Period-Accurate Character Names: How to Pick Names That Fit the Era

To give a character a period-accurate name, work out the year they were born (not the year the story is set) and look up which names parents actually chose then. For the US, the Social Security Administration publishes the top names for every decade since the 1880s; for England and Wales, the Office for National Statistics publishes top 100 lists going back to 1904. For older periods, get the naming system right first (titles, patronymics, surnames or none) and then the names.

Our guide to naming characters covers the general tests a good name has to pass. This post is the data behind the era test: what the records show, decade by decade, and where period novels most often slip.

Name for the birth year, not the story year

The most common error in historical and family-saga fiction is naming every character from the story's own decade. A novel set in 1955 has teenagers, parents and grandparents, born decades apart, and each generation's names came from a different fashion.

Take a family in a 1955 American novel. The 16-year-old daughter was born in 1939; her mother, at 33, was born around 1922; her grandmother, at 68, around 1887. That's three different naming worlds. In the SSA data, the grandmother belongs to the decade of Mary, Anna and Emma; the mother to the era of Mary, Dorothy and Helen; and Linda, the second most popular girl's name of the 1950s, is right only for a baby born that decade, not for anyone old enough to have lines in the story.

The arithmetic is trivial, but it's easy to skip. Do it for every named character before you choose the name.

What the US records show, decade by decade

The SSA's decade pages list the 200 most popular names for boys and girls, drawn from Social Security card applications. The 1920s page, for example, draws from a universe of 11,372,758 boy births and 12,402,259 girl births.

Three patterns stand out.

Boys' names change slowly; girls' names change fast. John, James and William appear in the top five for the 1880s and William is still there in the 2010s. The girls' lists turn over almost completely: Dorothy, Betty and Helen in the 1920s, Linda, Patricia and Deborah in the 1950s, Jessica, Jennifer and Ashley in the 1980s. A girl's name dates a character more precisely than a boy's, which makes it more useful and more dangerous.

Mary dominated for decades. Mary is the top girl's name in the 1880s, 1920s and 1950s tables. In the 1980s it had dropped to 35th, and in the 2010s to 127th. A cast of older American women with no Mary in it is, statistically, a little odd.

Names recycle. Emma was third in the 1880s and first in the 2010s. Olivia was 195th in the 1980s and second in the 2010s. Noah was 155th in the 1880s and first in the 2010s. A writer using an 1880s Emma is accurate, but a modern reader may hear a 2010s name. We come back to that problem below.

Older casts should share names

The other thing the records show is concentration. Divide the count for the top name by SSA's stated total for each decade, and the share given the single most popular name falls steadily.

In the 1880s data, John accounted for 89,949 of 1,177,152 boy births, about 7.6%. In the 2010s, Noah accounted for 183,378 of 20,110,846, about 0.9%. Girls follow the same path: Mary at about 6.5% in the 1880s, Emma at about 1.0% in the 2010s.

For fiction, this means a 19th-century village with three Johns and two Marys is realistic, and authors can use that (nicknames, middle names, "Young John") as texture rather than avoiding it. A contemporary cast where two teenagers share a first name is less likely, and readers will notice.

One caveat matters for the early decades. SSA's background page notes that "many people born before 1937 never applied for a Social Security card, so their names are not included in our data." The pre-1937 tables are a large sample, not a full count.

England and Wales: the ONS lists

The ONS publishes a historical top 100 for England and Wales at 10-year intervals from 1904 to 2024, compiled from first names recorded at birth registration.

The British lists diverge from the American ones in ways that matter for a British-set book. Florence, Doris and Edith sit just under Mary at the top of the 1904 girls' list; none of them appears in the US top fives above. In 1954, the British girls' top five (Susan, Linda, Christine, Margaret, Janet) overlaps the American only on Linda and Susan. And 2024's list shows recycling again: Arthur, fifth in 1904, is fourth in 2024, and Florence, second in 1904, is sixth.

If your characters are British, don't borrow American data, and vice versa. Christine and Janet, both in the 1954 English top five, never reach the US top five in the decades above.

Victorian England: registration samples

National ranked lists for England begin in 1904, so for the Victorian period you need studies built on registration or census records. The Guild of One-Name Studies publishes men's and women's forename tables built from samples of FreeBMD birth registrations at 15-year intervals, plus samples of bridegrooms' and brides' names from 1867 to 1886.

What they show is useful for any book set between 1837 and 1901:

• Mary is first for girls in every sample year from 1838 to 1898. Elizabeth, Sarah, Ann and Jane fill out the early lists.
• The late-Victorian names arrive late. Florence, Edith and Ethel are absent from the 1838, 1853 and 1868 top 10s, but Florence is fourth in 1883 and second in 1898. A Florence or an Ethel born in the 1840s is possible but unusual; a late-1880s baby with either name is ordinary.
• For boys, William and John lead throughout, with George, Thomas and James close behind. Arthur and Albert join the top 10 later in the century.
• Names concentrated, then spread. The top 10 boys' names cover 77% of the 1838 sample and 55% of the 1898 sample. For girls the fall is steeper, from 67% to 36%.

The Guild also notes that the top names with the greatest decline since Victorian times "are Mary, Ann, Jane and Agnes," and puts them at "now 1% of the frequency then." That's the reverse of the Emma problem. An authentic early-Victorian name can read to modern ears as plain or old-fashioned, which may be exactly the effect you want.

Our Victorian English name generator is a quick way to get a shortlist in this register; check the ones you like against the tables.

Medieval England and France

For medieval England, the same Guild tables include a 13th-century sample: an analysis of 4,001 men's and 1,407 women's names from the Essex Feet of Fines. William leads at 13%, then John at 10%, Robert at 9% and Richard at 8%. Among women, Alice (11%), Matilda (10%), Joan (6%) and Agnes (6%) lead, with Emma, Isabella, Margery, Christiana, Rohesia and Juliana in the top 10. A small stock of names, shared widely, is again the realistic pattern.

Spelling is where medieval names reveal their period. The Dictionary of Medieval Names from European Sources, edited by Sara L. Uckelman at Durham University, aims to contain all given names recorded in European sources written between 600 and 1600, and its entries cite each spelling with its date and source. Its entry for John shows the range: Middle French "Jehan" in 1351, Middle English "Iohn" in 1387, and diminutives such as "Iankyn" in the early 1400s. A medieval Jehan or a Iankyn signals the period more strongly than "John," and a dated spelling from DMNES is a defensible choice if an editor queries it.

For quick shortlists, try the medieval English and medieval French presets, then confirm spellings in DMNES.

For France after 1900, INSEE publishes a ranking of first names given in France since 1900, which covers the end of the Belle Époque; our Belle Époque French generator gives you a starting list to check against it.

Older and non-English settings: get the system right

Before modern registration, and outside the English-speaking world, there's often no ranked list to consult. What you can get right is the structure of a name, and readers who know the period notice structure faster than they notice sound.

Heian Japan

Even the most famous Heian woman writer is known to history by a sobriquet. Britannica says the real name of the author of The Tale of Genji "is unknown," that she may have taken "Murasaki" from the heroine of her novel, and that "Shikibu" reflects her father's position at the Bureau of Rites. Sei Shōnagon, Britannica notes, was the daughter of the poet Kiyohara Motosuke. Our suggestion: a Heian lady known in your story by a sobriquet tied to a relative's office follows that record more closely than one called by a modern-sounding given name. The Heian Japanese generator can supply raw material.

Meiji Japan

According to a nippon.com timeline, the general public were allowed to start using surnames in 1870, and mandatory surnames began in 1875 "for tax collection and military registration purposes." A commoner in an 1860s-set story without an official family name is accurate; one in the 1880s probably has one. The Meiji Japanese generator fits the later setting.

Imperial Russia

In Russia, Britannica explains, everyone has a patronymic as well as a given name and surname: sons take a form ending in -ovich (Ivanovich) and daughters one ending in -ovna (Ivanovna). Get the patronymic right, and match it to the father's name throughout the book; it's an easy continuity error. Try the Imperial Russian generator for combinations.

Ancient Rome

The classical Roman "tria nomina" was praenomen, nomen gentilicium and cognomen. Benet Salway's survey in the Journal of Roman Studies treats it as a transitory stage rather than a culmination, typical of the best-documented class in the late Republic and early Empire, and notes that the cognomen later displaced the nomen as the most consistently attested element. A footnote cites Probus reporting that Varro thought there were only about thirty praenomina. So Roman men in your story should share a very small pool of first names, and the name they're called by may shift with the century. The Old Roman generator is a starting point.

Anachronism: the traps the records expose

The first and last traps are the most common. Jennifer is a good example of a late boom: absent from the SSA top 200 for the 1880s, 98th in the 1950s, second in the 1980s. Ashley, Brittany and Madison follow similar curves, appearing in the top lists of the 1980s or 2010s but not in the 1950s top 200. Any of these on a character born before the 1950s needs a reason in the story.

The recycling trap is subtler, because the name is accurate. When a researched name reads as modern (Emma, Olivia, Arthur) or as cartoonishly old (Agnes, Mildred), you have a choice between historical accuracy and the reader's ear. Our naming guide discusses that trade-off; the data here tells you which side of it you're on.

Spelling is the quiet one. SSA's background notes say different spellings of similar names are not combined: Caitlin, Caitlyn, Kaitlin, Kaitlyn, Kaitlynn, Katelyn and Katelynn each get their own rank, and ONS works the same way. A sound can be period-right while a spelling is decades out.

Putting it together

The top of a list describes the average baby. Real families deviate: an immigrant household, a devout one, a family that names every eldest son after his grandfather, parents who read novels. Use the records to know what was normal, then decide on purpose where each character departs from it. If a name's meaning matters to you as well as its period, our data study of character name meanings covers that layer.

Once names are chosen, write them down with birth years, nicknames and exact spellings, especially in a multi-generation saga where the same name passes down a family. That record is what stops "Kathryn" from turning into "Katherine" in chapter 30.