Beyond Indo-European

Most of this course has used Indo-European, because it is the best-studied family and contains English. This chapter looks at what lies outside it, and at languages that fit the family tree badly or not at all. It also keeps a promise from Chapter 5: Hungarian is not on the Indo-European tree, and here is where it belongs.

The Uralic Family and Hungarian

Wikipedia's article on the Uralic languages describes a language family native to northern Eurasia, with an estimated 21.3 million native speakers, concentrated in Central, Northern and Eastern Europe and Northern Asia. The page says the family is "not part of what is now known as the Indo-European family."

Erzya, Moksha, Mari, Udmurt and Komi each have over 100,000 speakers, all in European Russia.

The branches

• The family has nine undisputed groups: Sámi, Finnic, Mordvinic, Mari, Permic, Hungarian, Mansi, Khanty and Samoyedic.
• Finno-Ugric is a disputed subdivision that excludes Samoyedic, and it is sometimes used as a synonym for Uralic.
• The page presents Donner's 1879 model, a traditional tree with Ugric and Finno-Permic branches, but notes that higher-order groupings are increasingly doubted.
• Hungarian belongs to the Ugric branch with Mansi and Khanty (Ob-Ugric); some scholars instead group Hungarian with Mansi.

Cognates: the evidence at work

The page's table of Proto-Uralic cognates shows the comparative method (Chapter 3) at work on Hungarian:

Look for the correspondence sets: Finnish k- against Hungarian h- (kala, hal) but Finnish k- against Hungarian k- (käsi, kéz); Finnish -s- against Hungarian -z (vesi, víz; käsi, kéz). The first pair differs by the vowel that follows, which is the kind of complementary distribution that Chapter 3 described, though I have not checked that this is the source's analysis. The page notes that cognates are not translations, since meanings can shift and loanwords can replace them. Numerals are less settled: Janhunen's proposal reconstructs Proto-Uralic *kektä "2" and *wixti "5", and Hungarian hét "7" is a replacement form.

How was the relationship found?

• Finnish-Hungarian affinity was first proposed in the late 17th century by Fogel, Stiernhielm and Skytte.
• In 1717, Rudbeck proposed about 100 Finnish-Hungarian etymologies, of which about 40 are still considered valid.
• Sajnovics (1770) argued for a relationship based on grammatical features, and Budenz later attempted a Proto-Finno-Ugric reconstruction.
• Shared traits include agglutination, extensive case systems (Finnish has 15) and vowel harmony in many languages.

Where did Proto-Uralic come from?

Several homelands have been proposed, including the Volga region, western Siberia and central Siberia. Grünthal et al. (2022) place it east of the Urals, near the Minusinsk Basin, and reject a Volga/Kama origin. Zeng et al. (2025) link Uralic speakers to ancient populations in Yakutia around 4,500 years ago. As in Chapter 4, the scholars disagree, so this course gives the views and does not choose.

Language Isolates

Chapter 5 defined an isolate as a language with no proven relatives, a family of one. Basque is the standard example, and the source estimates about 129 isolates today. "No proven relatives" is a statement about evidence, not a claim that the language came from nowhere: the comparative method needs enough regular correspondences, and for an isolate they have not been found.

Creoles

Wikipedia's article on creole languages defines a creole as a stable contact language that forms when a language simplifies and mixes into a new form, often a pidgin, which then expands into a full language with native speakers in a fairly brief period. A creole has a consistent grammar and a large stable vocabulary, and children acquire it as a first language.

• A creole's lexicon is largely supplied by the parent languages, especially those of the most dominant group. Phonetic and semantic shifts are common, and the grammar often has new features that differ substantially from the parents'.
• Creoles do not fit the 19th-century neogrammarian tree model, which assumes regular sound changes. Phylogenetic classification usually relies on inherited core vocabulary and grammar; in creoles the core lexicon often has mixed origins and the grammar is largely original.
• The page says the question of which language is the parent of a creole often has no definitive answer and can lead to long-running controversies. The European dialect origin hypothesis treats French creoles as phylogenetically based on French; the monogenetic theory traces Atlantic creoles to a single Portuguese-based pidgin; others argue creoles are structurally no different from other languages and that "creole" is a sociohistoric rather than linguistic concept.
• Examples on the page include Haitian Creole ("the most creole of creoles" in McWhorter's view) and Sranan, which has existed for centuries in a diglossic relationship with Dutch.

Planned Languages: Esperanto

Esperanto is a constructed (planned) language, which Wikipedia calls "the world's most widely spoken constructed auxiliary language". L. L. Zamenhof created it and published the first grammar, Unua Libro, in 1887 under the pseudonym "Doktoro Esperanto". Its vocabulary, syntax and semantics come mainly from Indo-European languages: about 80% of its vocabulary derives from Romance languages, with the rest from Germanic, Greek and Slavic languages. The infobox lists it as an international auxiliary language and an a posteriori language.

• Native speakers: the infobox gives "c. 1,000 (2022)", but the body text says an estimated 2,000. The page's own figures disagree.
• Second-language speakers: the infobox gives 30,000 to 2 million. Individual estimates are about 2 million (Culbert), above 2 million (Wandel) and 30,000-180,000 (Vendelbo); the body text also cites two estimates of around 100,000 Esperantists.

The page does not say whether Esperanto sits outside every language family. My reasoning: it was not inherited by descent from a proto-language, so the family model does not apply to its origin, and it borrows from several families. That is a conclusion drawn from the facts above, not the source's statement. Compare the course's earlier material: an Esperanto word such as the ones in Esperanto Word-Building 1 comes from Romance, Germanic or Slavic roots chosen in 1887, not by regular sound change.

Try It Yourself