A small telescope protocol

The HLC telescope protocol turns an exploratory conversation into a sequence with an aim, a comparison, and an endpoint. Its central proposal is to identify familiar framings before asking for alternatives. The model’s tendency to produce recognizable language becomes material for a survey rather than merely a source of answers.

Set the observation

Choose a question narrow enough to inspect. “What makes a library accessible to shift workers?” is easier to evaluate than “What is the future of public knowledge?” State what sources you have, what kind of outcome you need, and what would count as a failure. This establishes an explicit posture.

Survey and vary

1. List several common framings of the question and their assumptions.
2. Compare the list with available sources or your own examples.
3. Select one assumption to vary, rather than rejecting every familiar idea.
4. Explore a boundary with a second domain or a different lens.
5. Describe a few candidate concepts in concise form.
6. Return to prose and test their meanings, examples, and limitations.

The attractor survey and boundary search pages expand the middle steps. Mandala format can serve as an intermediate sketch when a dense field of associations is useful, but ordinary prose may be clearer for many tasks.

Close the observation

Select one result to keep, one to reject, and one unresolved question. Explain why. A retained concept should survive a plain-language definition and an example; a rejected one should leave a record of the failure rather than disappear into a polished narrative. This is integration.

The telescope analogy has a limit. The protocol is not a way to measure how rare a concept is inside a model, and the model cannot certify that an idea has never occurred before. It organizes exploration. Claims about novelty or the world still require the separate work described in Novelty needs evidence.

Decide when the session is complete

A small observation should have a small stopping rule. You might stop after comparing two lenses, after finding one usable distinction, or after identifying the source needed to resolve a disagreement. The limit keeps the protocol from rewarding continuation for its own sake.

Before ending, ask which part of the result depends most heavily on an unverified assumption. If that assumption concerns a fact, identify the record that could check it. If it concerns a value, name the people or purposes among which a choice is being made. If it concerns a definition, try a counterexample. These are different kinds of unfinished work, and each deserves a different next step.

A session that locates a precise uncertainty can be complete even when the uncertainty remains unresolved. Completion means the bounded inquiry produced an inspectable result, not that the subject has been exhausted.