The Nostratic hypothesis is a highly controversial proposal in historical linguistics. It posits a linguistic macrofamily, "Nostratic," which groups together several major language families of Eurasia.
Core Concept
The hypothesis suggests that these language families descend from a single common ancestor, "Proto-Nostratic," which would have been spoken roughly 12,000 to 15,000 years ago, during the Upper Paleolithic era.
Included Language Families
The specific families included vary by proponent, but the core "Moscow School" definition (by Vladislav Illich-Svitych and Aharon Dolgopolsky) typically includes:
Indo-European (e.g., English, Spanish, Hindi, Russian)
Uralic (e.g., Finnish, Hungarian, Estonian)
Altaic (a disputed grouping itself, often comprising Turkic, Mongolic, and Tungusic languages)
Kartvelian (South Caucasian languages, e.g., Georgian)
Dravidian (South Indian languages, e.g., Tamil, Telugu)
Afroasiatic (often included, e.g., Arabic, Hebrew, Hausa)
Evidence
Proponents base the hypothesis on:
Systematic Sound Correspondences: Attempting to show regular sound changes between the proposed cognates (shared root words) across the different families.
Shared Core Vocabulary: Identifying a list of proposed cognates for basic concepts like pronouns (e.g., $m-$ for 'I', $t-$ for 'you'), body parts, and natural elements.
Grammatical Parallels: Arguing for shared morphological elements, such as pronominal systems and case endings.
Status and Criticism
The Nostratic hypothesis is not accepted by mainstream historical linguistics. The skepticism is based on several major criticisms:
Time Depth: Most linguists believe the comparative method, the standard tool for proving genetic relationships between languages, is unreliable beyond a time depth of 6,000 to 8,000 years. Proto-Nostratic would be far older.
Methodology: Critics argue that the proposed cognates are often based on loose semantic and phonetic resemblances. They contend these similarities are more likely due to:
Chance: Given the vast number of words in the parent languages, chance resemblances ("look-alikes") are statistically likely.
Borrowing: The words could be ancient loanwords (Wanderwörter) that spread across Eurasia, rather than inherited forms from a common ancestor.
Unstable Groupings: The hypothesis relies on other disputed macrofamilies, most notably the "Altaic" family, which many linguists believe is a Sprachbund (a group of languages with shared features due to contact) rather than a genetic family.
Here are prominent examples used to support the Nostratic hypothesis, along with the common criticisms of them.
1. The Pronominal Evidence (M-T Pattern)
This is widely considered the strongest evidence by proponents. It observes that a vast number of language families in the proposed group use a first-person pronoun (for "I" or "me") based on an $m$-sound and a second-person pronoun ("you" or "thee") based on a $t$-sound.
| Proposed Proto-Nostratic | Language Family | 1st Person ("I/Me") | 2nd Person ("You/Thee") |
*mi (or *män-) | Indo-European | Proto-Indo-European *me- (English me, French moi) | Proto-Indo-European *tu- (English thou, French toi) |
| Uralic | Finnish minä | Finnish sinä | |
| Turkic | Turkish ben (from earlier *män) | Turkish sen | |
| Mongolic | Mongolian bi (from *mi) | Mongolian či (from *ti) | |
| Kartvelian | Georgian me | Georgian šen (from *t'wen-) |
Criticism:
Critics argue that pronouns are "special" words. The sounds $m$ (a nasal) and $t$ (a dental stop) are among the most basic and common sounds in human languages. Furthermore, "nursery words" (like "mama" and "papa") show that certain sound-meaning connections can arise independently in different languages without a common ancestor. The $m$/$t$ pattern could be a coincidence or a very ancient, borrowed feature, not a sign of genetic inheritance.
2. Lexical Cognates (Shared Roots)
Proponents have compiled hundreds of proposed shared root words. These often relate to basic concepts like body parts, natural phenomena, and simple actions.
| Proposed Nostratic Root | Meaning | Proposed Cognates |
*K'ES- | "to cut" | Indo-European: Proto-Indo-European Turkic: Proto-Turkic |
*BER- | "to bear, carry" | Indo-European: Proto-Indo-European Turkic: Proto-Turkic |
*K'ÜN- | "who, what" (interrogative) | Indo-European: Proto-Indo-European Uralic: Proto-Uralic |
*BAR- | "grain, seed" | Indo-European: Proto-Indo-European Afroasiatic: Proto-Afroasiatic |
Criticism:
This is the most heavily criticized part of the hypothesis.
Semantic Looseness: Critics point out that the meanings are often too flexible. In the
*BAR-example, proponents may link words for "barley," "seed," "pebble," and "grit" under one root, which critics see as "cherry-picking" vague resemblances.Chance Resemblance: With thousands of words in each proto-language, a certain number of short, basic words will look similar just by random chance.
Onomatopoeia: Words for actions like "cut," "spit," or "blow" can be onomatopoeic and thus arise independently.
Ancient Borrowing: These words could be ancient loanwords (Wanderwörter) that spread across Eurasia through contact, not from a single ancestor.