On the Speed of Language Change

In the prospect that you follow an interest in linguistics, you might have run over this video (How long can a language last before it's unrecognizable?), second of two parts, on NativLang’s channel.

It delves, though in a glancing fashion (as is both custom and intended of the format), into the speed at which languages in general tend to evolve, and ultimately become unrecognizable from their former selves. I do very much love the approach taken by this one (and NativLang in general!), but I feel in this particular instance it kind of glosses over the biggest point.

The most striking example I see put forward is how Norwegian and Icelandic present a strong refutation of the premise that with enough linguistic data, one can reliably estimate a language or language family’s age, since those two show completely different stages of evolution even after having split away from the same source, roughly a thousand years ago. While certainly not wrong, this and most of the narrative followingly seems to pit two sides of the debate onto one-another: a mathematically convenient but unachievable formula, versus a more prosaic, seemingly unpredictable reality. Passing mentions are made of the effects of small population sizes and eras of major social upheavals on the matter, but the discussion ends short of them due to the format’s nature.

Landing back to my take on this. As a hobbyist linguist, I have a personal specialty in Greco-Latin languages, especially on the evolutionary aspect, for example how features unfold from Proto-Italic, to Old, Classical and Vulgar Latin, toward Proto-Romance and finally land within the Romance spectrum. Within the constraints of the languages I’m therefore most familiar with, I can attest a clear pattern: Greek and Latin are both marked by relative linguistic stability in pre-classical (though lightly attested) times, with limited phonological and grammatical changes occurring slowly throughout the millenniums, then followed by sudden and absolutely major changes in limited timespans, with different onsets for each.

Greek, first under the conquests of Alexander, with its resultant major societal unification and increase in exchanges and trade around the Aegean, undergoes a phase of dialectal levelling leading to the emergence of a ‘Koine’⁠—already similar forms get rapidly assimilated into a prestige norm, strongly influenced by literary forms, yet sharing features of many of its constituents therein. The transition occurs relatively quickly, mostly within the constraints of Classical Greece, and leads to some shifts, mostly around phonology and especially toward simplification of vowels and diphthongs, one of the most divergent features between the original dialects.

Later, through basically the entire right half of the Mediterranean having come under control of Greek states, huge swaths of foreign speakers enter its linguistic community, often through local merchants and elites picking up Greek as a trade and renown language, while keeping bilingualism with their region’s original speech. We start to see rapid and substantial shifts and simplifications in phonology, loss of vowel length and as a result pitch accent, replacement of core vocabulary, borrowing of local manners of speech, loss of some grammatical cases and suppression of irregularities; all features forming the heart of the hardest things in a language to learn.

Latin, for its part, remained united up until the onset of the 1st Century BCE, with the appearance of a vulgar form right at the time of the first major conquests outside the Italic-speaking peninsula, into Gaul, Hispania, North Africa, and in the end quite literally the entire known world at the time (kudos to the romans). This “vulgar” romanice shows accelerated phonological changes (initially less marked than Greek’s), loss of all grammatical cases (partly influenced by phonological mergers, yet clearly far more exhaustive than anticipated from them), replacement of most of the core vocabulary and the near entirety of classical function words, simplification of irregularities, and so on.

By the 8th Century, the distinction between Classical and Vulgar registers came to be so far-reaching as to qualify as two completely different languages, yet the new Proto-Romance remained united as a linguistic community until very late, with current romance descendants sharing the vast majority of the post-expansion innovations.

As seen, this does seem to concur with one of the aforementioned but underdeveloped hypotheses of the video: major societal changes are drivers of major linguistic shifts. Close-knit communities⁠—sometimes on the scale of small neighboring villages⁠—, cruise along without much change for centuries on end, and when faced with massive inflow of second language learners from disparate areas, their languages adapt to become easier to learn and exchange with.

Beyond this and later in medieval times, following frankish conquests of the north of Gaul and the establishment of germanic kingdoms, we see the area around the center of modern-day France become a hotbed of exchanges and the stem of most linguistic innovations within the romance continuum. Early medieval french dialects⁠ develop major and most of the sound shifts⁠—especially northern ones most in contact with germanic settlers and nobility⁠—, which are then carried over radially to other romance speakers inside Europe. Only a select few features survive as the distance progressively grows, up to Spain and sometimes crossing Italy’s mountain ranges, but rarely if any into the remote reaches of the peninsula, and especially crossing water to Sardinia.

Similarly, the original example contrasts the continental Scandinavian dialects⁠—Norwegian, Swedish, Danish⁠—to Icelandic, the former which had regular contact with neighboring Germanic, Slavic and Finnish linguistic communities and came to develop various innovations, while the latter kept remote.

As a final case, English itself had its own event of ‘major upheaval’. Following the Viking conquests and the establishment of the Danelaw in the 9th Century onwards, substantial exchanges between Old Norse settlers in the northeast of Britain and local Old English speakers began. With the two languages sharing many features but differing mainly through grammatical cases and verb inflections, leading into the Middle English period those slowly fell out of use to be replaced by a more rigid, yet more unanimous system based on analytical formations and word position alone. Coupled with later French influences, it is therefore sometimes said that the language became (at least partially) a creole of those multiple sources, a major shift considering its relative stability before then, and of its sister continental germanic languages, which often maintain many of those features to this day.

To conclude, while I would certainly agree there certainly is no simple formula to derive the rate of language change, or conversely its age based on comparative analysis from features and lexicon alone, neither would I say the matter is truly so obscure as to be unpredictable. There is a clear contrast in the speed at which phonological, grammatical and lexical evolution occurs, not just between societies undergoing major cultural and population expansion or exchanges and those that remain stable, but also more broadly between ‘insular’ and ‘connected’ ones. Be it entire islands like England, Iceland and Sardinia, or small, disparate rural villages, isolation nurtures stability and exchanges nurture change, but in a manner I feel ‘social upheaval’ is too broad to describe. Scale seems to matter little, instead being the events that do so: while languages naturally evolve over time as to self-optimize, accelerated shifts do not occur in a vacuum but in reaction to problems at hand—trade, communication and easier learning, expansions, bilingualism and influence of substrates or neighbors, etcetera.