Language as a Clue to Prehistory

[lpetrich]

Abstract Profiles of Structural Stability Point to Universal Tendencies, Family-Specific Factors, and Ancient Connections between Languages | PLOS ONE

Uses the World Atlas of Language Structures - WALS Online - Home - to try to find out what language structures are the most and least stable by finding out how much they are conserved in language families.

WALS is very comprehensive, though limited to present-day languages, and it needs to be expanded to past ones and to reconstructed protolanguages.

A curious complication that they note: Languages Evolve in Punctuational Bursts | Science

Linguists speculate that human languages often evolve in rapid or punctuational bursts, sometimes associated with their emergence from other languages, but this phenomenon has never been demonstrated. We used vocabulary data from three of the world's major language groups—Bantu, Indo-European, and Austronesian—to show that 10 to 33% of the overall vocabulary differences among these languages arose from rapid bursts of change associated with language-splitting events. Our findings identify a general tendency for increased rates of linguistic evolution in fledgling languages, perhaps arising from a linguistic founder effect or a desire to establish a distinct social identity.

Back to the first-mentioned paper.

We found that across all language families and datasets, the correlation between path length and number of nodes is very high (range 0.65–0.80, mean = 0.75, sd = 0.046), suggesting that punctuational bursts might explain about 50% of structural change.

 

[lpetrich]

The most stable features (most stable first):

Absence of Common Consonants -- Front Rounded Vowels -- The Optative -- Vowel Nasalization -- Obligatory Possessive Inflection -- Order of Genitive and Noun -- N-M Pronouns -- Nominal and Locational Predication -- Uvular Consonants -- M-T Pronouns -- Order of Object and Verb -- Order of Numeral and Noun -- Numeral Classifiers -- Order of Subject and Verb -- Tone

The least stable features (least stable last):

Locus of Marking in the Clause -- Voicing in Plosives and Fricatives -- Symmetric and Asymmetric Standard Negation -- Applicative Constructions -- Relationship between the Order of Object and Verb and the Order of Adjective and Noun -- Order of Person Markers on the Verb -- Indefinite Articles -- Asymmetrical Case-Marking -- Definite Articles -- Third Person Pronouns and Demonstratives -- Position of Polar Question Particles -- Number of Cases -- Ordinal Numerals -- Consonant-Vowel Ratio -- Consonant Inventories

The most stable feature: Absence of Common Consonants - absence of bilabials and/or fricatives and/or nasals. For all three to be absent, there must be no /p/ or /b/ or /f/ or /v/ or /m/.

The least stable feature: Consonant Inventories - how many consonants treated as distinct and not variations of the same consonant

 

[lpetrich]

For each language family, the authors worked out its "stability profile", the relative stability of each feature in the family. I find it curious that the authors used variations instead of average values, because average values are also important information. Northern Eurasian families often have m-t personal pronouns and western North American ones often have n-m ones, but by doing so, both would rank high in stability.

Different families have different stability profiles, and they compared the profiles of different families and they found some odd patterns. These profiles vary by geography, with regional clusters of similar profiles.

They found that the Americas are one cluster, with well-defined subclusters in North America and South America, though not in Central America. Siberia clustered with North America; Siberia: Tungusic, Yukaghir, Chukotko-Kamchatkan.

Core Eurasia is a well-defined cluster: Indo-European, Uralic, Altaic (Turkic, Mongolic), North Caucasian (NE, NW), as is Eurasia more generally: Core Eurasia + Siberia.

For Nostratic:
v1 (narrow) - IE, Uralic, Altaic - significant
v2 (broad) - IE, Uralic, Dravidian, Afroasiatic - not significant
So there is at least some support for Narrow Nostratic or Eurasiatic, even if not for Broad Nostratic.

Austro-Tai is somewhat significant, as are the Papua New Guinea families taken together, but neither is very close to Australian.

Likewise, African outside of Khoisan is somewhat significant, but adding Khoisan is not.

 

[lpetrich]

1st Americans came over in 4 different waves from Siberia, linguist argues | Live Science - "The languages of the earliest Americans evolved in 4 waves, according to one expert."
noting
Founder effects identify languages of the earliest Americans - Nichols - American Journal of Biological Anthropology - Wiley Online Library by Johanna Nichols

Nichols surveyed 16 features of these languages, including syllable structure, the gender of nouns and the way consonants are produced when speaking.

That latter one is the point of articulation, the place where the consonant sound is produced. She also included labialization (+w) and palatalization (+y). Point of articulation is fairly stable, while voicing is not as stable.

The languages split into two main groups: an early one where the first-person pronoun has an "n" sound while the second-person pronoun has an "m" sound, and a later group with languages that incorporate a sentence's worth of information in just one word.

That latter one is polysynthesis. By English standards, it is packing pronouns, auxiliary verbs, and common adverbs into a verb.

Further linguistic analysis indicated that people arrived in the Americas in four distinct waves. The first occurred around 24,000 years ago, when massive glaciers covered much of North America. Nichols found no unique language features, suggesting a diverse set of people and languages entered North America at that time. A second wave of people around 15,000 years ago brought languages with n-m pronouns, while a third wave 1,000 years later brought languages with simple consonants. A fourth wave around 12,000 years ago then brought complex consonants.

 

[lpetrich]

From Johanna Nichols's paper:

No features likely to be unique to the first opening have been found; it appears to be a well-mixed, diverse set of languages.

Or else that first arrival is long enough ago to allow for plenty of divergence in language features.

The second opening is marked by n-m pronouns and the Penutian languages (nearly all of which have n-m pronouns). n-m pronouns are found densely in different families (Penutian and other) of Oregon-California, but only once north of the Columbia in the small Tsimshianic family.

She then notes that personal pronouns and their paradigms are very resistant to borrowing. That is some of the main evidence for the Afroasiatic language family, for instance.

The third opening is distinguished by later-stratum features and simple consonant systems; the fourth opening is similar but is saliently marked by complex consonant systems and numeral classifiers.

... Languages of the fourth opening fit typologically with some of the Paleosiberian languages of eastern Siberia and the whole set defines a North Pacific Rim linguistic population.

... The AET languages from a later entry to some extent fit typologically with the descendants of the fourth opening, and both AET and Eskimo-Aleut have North Pacific Rim affinities.

AET = Athabaskan-Eyak-Tlingit (Na-Dene)

How stable are these features? Are any of the structural features at least as stable as personal pronouns?