%; O’ 1.67%→3.2%), O’Sullivan’s sweep (via the Irish-genealogy-toolkit compilation), Mac Gormáin/O’Gorman.
• FaNUK: Hanks, Coates & McClure, Oxford Dictionary of Family Names in Britain and Ireland (2016) and the FaNUK methodology paper — garbling and twentieth-century fixation; Brown/Browne; Faber/Fawber; Lefèvre/Feaver; Jacobi; Li/Lee; Patel; Muhammad; “Parish → Paris”; non-transmission of surnames as Y-chromosome counterparts.; George Redmonds, Surnames and Genealogy: A New Approach (1997/2002) — the transformation catalogue and alias evidence; Redmonds, “Migration and the Linguistic Development of Surnames,” Family History 11 (1980) — Lightowler; Redmonds, King & Hey, Surnames, DNA and Family History (2011/2015) — Wolstenholme→Houseman; Tordoff.
• Richard McKinley, A History of British Surnames (1990) — regional and class differences; the -er suffix; David Hey, Family Names and Family History (2000); Colin Rogers, The Surname Detective (1995) — the Banwell ratio; Steven Archer’s Surname Atlas; Wolfgang Viereck, “On the Geography of English Family Names,” Dialectologia 3 (2009) — Pett/Putt phone-book fossils; Kristensson’s dialect surveys and the McClure–Kristensson scribal debate (1973/1976).
• Sturges & Haggett, Inheritance of English Surnames (1987) — 430 of 1,000 surviving 23 generations; USCIS (Marian L. Smith), “American Names/Declaring Independence” and “Immigrant Name Changes”; JewishGen (Joel Weintraub), “The Ellis Island Name Change Myth” — the Focht→Ford chain; the Welsh Government diaspora study (Webber) — index maps; Mariann Slíz, “Occupational Names in the Hungarian Family Name System” (2015) — official name-change regimes.
• Companion pillars in this series: Welsh Surnames; Surname Spelling Variations; Occupational Surnames; Patronymic and Matronymic Surnames; Locational and Topographic Surnames; Surname Frequency and Distribution; Surname History vs. Family History; Common Surnames in the United States.

Individual surname profiles on Select Surname List now carry a “journey” note: the name’s formation, its principal spellings by period and place, its migration fingerprint — and the specific records in which a reader can test each layer for their own line.