This page frames and prepares a system-theoretical continuation of Niklas Luhmann’s
“Kommunikation mit Zettelkästen” (1981) under contemporary conditions. niklas-luhmann-archiv 
, luhmann_1981_AB46
The question is not whether large language models “think”, “understand”, or “replace” human cognition. The question is whether interaction with such systems can be analysed as communication in Luhmann’s strict sense: as the generation of information through system-internal comparison under conditions of contingency. The working hypothesis is that interaction with LLMs can be described as communication with an artificial partner that is structurally comparable to, but not identical with, Luhmann’s Zettelkasten.
Can communication with large language models be analysed as a continuation of Luhmann’s “communication with card indexes” — and if so, what changes?
claim
Communication with LLMs can be described as communication with an artificial partner that generates information through non-identical comparison schemata, producing surprise, contingency, and innovation — but without the durable structural memory characteristic of the Zettelkasten.
support
Luhmann’s original move (1981) was not to treat the Zettelkasten as a storage device, but as a communication partner with an Eigenleben.
The decisive features were: * operational closure * internal comparison schemes * productive surprise * order emerging from a combination of order and disorder
These features can be re-specified for LLMs: * Information remains an inner-systemic event. * Surprise arises because the system’s comparison horizons are not identical with those of the user. * Communication succeeds “on occasion”, not through guaranteed correctness. * Innovation occurs at the level of relationing relations, not retrieval of stored content.
At the same time, an important difference appears: The Zettelkasten stabilises communication through durable addresses, traceable genealogies, and cumulative internal structure. LLMs maximise combinatorial productivity while minimising durable internal memory. They behave as highly sensitive but structurally amnesic communication partners.
distinction
Zettelkasten ≠ LLM
Zettelkasten: * fixed addresses * inspectable internal links * cumulative second memory * slow, durable surprise LLM: * statistical positionality * opaque internal reference structure * retrainable capacity instead of memory * synchronous, high-frequency surprise
The continuity lies at the level of communication theory. The break lies at the level of memory and historicity.
scope
This page does not: * evaluate correctness or truth of LLM outputs * argue for or against AI adoption * anthropomorphise artificial systems It prepares a system-theoretical rewrite of Luhmann’s Sections I–IV for LLM-based communication and situates that rewrite within ongoing work on: * Zettelkasten theory * FedWiki graphs * Narrative Walks * artificial communication partners
related
* Kommunikation mit Zettelkästen (Luhmann, 1981) * Zettelkasten * Communication * Artificial Partner * Narrative Walks * System Theory * LLMs
pages/communication-with-an-artificial-partner
–– dmx