Did other countries have the concentration of first names that England had in past centuries?

Question

Anyone who has done genealogical research on England in the 17th through 19th centuries will have noticed it: a parade of Johns, Williams and Thomases, accompanied by Marys, Elizabeths and Annes. There is a huge concentration of first names: Douglas Galbi's study in 2002 of English first names showed that nearly a quarter of men in past centuries were called John, and the top three names accounted for 40–50% of all people. The same pattern applied to women's names, with Mary dominating even though it had been rare in the mediaeval period.

There is actually a reasonable explanation for this pattern: a tendency to name sons after fathers and grandfathers, together with random growth of lineages. If every father named his first son after himself or his own father, lineages of first names will tend to grow randomly, depending on family size. Think about it like this: every (legitimate) son of John Smith will be a Smith, but they are also more likely than average to be a John. Every grandson of that John Smith is also more likely to be a John, even if their father wasn't. If that tendency runs over many generations, together with randomly varying numbers of male and female children, you will get something close to the random growth pattern known as Gibrat's Law, which tends to result in a power law.

This is interesting to genealogists because knowing that certain names were very common means that you can't conclude immediately that two people of the same first (common) name are the same, but you might be on safer ground with unusual first names, which also tend to pass down the generations. (For example, in my in-laws' tree there are six generations of Sibellas, including daughters of sons of a Sibella). But it's also interesting because the macro pattern of a small set of very common names is consistent with a tendency for names to be passed down the generations, which in itself is an interesting observation.

Is this pattern demonstrable outside of England?

Answer 1

Oooh, I like this question. And I have data to contribute!

This data set is from the Israel Genealogy Research Association's "All Israel Database" (http://genealogy.org.il/AID/index.php) which, as of August 2015, covers over 576,000 records that were collected in, or chiefly about, the land of Israel in its Ottoman period, British Mandate period, and modern period. Because of privacy laws, almost none of the records is newer than 1965, i.e. fifty years ago. Most of the records in the database come from the late nineteenth and early 20th centuries, from a variety of libraries and sources. They cover all religions and ethnic groups in the land, but practically speaking, at the moment they are hugely over-represented with mostly-Jewish source material -- or perhaps a more accurate description would be source material that was far more likely to feature the names of Jews, such as the lists of illegal immigration during the Mandate period, i.e. the Atlit Database (מסד נתונים עתלית). More content is going online all the time, including content explicitly from other ethnic and religious groups; one of the newest big databases is for baptisms and burials at Christ Church in Jerusalem, for example.

Finally, some of the original material was written in Hebrew and some in English, so a standardized transliteration system was put together to handle bilingual indexing in a consistent way. For example, the name "רחל" usually used to get transliterated as "Rachel" (English) or "Ruchel" (Yiddish), but they chose to transliterate it using standardized modern spelling as "Rahel". Similarly, Moses gets transliterated as Moshe, Abraham as Avraham, etc.

So, with that long but necessary preamble out of the way, here are the top fifty English language given names in the 570,000+ records of the database so far, either spelled exactly as they were in their original English sources or else transliterated in a standardized way from Hebrew sources. The numbers in parentheses are the actual count of how many times they appear in the database, not a percentage, as a single record can hold multiple names (i.e. parents and child), but the Power Law pattern is pretty clear:

01. Moshe (32,913)
02. Avraham (28,445)
03. Yosef (25,225)
04. Yaakov (23,921)
05. Yitzhak (22,996)
06. David (19,704)
07. Haim (17,884)
08. Shlomo (14,509)
09. Shmuel (14,063)
10. Sara (12,046)
11. Mordekhai (11,795)
12. Tzvi (11,408)
13. Israel (11,176)
14. Aharon (10,240)
15. Yehuda (10,088)
16. Rahel (9,845)
17. Esther (9,738)
18. Meir (9,715)
19. Miriam (8,866)
20. Hana (8,559)
21. Eliahu (8,446)
22. Arie (7,882)
23. Rivka (7,704)
24. Lea (6,984)
25. Eliezer (6,814)
26. Shimon (6,408)
27. Dov (6,200)
28. Zeev (5,760)
29. Shalom (5,720)
30. Haya (5,456)
31. Yehoshua (5,060)
32. Menahem (5,021)
33. Benyamin (4,819)
34. Shushana (4,727)
35. Barukh (4,051)
36. Malka (3,803)
37. Tzipora (3,721)
38. Leib (3,641)
39. Pinhas (3,576)
40. Efraim (3,488)
41. Mikhael (3,400)
42. Devora (3,300)
43. Rafael (3,275)
44. Natan (3,260)
45. Simha (3,251)
46. Reuven (3,160)
47. Yehudit (3,122)
48. Nisim (2,971)
49. Zalman (2,945)
50. Nahum (2,893)

And here's the top fifty Hebrew language given names in the database so far, either spelled exactly as they were in their original Hebrew sources or else transliterated in a standardized way from English sources:

01. (משה (34,212
02. (אברהם (30,718
03. (יוסף (28,334
04. (יעקב (27,020
05. (יצחק (25,530
06. (דוד (20,057
07. (חיים (19,621
08. (שלמה (15,004
09. (שמואל (14,940
10. (שרה (13,744
11. (צבי (13,655
12. (מרדכי (13,567
13. (רחל (12,069
14. (ישראל (11,823
15. (אסתר (10,785
16. (מאיר (10,325
17. (חנה (10,255
18. (יהודה (9,895
19. (מרים (9,475
20. (אליהו (8,879
21. (רבקה (8,583
22. (אריה (8,382
23. (אהרן (8,045
24. (לאה (7,882
25. (אליעזר (7,256
26. (שמעון (6,637
27. (חיה (6,505
28. (זאב (6,296
29. (שלום (5,891
30. (מנחם (5,772
31. (דב (5,620
32. (שושנה (5,522
33. (בנימין (5,388
34. (יהושע (5,158
35. (ברוך (4,958
36. (מיכאל (4,139
37. (מלכה (4,051
38. (פנחס (4,034
39. (צפורה (4,032
40. (דבורה (3,964
41. (שמחה (3,698
42. (נתן (3,652
43. (אפרים (3,616
44. (ציון (3,615
45. (יהודית (3,569
46. (רפאל (3,475
47. (ראובן (3,419
48. (בן (3,180
49. (נחום (3,175
50. (אהרון (3,078

If you can read Hebrew a little, you'll notice that the two lists do not correspond exactly! For example, in English sources, Rahel is #16, while in Hebrew, it's #13. And some of the names start to diverge more and more because of genuine spelling variants, as the lists go on; for example, on the Hebrew list, #23 and #50 are both acceptable ways to spell the name Aaron/Aharon, which is way up at #14 on the English list.

So yes, Power Laws certainly seem to be in effect here too, with certain given names seen far more frequently than others.

However, your comment makes the assumption that certain names came to prominence in England in your initial data set because first-born sons were named for their fathers, daughters for mothers, etc. This data set would break that assumption, because Ashkenazi Jews (from Eastern Europe) do not name for living relatives, but rather for a recently deceased close relative, such as grandparent or great-grandparent. And Sephardic Jews (from Iberia originally, dispersed mainly to the Mediterranean and Caribbean after 1391/1492) have another tradition entirely, where they name a first-born son for a paternal grandfather, first-born daughter for a grandmother, repeat with the maternal side, and only then start adding in new and unique names. Most of the surnames in the IGRA database tend to heavily favor traditionally Ashkenazi surnames (Friedman, Katz, Goldberg, Shapira/Shapiro, etc.) so I would assume their given name naming patterns probably predominate in the data set too.

Also potentially impacting the name re-use: gender! Traditionally, Jewish males needed a Hebrew/Biblical name, so they could be called up to read to the congregation from the Torah (Old Testament), but women were not called up to the Torah to read and lead prayers and so could be named for a Biblical figure or just a common vernacular language name, in whatever country they lived. So you'll notice the relatively few female given names that managed to crack this Top 50 list were Biblical names, but there's a lot more diversity in the womens' names, so the Power Law is weaker, more diffuse. In fact, #68 on the list would be "Frida" and #88 would be "Roza" -- neither name Biblical, obviously, but more just a reflection of common names for women in the early 20th century. You even get hundreds of Charlotte's and Margaret's as the list goes on, too, and eventually it stops looking like what you'd think of as a traditionally Israeli or Jewish name list and just becomes more reflective of the fashionable female names of the time period.

Finally, I'd be remiss if I didn't point out that the All Israel Database is entirely free to use and search, although you do need to sign up and make a user account first. It's quick and free, though. They have an amazing amount of other content online, including research guides and webinars, and IMHO it's definitely worth becoming an annual member to help support their work: http://genealogy.org.il/

Answer 2

Bottom Line: There is some evidence that the tendency for a few given names to dominate is also seen elsewhere in Europe, even if the top names are not John and Mary

Evidence Point 1: Lowland Scotland

Lowland Scotland is not that far removed from England anyway, but I now have evidence that the same concentration exists there too. I have (laboriously) transcribed the Parish Lists of Wigtownshire and Minigaff, 1684, which amounts to more than 9000 names. The same tendency for a few names to dominate applies, but for women the names that dominate are different than in England. Here are the top-twenty male and female names, after regularising spellings. The domination of Janet (here together with its variants Jonet, Jannet and Jennet) is stark. Characteristic Scottish names like Fergus, Ochtrie, Niven, Grizell/Grissel and Elspeth are reasonably popular, and historic Scottish names like James, Alexander and Agnes even more so, but like in England, John and William are in the top two.

John        27.90   Janet       31.60
Alexander   12.40   Margaret    18.60
William     11.00   Agnes       8.81
James       8.27    Jean        6.30
Patrick     7.01    Marion      5.59
Andrew      6.24    Elizabeth   5.04
Thomas      4.46    Isabel      3.49
Robert      3.90    Grizell     3.21
Gilbert     3.00    Helen       3.00
Hugh        2.10    Catherine   2.68
George      1.92    Christian   1.64
Archibald   1.51    Elspeth     1.49
David       1.31    Mary        1.43
Anthony     0.77    Ellen       0.96
Michael     0.70    Sarah       0.85
Fergus      0.59    Bessie      0.85
Adam        0.54    Barbara     0.57
Ochtrie     0.36    Marie       0.53
Niven       0.32    Anna        0.51
Ninian      0.27    Katrine     0.43

I note that the less common but characteristically Scottish names like Katrine, Niven, Ochtrie etc tended to come in runs in the parish lists, meaning that they were more like to be in the same village / family / household. Even if they didn't have the same surname, they might have included sons of daughters of a man with the same name (and analogously for female names), and/or people named after a prominent person of the previous generation, who was not related.

Statistical testing wasn't consistent with a true power law in these frequency data, but it was reasonably close to one.

Evidence Point 2: Swiss Emigrants

I have extracted all the given names from the archive.org text of Lists of Swiss emigrants in the eighteenth century to the American colonies. There are only about 2500 names, many more men than women. A lot of cleaning up of alternative spellings was needed: there were at least 10 different spellings of Catherine, for example, and I had to allow for the use of the diminutive "li" at the end of some names (this is a typically Swiss thing, used even today about all sorts of people and objects).

It is not clear how representative of Swiss society these people were, given that emigration was actively discouraged and the commentary in the book suggests that many of the people had fallen on hard times or were regarded as in some way dissolute.

With those caveats in mind, here are the top 20 names for males and females from this source.

Heinrich        15.40   Anna         24.70
Jacob           12.30   Barbara      13.90
Hans            11.70   Elisabeth    13.10
Hans-Jacob      7.83    Verena       10.20
Felix           4.88    Margaretha   7.77
Johannes        4.73    Regula       6.17
Hans-Ulrich     4.21    Susanna      4.34
Caspar          4.21    Magdalena    3.20
Rudolf          4.14    Catherine    2.86
Hans-Heinrich   3.84    Ursula       1.94
Hans-Conrad     2.51    Maria        1.94
Ulrich          2.22    Anna-Barbara 1.94
Conrad          2.00    Cleophela    1.14
Hans-Rudolf     1.11    Babeli       1.14
Andreas         0.96    Dorothea     1.03
Heiri           0.89    Barbel       0.80
Rudi            0.74    Veronica     0.69
Salomon         0.59    Esther       0.69
Junghans        0.59    Angelica     0.34
Franz           0.59    Adelheit     0.34

Again, the top few names account for a large fraction of the total, especially when you allow for all the hyphenated names starting with Hans. If they are counted together with Hans, that name accounts for nearly 32% of all males in this source, and the top five male and female names both account for nearly 70% of their respective totals.

Evidence Point 3: Mainly Germany

Finally, using a dataset of mainly German male immigrants to Pennsylvania in the 18th century, we see a dominance of Johan/Johannes. This was a very large dataset so I have shown the top 36 (35 if you take Johan and Johannes as one). The top 6 (five if Johan and Johannes are merged) account for nearly half the males in this source. However it should be noted that some of these names represent the merged totals of as many as 20 alternative spellings.

Johan       20.30
Johannes     8.70
Jacob        7.21
Georg        4.89
Peter        3.33
Michael      2.94
Heinrich     2.72
Hans         2.58
Christian    2.54
Philip       2.20
Hans-Georg   2.11
Andreas      1.94
Friederich   1.91
Christoph    1.69
Conrad       1.64
John         1.55
Nicholas     1.54
Martin       1.37
Mathias      1.29
Job          1.22
Adam         1.13
Henry        0.97
Daniel       0.95
Caspar       0.95
Wilhelm      0.88
Ludwig       0.86
Valentin     0.80
Hans-Jacob   0.80
Joseph       0.77
Abraham      0.73
Hans-Michael 0.70
Anthony      0.59
Ulrich       0.57
David        0.54
Leonhard     0.53
Balthasar    0.53

If anyone would like the original raw data or the Mathematica notebook I used for the analysis, please let me know in comments.

Answer 3

Excellent work showing the cross-cultural regularity. Collecting the name data and summarizing it reasonably is a huge amount of work. Thank you for your contribution to public knowledge!

Unfortunately, the regularity is difficult to understand. You stated:

There is actually a reasonable explanation for this pattern: a tendency to name sons after fathers and grandfathers, together with random growth of lineages. If every father named his first son after himself or his own father, lineages of first names will tend to grow randomly, depending on family size.

That doesn't change the distribution of first names in the subsequent generation in any obvious way. Imagine every father having only one son, and every father naming his son after himself. Then the distribution of first names remains the same in the son's generation (and overall). The dynamics of name distributions depends mainly on communication across families, not naming inertia within lineages.

Much evidence indicates that, with the greatly enriched communicative environment of the Internet, symbolic concentration has increased, at least over short-time horizons. Consider the current economics of blockbusters: http://purplemotes.net/2013/11/04/elberse-blockbusters-review/

I continue to believe that name concentration is an important symbolic regularity to understand. But I don't think anyone really understands what changed with names about 1800 (at least in England), and what's happening with media and the Internet today.