[Linganth] Ling Anth and LLMs Reading Group - Dec 15 - Jack LaViolette & Mikael Brunila - Transforming the Invention of Culture

[Language Machines]

Dear Language Machines Network,

We'd like to invite you to our next reading group meeting. We'll discuss
Jack LaViolette (Columbia) & Mikael Brunila's draft article *"Transforming
the Invention of Culture: Masking, Attention, and Semiotic Alienation in
the Age of LLMs"* on Monday, December 15, 2025 from 18:00-20:00 CET. See
below for the schedule of future meetings.

Abstract: Building on recent scholarship that understands Large Language
Models (LLMs) as the materialization of a "distributionalist" language
ideology, this paper unpacks the specific mechanics of this operation for
an anthropological audience. We theorize the Transformer architecture that
powers all modern LLMs by drawing on the symbolic anthropology of Roy
Wagner, whose dialectic of invention and convention offers a precise
structural vocabulary—"masking," "attention," and "context"—that uniquely
mirrors the technical nomenclature of contemporary machine learning. By
placing these two regimes of semiosis on a collision course, we distinguish
the generative acts of human speakers from the probabilistic calculations
of the machine.

We demonstrate that while the Transformer’s "attention mechanism" simulates
context by scanning a totalizing archive of fossilized text, it
fundamentally inverts the Wagnerian process of meaning-making. For the
human subject, masking obscures the arbitrariness of convention to allow
actors to experience culture as phenomenologically real, whereas for the
model, masking serves only to enforce statistical probability, foreclosing
invention in favor of a closed loop of prediction. We further contend that
alignment techniques like Reinforcement Learning from Human Feedback (RLHF)
entrench this deficit, producing what Wagner calls a "wax museum" of
culture—static forms severed from the motivating resistance of social life.
Finally, we examine how the AI industry manipulates the Wagnerian
opposition between "Nature" (the innate) and "Culture" (the artificial). By
masking the statistical constructedness of these models behind a rhetoric
of biological inevitability, the industry presents algorithmic output as
the organic intuition of a sui generis mind. We conclude that as our
communicative practices increasingly depend on such automated replication,
we risk a distinctly semiotic form of alienation.

Please RSVP by filling out this form
<https://docs.google.com/forms/d/1d_dBldMkRgVOmsdCERvYktNpLAAE44NktxOip1IC9h8/edit>
to
receive the Zoom link and PDF of the article when it's available (not to be
circulated please)

If you are interested in presenting a draft paper yourself, please get in
touch. We have open slots starting in March.

We look forward to seeing you there!

Siri and Anna

January 19, 2026

Leif Weatherby

Talking about his book Language Machines: Cultural AI and the End of
Remainder Humanism
<https://www.upress.umn.edu/9781517919320/language-machines/>

February 16, 2026

Spencer Kaplan

Artificial Agency/“You can just do things”
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/linganth/attachments/20251201/f25590aa/attachment.htm>