Ten artykuł nie został jeszcze przetłumaczony na Polski — czytasz oryginał w języku English. Dostępny także w:Deutsch, English, Українська
Iceland's 59 Surnames That Are Not Supposed to Exist
Iceland banned surnames in 1925. The relevant sentence in the law is one line long and admits of no argument: "Ættarnafn má enginn taka sér hér eftir" — no one may henceforth take a family name. It has been on the books for a century.
And yet the former prime minister was called Geir Haarde. The Nobel laureate was Halldór Laxness. The man who created LazyTown is Magnús Scheving, and the actress in Journey to the Center of the Earth is Anita Briem. None of them is breaking the law.
Our Icelandic surname dataset holds 343 entries. 284 of them are patronymics ending in -son. The other 59 are real, heritable family names — the kind Iceland is famous for not having. This article is about that 17% of the list, why it exists, and why the most repeated statistic about it is wrong.
First, the part everyone gets right
Icelanders mostly do not have surnames. They have patronymics, and a patronymic is not a family name — it is a statement about one person's father.
Jónsson means "Jón's son." Jónsdóttir means "Jón's daughter." If Jón Einarsson has a son called Ólafur and a daughter called Anna, they are Ólafur Jónsson and Anna Jónsdóttir. Their father is Jón Einarsson. Nobody in that house shares a last name, and nothing is inherited: the name resets every generation.
This is why the head of our list looks the way it does. Here are the ten heaviest patronymics in the file. The weight column is the relative frequency the dataset assigns each entry; the 343 entries carry a combined weight of 2,335.
| # | Male form | Female form | Weight | Means | Share of file |
|---|---|---|---|---|---|
| 1 | Jónsson | Jónsdóttir | 100 | son of Jón | 4.28% |
| 2 | Sigurðsson | Sigurðardóttir | 78 | son of Sigurður | 3.34% |
| 3 | Guðmundsson | Guðmundsdóttir | 70 | son of Guðmundur | 3.00% |
| 4 | Gunnarsson | Gunnarsdóttir | 60 | son of Gunnar | 2.57% |
| 5 | Ólafsson | Ólafsdóttir | 54 | son of Ólafur | 2.31% |
| 6 | Einarsson | Einarsdóttir | 50 | son of Einar | 2.14% |
| 7 | Kristjánsson | Kristjánsdóttir | 46 | son of Kristján | 1.97% |
| 8 | Magnússon | Magnúsdóttir | 44 | son of Magnús | 1.88% |
| 9 | Árnason | Árnadóttir | 40 | son of Árni | 1.71% |
| 10 | Björnsson | Björnsdóttir | 38 | son of Björn | 1.63% |
Those ten carry a combined weight of 580 — 24.8% of the file. That concentration is not really a fact about surnames at all. It is a fact about given names: Jónsson is common because Jón was a common name for a father, and the whole distribution is a snapshot of what Icelandic men were called one generation ago.
There is a practical consequence that outsiders find startling: the Icelandic phone book is alphabetised by first name. Sorting by patronymic would scatter every family across the book and pile tens of thousands of unrelated people under "Jónsson." So Iceland does not bother. The directory — now mostly the website ja.is — lists people under their given name, with the patronymic, occupation and address alongside to tell the many Guðrúns apart. Icelanders address everyone by first name anyway, teachers, doctors and prime ministers included.
The trap in the middle of row 2
Look again at Sigurðsson / Sigurðardóttir. If you assumed the two forms were the same string with a different suffix bolted on, you would produce Sigurðurson — which is wrong — or Sigurðsdóttir, which is also wrong.
The rule is that a patronymic is the father's name in the genitive case, plus -son or -dóttir. And Sigurður has two competing genitives: Sigurðar and Sigurðs. Usage settled unevenly. The daughter-form standardised on the -ar genitive (Sigurðardóttir); the son-form mostly settled on the -s genitive (Sigurðsson), though Sigurðarson still exists.
So the male and female forms of the same patronymic can be built from different genuine genitives of the same name. This is exactly why a name dataset has to store the two forms as a matched pair rather than derive one from the other with string surgery. Iceland is one of the few places where the male and female surname lists are genuinely two lists.
Now the part everyone gets wrong
The stock statistic, repeated across naming sites, popular articles and encyclopedia footnotes, is that about 10% of Icelanders have a family name.
Statistics Iceland has actually measured it. In its report Mannanöfn og nafngiftir á Íslandi, using registry data for 1 January 2017, the breakdown of what Icelanders carry as a last name is:
| Type of last name | Share of population |
|---|---|
| Patronymic or matronymic | 82% |
| Icelandic family name (ættarnafn) | 4% |
| Other surnames, mostly of foreign origin | 14% |
Four per cent. About 12,000 people. Not ten.
The 10% figure looks like it survives by conflating the second row with the third. Add the ~14% who carry a foreign surname — the Kowalskis, the Nguyens, the Smiths who moved to Reykjavík — and you get a much bigger number for "Icelanders who don't use a patronymic." But those are not ættarnöfn. They are imported names that Iceland's naming law explicitly lets people keep. The category the 1925 law is about is the 4% one.
And the trend runs the other way from what you would guess: the probability of bearing a family name is 3.8% among Icelanders aged 50 and over, but 3.2% among those younger. Family names in Iceland are quietly shrinking.
What our file says
Our 59 family names carry a combined weight of 97 out of 2,335 — 4.15% of the file.
That is the registry's 4%, near enough to be uncomfortable. The dataset was built on the assumption that the ~10% figure was correct, which means it was expected to be underweighting these names by more than half. It is not. It landed on the measured number by accident.
Here is the head of the 59, next to the actual bearer counts Statistics Iceland published for 1 January 2017:
| Family name | Weight in file | Bearers (Hagstofa, 2017) |
|---|---|---|
| Blöndal | 7 | 324 |
| Briem | 4 | 156 |
| Hansen | 3 | 282 |
| Bergmann | 3 | 158 |
| Möller | 3 | 198 |
| Thorlacius | 3 | 155 |
| Petersen | 3 | — |
| Thorarensen | 2 | 293 |
| Thoroddsen | 2 | 206 |
| Nielsen | 2 | 180 |
| Kvaran | 2 | — |
| Zoëga | 2 | — |
| Waage | 1 | 171 |
The top of the list is right: Blöndal really is Iceland's most common family name, and our file makes it the heaviest of the 59. The middle is approximate — Thorarensen is genuinely the second most common and sits at weight 2, while Hansen and Möller outrank it in our data. At these magnitudes the whole category is tiny: Iceland's largest family name has 324 bearers, which in a country of ~394,000 is about one person in 1,200.
For scale, the Árni Magnússon Institute's register lists 640 family names in total, of which roughly 400 were actually borne by someone at the start of 2017. Our 59 are the head of that list, not the whole of it.
Why the 59 exist: a law that arrived too late
Family names came to Iceland from Denmark. Under Danish rule, an Icelander who studied in Copenhagen might come home with a surname, because having one signalled that you were educated and probably well-off. Some Latinised a personal name (Þorlákur → Thorlacius), some Latinised a place (Víðidalur → Vídalín, the first Icelandic family name, taken around 1700), some adapted a valley name (Blöndudalur → Blöndal, Brjánslækur → Briem), and some simply Danicised their patronymic into -sen (Petersen, Nielsen, Hansen — all of which are in our file).
The count grew: 108 family names in 1855, 297 by 1910. Then the Althing made it official. Law no. 41/1913, in force from 1 January 1915, legalised the adoption of new family names and charged a fee for the privilege; the approved names were entered in the Ættarnafnabók, the Family Name Book, between 1915 and 1925. Roughly 200 more were approved in that decade.
The backlash was immediate and had a leader: Bjarni Jónsson frá Vogi, who wanted family names eradicated. His objections were partly linguistic — a name in -son gives no clue whether it is a patronymic or a frozen family name — and partly social: surnames were an upper-class affectation, and nothing stopped anyone from adopting a well-known family's name. His 1923 bill to abolish them passed the lower house and died in the upper one. He came back in 1925 with a compromise that passed: everyone who already had a family name could keep it; nobody new could take one. That is law no. 54/1925, whose Article 2 is the sentence at the top of this article, and whose Article 8 repealed the 1913 law that had started it.
So the 59 in our file are not an error. They are a grandfather clause with a hundred years of momentum. Every one of them is a name that existed before 1925 and has been legally inherited ever since. Iceland did not abolish its surnames; it froze them and waited.
The law barely worked
Worth saying plainly: law 54/1925 was close to a dead letter. When it took effect there was no comprehensive national population register and no authority tracking what anyone called themselves, so people carried on adopting family names regardless. The scholar Halldór Halldórsson is said to have called it the most frequently violated law in Iceland, alongside the ban on importing nylon stockings. Real enforcement only arrived with law no. 37/1991 (which legalised the old family names, lawful and unlawful alike, and re-banned new ones), and was carried forward into the Personal Names Act no. 45/1996, which remains in force.
There is also a genuine exception, and it produces an irony Icelanders have noticed. Before 1996 a foreigner naturalising in Iceland generally had to adopt an Icelandic name. Since 1996, immigrants keep their names — and their children inherit them. So the surname category is now the one part of the Icelandic system that can still grow, but only through immigration. Jón Gnarr, the comedian and former mayor of Reykjavík, made the point sharply: a newcomer may keep a non-conforming name that a native-born Icelander is forbidden to take. Gnarr had to go to court in 2018 to have his own second name recognised as a family name at all.
The one-line version
"Icelanders don't have surnames" is 96% true and 4% false, and the 4% is legislated, documented and named. Our file reflects it: 284 patronymics, 59 family names, 4.15% of the weight against a measured 4% of the population — a category the state closed in 1925 and could not quite shut.
Data as of 2026-07-17. Figures described as "in the file" are computed from our is_IS surname dataset: 343 entries, combined weight 2,335; 284 patronymics; 59 family names carrying weight 97 (4.15%); top ten patronymics carrying weight 580 (24.8%).
Sources:
- Mannanöfn og nafngiftir á Íslandi — Statistics Iceland (Hagstofa Íslands), Hagtíðindi, 26 February 2018; data as of 1 January 2017. Source of the 82% / 4% / 14% split, the ~12,000 bearers, the 3.8% vs 3.2% age gap, and the Blöndal / Thorarensen / Hansen counts.
- Ættarnöfn á Íslandi — Árni Magnússon Institute for Icelandic Studies. The Ættarnafnabók, registration 1915–1925, 640 registered names, ~400 borne.
- Lög um mannanöfn og áhrif þeirra á nafnaforðann — Árni Magnússon Institute. Law nos. 41/1913, 54/1925 (Article 2 quoted, Article 8 repeal), 37/1991, 45/1996; Bjarni Jónsson frá Vogi; the ~200 names approved 1914–1925; the law as a dead letter.
- Icelandic name — Wikipedia. 108 family names in 1855 / 297 by 1910; the 1913 legalisation and 1925 ban; named bearers (Haarde, Laxness, Guðjohnsen, Scheving, Briem, Kvaran); phone book alphabetised by first name; the 2019 gender-neutral
-buramendment. - Personal Names Act, No. 45 of 17th May 1996 — Government of Iceland, official English text.
- Sigurður — Wiktionary. The two genitives Sigurðar / Sigurðs and the resulting
SigurðssonvsSigurðardóttirasymmetry. - A 150-Year Debate over Surnames vs. Patronymics in Iceland — Kendra Willson, Genealogy 7(4):85 (2023). Legislative history and the class connotations of surnames.
A note on verification. The match between our file's 4.15% and Statistics Iceland's 4% is a genuine agreement, not a self-check: the dataset was not calibrated against Hagstofa, and was in fact built expecting the widely-repeated ~10% figure. Where a dataset's weights are derived from a registry, agreeing with that registry is a consistency check and not independent confirmation — a distinction we try to keep honest. The bearer-count column above is such a check on ordering, and it shows our middle ranks are approximate.