ARCHaiC 8
Open and Closed
The rapid expansion of AI marks a historic shift in how we write, communicate, and think. It’s a distinct set of circumstances, but only the latest in a series of “hinge moments” in linguistic history, though some of these “moments” may have been thousands or even millions of years in the making.
MRI scans reveal surprising similarities between humans and chimpanzees in language-related areas of our brains. It’s possible a neurological foundation for language might date to our last common ancestor, which would have been 6-7 million years ago.
All creatures communicate, but other animals use “closed language systems” with a fixed number of signals, each linked to a specific purpose, sounds not combining to form new ideas. Only modern humans use an “open language system” that allow us to join parts, make new words and sentences, form new meanings. The opportunity to create is infinite, if we take advantage of it.
The first known human species, Homo habilis, appeared in Africa almost 3 million years ago. Homo habilis was a tool maker who likely needed to share some instructions to pass along skills. Brain impressions and the shape of the vocal tract suggest the beginnings of proto-language skills but not fluent speech.
Of the twenty-or-so previous human species besides our own, only Neanderthals and a couple of others possessed the physical/genetic traits required for speech.
When Homo sapiens first emerged in Africa about 300,000 years ago, we likely still relied on gestures and a limited number of vocalizations. Anatomic and genetic data suggests that by 100,000 years ago we were fully wired for modern human speech. Complex burial rituals and advanced artwork offer evidence that by 50,000 years ago we had a true open language system.
With an open language system came more complex cultures, accelerating technology, and survival advantages.
Picture an isolated band of humans, 40 or so thousand years ago, advancing further north into ice age Europe for the first time. Perhaps they’d been pushed from their previous settlement by a territorial dispute, or that resources had dwindled for some other reason.
Their position is precarious. Wind whips incessantly across the rocky plains, mammoth-hide shelters must be reinforced. Food supplies are dwindling and they can’t afford internal conflict. Nuanced communication is absolutely essential.
The territory into which they are heading is utterly unfamiliar to them. A small scouting party has been sent ahead to gather information. The main group spends several nervous days waiting, uncertain. The scouting party finally returns and happily shares what they’ve learned with the others. They discuss options with elders and plot a course forward.
Can we know what language they spoke? The Indo-European language family (which includes English) spread into Europe less than 10,000 years ago. It replaced whatever was there previously, so we literally have no idea what language these ice age migrants would’ve spoken. But we can use our imagination, a few educated guesses, and perhaps some computer assistance.
Not all languages have definite articles, adjectives, adverbs, prepositions or conjunctions. But all languages do have subjects, objects, and verbs, even if the order varies language-to-language.
Sticking with nouns and verbs, I feed Anthropic’s Claude a vocabulary list of a dozen or so words and ask it to generate corresponding words in a theoretical ice age language, emphasizing that the words not resemble anything from the Indo-European language family or any other known language families.
The results are impressive. Claude not only generates words, but explanations behind each one. A few examples:
Fire = gura (round, warm sound)
Food = mun (nasal, close to “eating” mouth sounds)
Fight = brakka (harsh, aggressive)
I ask for more words, then feed Claude sentences that convey meaning using only nouns and verbs. Claude decisively standardizes my sentence structure, firmly establishing a subject-object-verb word order common in archaic languages. It implements grammatical rules such as:
suffix -ta makes a verb past tense
suffix -na makes a noun plural
prefix ni- on the verb makes it negative
Here are the results:
Sa yunna-bora lem-ta, throm tok-wassa-du, throm-tselm gan-na-tsa dorak-brakka-ta. Sa durn-ta.
“We moons-many walked, across vast rock-plain, against browridge-people weapon-fought. We survived.”
Sa yunna-bora lem-ta, nolu seka-ta. Throm-tselm gan-na ni-hamsu-ta. Sa shiv am-ta, kelu thagor-ta, kelu am-ta. Sa hamsu vor-ta, noru. Tutu-yalen.
“We moons-many walked, valley found. Browridge-people not-settling. We plants ate, deer hunted, deer ate. We settle can, safe. Baby-season.”
I picture this as the elevator speech, delivered breathlessly upon arrival back at camp. Now imagine the more complete description with all the details, shared around the fire that night with the whole group in the oral storytelling tradition.
Elders weigh in, comparing new information with generational survival knowledge. Questions are asked by eager youngsters. Side conversations take place, hopes, speculations, dreams for the future…
Whether human development of an open language system fits neatly into a “moment” or not, it constitutes a transition as significant as tool use and the mastery of fire. It set us on the path toward abstract thought, philosophy, the lightbulb, the atomic bomb, the moon landing, Artificial Intelligence.
Is AI now using an open language system much like we do? It gives the outward appearance of doing so, but because its linguistic output is based on computation rather than awareness and experience, many linguists would say no.
Maybe the question is less about what AI is doing and more about what we’re doing with it.
History will decide whether the introduction of Large Language Models marked a new height in human linguistic achievement, or was the moment we began to hand our evolutionary advantages over to a machine.
—
Sources:
Hecht, Erin E. “How Did Language Evolve? Comparing Human & Chimpanzee Brains.” Harvard Brain Science Initiative, 2026. https://brain.harvard.edu/hbi_news/how-did-language-evolve-comparing-human-chimpanzee-brains/
Hecht, Erin E. et.al. “Individual Variation in the Chimpanzee Aruate Fasciculus Predicts Vocal and Gestural Communication.” National Library of Medicine. April 17, 2025 https://pubmed.ncbi.nlm.nih.gov/40246833/
Roos, Dave. “When Did Ancient Humans Start Talking?” History. July 13, 2026 https://www.history.com/articles/early-humans-language
Lieberman, Philip and McCarthy, Robert. “Tracking the Evolution of Language and Speech: Comparing Vocal Tracts to Identify Speech Capabilities.” Expedition Magazine, Penn Museum, volume 49 / number 2. 2007 https://www.penn.museum/sites/expedition/tracking-the-evolution-of-language-and-speech/
Levinson, Stephen C. and Holler, Judith. “The Origin of Human Multi-Modal Communication.” Philosophical Transactions of the Royal Society B, Vol. 369, Issue 1651. September 19, 2014. https://pmc.ncbi.nlm.nih.gov/articles/PMC4123681/
Krzanowski, Roman. “Intentionality and the Limits of LLMs.” Conference on Philosophy of Artificial Intelligence, Amsterdam. Oct. 2025 https://www.researchgate.net/publication/396440035_Intentionality_and_the_Limits_of_LLMs
Claude: https://claude.com
—
Selected vocabulary of the Nolu-Gan (Valley People):
fire = gura
rock = tok
plain = wassa
food = mun
valley = nolu
plant = shiv
deer = kelu
hunt = thagor
fight = brakka
eyebrow = tselm
weapon = dorak
grow = velu
settle = hamsu
safe = noru
baby = tutu
season = yalen
we = sa
walk = lem
many = bora
moon = yun
stretching/vast = throm
people = gan
survive = durn
find = sek
not (prefix) = ni-
eat = am
can/able = vor
—
—
Click below for…