Evolutionary forces in language change depend on selective pressure, but also on random chance

english-language-evolution

A new interesting paper from Nature: Detecting evolutionary forces in language change, by Newberry, Ahern, Clark, and Plotkin (2017). Discovered via Science Daily.

The following are excerpts of materials related to the publication (written by Katherine Unger Baillie), from The University of Pennsylvania:

Examining substantial collections of annotated texts dating from the 12th to the 21st centuries, the researchers found that certain linguistic changes were guided by pressures analogous to natural selection — social, cognitive and other factors — while others seem to have occurred purely by happenstance.

“Linguists usually assume that when a change occurs in a language, there must have been a directional force that caused it,” said Joshua Plotkin, professor of biology in Penn’s School of Arts and Sciences and senior author on the paper. “Whereas we propose that languages can also change through random chance alone. An individual happens to hear one variant of a word as opposed to another and then is more likely to use it herself. Chance events like this can accumulate to produce substantial change over generations. Before we debate what psychological or social forces have caused a language to change, we must first ask whether there was any force at all.”

“One of the great early American linguists, Leonard Bloomfield, said that you can never see a language change, that the change is invisible,” said Robin Clark, a coauthor and professor of linguistics in Penn Arts and Sciences. “But now, because of the availability of these large corpora of texts, we can actually see it, in microscopic detail, and begin to understand the details of how change happened.”

One change is the regularization of past-tense verbs. Using the Corpus of Historical American English, comprised of more than 100,000 texts ranging from 1810 to 2009 that have been parsed and digitized — a database that includes more than 400 million words — the team searched for verbs where both regular and irregular past-tense forms were present, for example, “dived” and “dove” or “wed” and “wedded.”

“There is a vast literature and a lot of mythology on verb regularization and irregularization,” Clark said, “and a lot of people have claimed that the tendency is toward regularization. But what we found was quite different.”

Indeed, the analysis pointed to particular instances where it seems selective forces are driving irregularization. For example, while a swimmer 200 years ago might have “dived”, today we would say they “dove.” The shift towards using this irregular form coincided with the invention of cars and concomitant increase in use of the rhyming irregular verb “drive”/“drove.”

Despite finding selection acting on some verbs, “the vast majority of verbs we analyzed show no evidence of selection whatsoever,” Plotkin said.

The team recognized a pattern: random chance affects rare words more than common ones. When rarely-used verbs changed, that replacement was more likely to be due to chance. But when more common verbs switched forms, selection was more likely to be a factor driving the replacement.

Language-evolution-hero
The grammar of negating a sentence has changed from “Ic ne secge” (Beowulf, c. 900) to “Ic ne sege noht” (the Ormulum, c. 1100) to “I seye not” (Chaucer, c. 1400) to “I doe not say” (Shakespeare, c. 1600) before returning to the familiar “I don’t say” (Virginia Woolf, c. 1900). A team from Penn used massive digital libraries along with inference techniques from population genetics to quantify the forces responsible for language evolution, such as in Jespersen’s cycle of negation, depicted here. (c) Cherissa Dukelow, 2017, license information below

The authors also observed a role of random chance in grammatical change. The periphrastic “do,” as used in, “Do they say?” or “They do not say,” did not exist 800 years ago. Back in the 1400s, these sentiments would have been expressed as, “Say they?” or “They say not.”

Using the Penn Parsed Corpora of Historical English, which includes 7 million syntactically parsed words from 1,220 British English texts, the researchers found that the use of the periphrastic “do” emerged in two stages, first in questions (“Don’t they say?”) around the 1500s, and then roughly 200 years later in imperative and declarative statements (“They don’t say.”).

old-medieval-modern-english
These manuscripts show changes from Old English (Beowulf) through Middle English (Trinity Homilies, Chaucer) to Early Modern English (Shakespeare’s First Folio). Penn researchers used large collections of digitized texts spanning the 12th to the 21st centuries to show that many language changes can be attributed to random chance alone. (c) Mitchell Newberry, 2017, license information below

While most linguists have assumed that such a distinctive grammatical feature must have been driven to dominance by some selective pressure, the Penn team’s analysis questions that assumption. They found that the first stage of the rising periphrastic “do” use is consistent with random chance. Only the second stage appears to have been driven by a selective pressure.

“It seems that, once ‘do’ was introduced in interrogative phrases, it randomly drifted to higher and higher frequency over time,” said Plotkin. “Then, once it became dominant in the question context, it was selected for in other contexts, the imperative and declarative, probably for reasons of grammatical consistency or cognitive ease.”

As the authors see it, it’s only natural that social-science fields like linguistics increasingly exchange knowledge and techniques with fields like statistics and biology.

“To an evolutionary biologist,” said Newberry, “it’s important that language is maintained through a process of copying language; people learn language by copying other people. That copying introduces minute variation, and those variants get propagated. Each change is an opportunity for a different copying rate, which is the basis for evolution as we know it.”

Featured image: copyrighted, modified from the Supplementary information of the article.

Image (c) Cherissa Dukelow, 2017, licensed under CC-BY-NC-SA 4.0 http://creativecommons.org/licenses/by-nc-sa/4.0/
Image (c) Mitchell Newberry, 2017, https://creativecommons.org/licenses/by-nc/4.0/, licensed under CC-BY-NC 4.0 (see materials at University of Pennsylvania for further sources).

Related:

Human ancestry: how to work your own PCA, ADMIXTURE analyses for human evolutionary and genealogical studies

yamna-corded-ware-bell-beaker

I wrote two days ago in the post anouncing the revised version (October 2017) of the Indo-European demic diffusion model, about dumping the information I had on doing PCA and ADMIXTURE analyses as ‘drafts’, without reviewing them, in the new section of this website called Human Ancestry.

I had some time today to review them, and to correct gross mistakes in the texts, so that they might be more usable now

I began to work with free datasets to see if I could learn something more about results of recent Genetic research by working with the available free software. For the moment, I don’t see it necessary to continue working with samples myself, because there are many professionals in Bioinformatics doing an excellent job with their publications – much better than I could do -, and publishing results early (as pre-prints) and with free licenses, which allow us to reuse and modify their material. To work again with their samples seems most of the time like reinventing the wheel.

After all, my interpretation of Indo-European migrations does not depend on my own analysis of free datasets – or on genetic analysis, or on archaeological fieldwork, for that matter – but on the study of all anthropological questions involved. I am actually more interested in Linguistics, and – only marginally – in Archaeology, as is the field of Indo-European Studies in general.

I did find certain interesting aspects that I have commented in the model, though: especially by labelling all samples and reading about them carefully (usually in the supplementary notes of the published papers), you can observe certain patterns and derive some information that others might have missed. Such examples include the Corded Ware outlier from Esperstedt (see more on the Corded Ware migration), or the differences in the three samples from early Khvalynsk.

Now that most data published seem to keep supporting what I have suggested – regarding the more complex nature of the steppe component (so-called ‘yamnaya component‘), and also regarding the migration from Yamna to Bell Beaker, and a migration of a different population (and probably language) with Corded Ware – I don’t find it worthy to spend more of my quite limited time in these tasks.

However, if I need to work again with datasets, I will try to complete the drafts the best I can. Especially regarding F3 Statistics and qpGraph, which I didn’t even try. If you want to help improve the sections, you are welcome of course.

If I find time, I might be of help with your work. And even though modern genealogy does not interest me (for the moment), I guess it can also be relevant to obtain conclusions on more recent migrations, so if I can be of any help to any interesting work, I will do it too.

plot3d-yamna
Plot 3D of datasets Minoans and Mycenaeans + Scythians and Sarmatians, using the same colours as in the Indo-European demic diffusion model.

Related:

  • The concept of “outlier” in studies of Human Ancestry, and the Corded Ware outlier from Esperstedt
  • New Ukraine Eneolithic sample from late Sredni Stog, near homeland of the Corded Ware culture
  • Human ancestry solves language questions? New admixture citebait

    human_ancestry

    A paper at Scientific Reports, Human ancestry correlates with language and reveals that race is not an objective genomic classifier, by Baker, Rotimi, and Shriner (2017).

    Abstract (emphasis mine):

    Genetic and archaeological studies have established a sub-Saharan African origin for anatomically modern humans with subsequent migrations out of Africa. Using the largest multi-locus data set known to date, we investigated genetic differentiation of early modern humans, human admixture and migration events, and relationships among ancestries and language groups. We compiled publicly available genome-wide genotype data on 5,966 individuals from 282 global samples, representing 30 primary language families. The best evidence supports 21 ancestries that delineate genetic structure of present-day human populations. Independent of self-identified ethno-linguistic labels, the vast majority (97.3%) of individuals have mixed ancestry, with evidence of multiple ancestries in 96.8% of samples and on all continents. The data indicate that continents, ethno-linguistic groups, races, ethnicities, and individuals all show substantial ancestral heterogeneity. We estimated correlation coefficients ranging from 0.522 to 0.962 between ancestries and language families or branches. Ancestry data support the grouping of Kwadi-Khoe, Kx’a, and Tuu languages, support the exclusion of Omotic languages from the Afroasiatic language family, and do not support the proposed Dené-Yeniseian language family as a genetically valid grouping. Ancestry data yield insight into a deeper past than linguistic data can, while linguistic data provide clarity to ancestry data.

    Regarding European ancestry:

    Southern European ancestry correlates with both Italic and Basque speakers (r = 0.764, p = 6.34 × 10−49). Northern European ancestry correlates with Germanic and Balto-Slavic branches of the Indo-European language family as well as Finno-Ugric and Mordvinic languages of the Uralic family (r = 0.672, p = 4.67 × 10−34). Italic, Germanic, and Balto-Slavic are all branches of the Indo-European language family, while the correlation with languages of the Uralic family is consistent with an ancient migration event from Northern Asia into Northern Europe. Kalash ancestry is widely spread but is the majority ancestry only in the Kalash people (Table S3). The Kalasha language is classified within the Indo-Iranian branch of the Indo-European language family.

    Sure, admixture analysis came to save the day. Yet again. Now it’s not just Archaeology related to language anymore, it’s Linguistics; all modern languages and their classification, no less. Because why the hell not? Why would anyone study languages, history, archaeology, etc. when you can run certain algorithms on free datasets of modern populations to explain everything?

    What I am criticising here, as always, is not the study per se, its methods (PCA, the use of Admixture or any other tools), or its results, which might be quite interesting – even regarding the origin or position of certain languages (or more precisely their speakers) within their linguistic groups; it’s the many broad, unsupported, striking conclusions (read the article if you want to see more wishful thinking).

    This is obviously simplistic citebait – that benefits only journals and authors, and it is therefore tacitly encouraged -, but not knowledge, because it is not supported by any linguistic or archaeological data or expertise.

    Is anyone with a minimum knowledge of languages, or general anthropology, actually reviewing these articles?

    Related:

    Featured image: Ancestry analysis of the global data set, from the article.

    Neolithic and Bronze Age Basque-speaking Iberians resisted invaders from the steppe

    gaul-asterix

    Good clickbait, right? I have received reports about this new paper in Google Now the whole weekend, and their descriptions are getting worse each day.

    The original title of the article published in PLOS Genetics (already known by its preprint in BioRxiv) was The population genomics of archaeological transition in west Iberia: Investigation of ancient substructure using imputation and haplotype-based methods, by Martiniano et al. (2017).

    Maybe the title was not attractive enough, so they sent the following summary, entitled “Bronze Age Iberia received fewer Steppe invaders than the rest of Europe” (also in Phys.org. From their article, the only short reference to the linguistic situation of Iberia (as a trial to sum up potential consequences of the genetic data obtained):

    Iberia is unusual in harbouring a surviving pre-Indo-European language, Euskera, and inscription evidence at the dawn of history suggests that pre-Indo-European speech prevailed over a majority of its eastern territory with Celtic-related language emerging in the west. Our results showing that predominantly Anatolian-derived ancestry in the Neolithic extended to the Atlantic edge strengthen the suggestion that Euskara is unlikely to be a Mesolithic remnant. Also our observed definite, but limited, Bronze Age influx resonates with the incomplete Indo-European linguistic conversion on the peninsula, although there are subsequent genetic changes in Iberia and defining a horizon for language shift is not yet possible. This contrasts with northern Europe which both lacks evidence for earlier language strata and experienced a more profound Bronze Age migration.

    Judging from the article, more precise summaries of potential consequences would have been “Proto-Basque and Proto-Iberian peoples derived from Neolithic farmers, not Mesolithic or Palaeolithic hunter-gatherers”, or “incomplete Indo-European linguistic conversion of the Iberian Peninsula” – both aspects, by the way, are already known. That would have been quite unromantic, though.

    Their carefully selected title has been unsurprisingly distorted at least as “Ancient DNA Reveals Why the Iberian Peninsula Is So Unique“, and “Ancient Iberians resisted Steppe invasions better than the rest of Europe 6,000 years ago“.

    So I thought, what the hell, let’s go with the tide. Using the published dataset, I have also helped reconstruct the original phenotype of Bronze Age Iberians, and this is how our Iberian ancestors probably looked like:

    Typical Iberian village during the Steppe invasion, according to my phenotype study of Martiniano et al. (2017). Notice typical invaders to the right.

    And, by the way, they spoke Basque, the oldest language. Period.

    Now, for those new to the article, we already knew that there is less “steppe admixture” in Iberian samples from southern Portugal after the time of east Bell Beaker expansion.

    portugal-bronze-age-admixture
    (A) PCA estimated from the CHROMOPAINTER coancestry matrix of 67 ancient samples ranging from the Paleolithic to the Anglo-Saxon period. The samples belonging to each one of the 19 populations identified with fineSTRUCTURE are connected by a dashed line. Samples are placed geographically in 3 panels (with random jitter for visual purposes): (B) Hunter-gatherers; (C) Neolithic Farmers (including Ötzi) and (D) Copper Age to Anglo-Saxon samples. The Portuguese Bronze Age samples (D, labelled in red) formed a distinct population (Portuguese_BronzeAge), while the Middle and Late Neolithic samples from Portugal clustered with Spanish, Irish and Scandinavian Neolithic farmers, which are termed “Atlantic_Neolithic” (C, in green).

    However, there is also a clear a discontinuity in Neolithic Y-DNA haplogroups (to R1b-P312 haplogroups). That means obviously a male-driven invasion, from the North-West Indo-European-speaking Bell Beaker culture – which in turn did not have much “steppe admixture” compared to other north-eastern cultures, like the Corded Ware culture, probably unrelated to Indo-European languages.

    portugal-bronze-age-haplogroup
    Summary of the samples sequenced in the present study.

    As always, trying to equate steppe or Yamna admixture with invasion or language is plainly wrong. Doing it with few samples, and with the wrong assumptions of what “steppe admixture” means, well…

    Proto-Basque and Proto-Iberian no doubt survived the Indo-European Bell Beaker migrations, but if Y-DNA lineages were replaced already by the Bronze Age in southern Portugal, there is little reason to support an increased “resistance” of Iberians to Bell Beaker invaders compared to other marginal regions of Europe (relative to the core Yamna expansion in eastern and central Europe).

    As you know, Aquitanian (the likely ancestor of Basque) and Iberian were just two of the many non-Indo-European languages spoken in Europe at the dawn of historical records, so to speak about Iberia as radically different than Italy, Greece, Northern Britain, Scandinavia, or Eastern Europe, is reminiscent of the racism (or, more exactly, xenophobia) that is hidden behind romantic views certain people have of their genetic ancestry.

    Some groups formed by a majority of R1b-DF27 lineages, now prevalent in Iberia, spoke probably Iberian languages during the Iron Age in north and eastern Iberia, before their acculturation during the expansion of Celtic-speaking peoples, and later during the expansion of Rome, when most of them eventually spoke Latin. In Mediaeval times, these lineages probably expanded Romance languages southward during the Reconquista.

    Before speaking Iberian languages, R1b-DF27 lineages (or older R1b-P312) were probably Indo-European speakers who expanded with the Bell Beaker culture from the lower Danube – in turn created by the interaction of Yamna with Proto-Bell Beaker cultures, and adopted probably the native Proto-Basque and Proto-Iberian languages (or possibly the ancestor of both) near the Pyrenees, either by acculturation, or because some elite invaders expanded successfully (their Y-DNA haplogroup) over the general population, for generations.

    Maybe some kind of genetic bottleneck happened, that expanded previously not widespread lineages, as with N1c subclades in Finland.

    There is nothing wrong with hypothetic models of ancient genetic prehistory: there are still too many potential scenarios for the expansion of haplogroup R1b-DF27 in Iberia. But, please, stop supporting romantic pictures of ethnolinguistic continuity for modern populations. It’s embarrassing.


    Featured image from Wikipedia, and Pinterest, with copyright from Albert Uderzo and publisher company Hachette.

    Images from the article, licensed CC-by-sa, as all articles from PLOS.

    The over-simplistic “Kossinian Model”: homogeneous peoples speaking a common language within clearly delimited cultures

    proto-greek-mynian-ware

    There seems to be a growing trend to over-simplistic assumptions in archaeology and linguistics, led by amateur and professional geneticists alike, due to the recent (only partially deserved) popularity of Human Evolutionary Biology.

    These studies are offering ancient DNA samples, whose Y-DNA and mtDNA haplogroups and admixture analyses are showing some new valuable information on ancient cultures and peoples. However, their authors are constantly giving uninformed conclusions.

    I have read a good, simple description of the Kossinnian model in the book Balkan Dialogues (Routledge, 2017), which has been shared to be fully read online by co-editor Maria Ivanova.

    Chapter 3, The transitions between Neolithic and Early Bronze Age in Greece, and the “Indo-European problem”, by Jean-Paul Demoule, offers a clear account of the difficulties found in tracing the arrival of Proto-Greek speakers to Greece or the “Coming of the Greeks”. The identifications of cultural breaks most commonly supported by academics as potentially signaling the arrival of Proto-Greeks are cited, including the Early Helladic III period ca. 2300 BC (with the diffusion of Mynian ware), or the Middle Helladic period ca. 2000 BC. The problem of finding a clear cultural break before the emergence of Mycenaean Greece (which obviously spoke an early Greek dialect) has led some to adopt a “Palaeolithic autochthonous theory” (Giannopoulos 2012), which offers still more problems than it solves.

    Of interest is his reference to Kossina in light of the recent popularity in resorting to DNA to answer all problems. It is mandatory for the field of Indo-European studies – regardless of what renown labs and journals of high impact factor are publishing – to avoid carrying on “in the steps of race based cranial measurement which enjoyed its floruit in the 19th century before fading into oblivion.”

    This is why, without denying the relationship between Indo-European languages, we need to question the validity of the overall model itself, which has shown itself to be over-simplistic in assuming the movement of permanent and long-lasting homogeneous “peoples”. More precisely, we have to criticize in details the “Kossinnian Model” underlying all those assumptions – “Kossinnian”, because of the German archaeologist Gustaf Kossina (1858–1931), well known for the famous sentence: “Cultural provinces, which are clearly delimited on the basis of archaeology, correspond in every era to specific peoples or tribes” (“Scharf umgrenzte archäologische Kultur-provinzen decken sich zu allen Zeiten mit ganz bestimmten Völkern und Völkerstämmen”). Four basic assumptions arise from this central idea:

    1. Changes in languages are due to population movements, usually involving conquest, and every migration implies a linguistic change.
    2. Archaeological “cultures” are homogenous ethnic groups, with defined frontiers, based on the model of 19th- and 20th-century nation-states and equally on the model of biological entities that reproduce by parthenogenesis.
    3. There is coincidence between language and material culture.
    4. Finally, languages are also homogenous biological entities which are autonomous and clearly delimited, and which can reproduce by parthenogenesis or by scissiparity.

    Unfortunately, none of these points is self-evident and each can be countered by a number of historical examples (Demoule 2014: 553–592).

    While I agree with the first part of the first statement attributed to the “Kossinnian model”, i.e. that languages are usually the product of population movements (either involving conquest or not), the other statements are obviously and demonstrably false, and are frequently assumed in comments, blog posts, forums, and even research articles – particularly in those based on genetic studies -, and this trend seems to be increasing lately.

    Welcome back!

    I have been trying to minimize contact with my own blogs, due to the huge amount of projects that I had – online as well as offline -, and the time-wasting nature of the dozen blogs I installed back in 2006-2009. They were (like this one) little more than dialectic in nature, with no particular aim.

    Right now I am tired of developing new ideas without publicizing them. I think I have information on some fields where other people might be interested in, and projects whose development could be interesting to share.

    For the moment, I have changed the WordPress theme to allow for an easy reading with smartphones and tablets.

    Welcome back to all subscribers!