Source
Hutchins (1995), LIT-1066 (Cognitive Science 19(3), 265–288), read in NOTE-867 from the author’s posted scan. The book Cognition in the Wild (LIT-1060) is Unavailable; only its posted Introduction was seen.
What was actually shown
The method. The unit of analysis is a functional system, “a constellation of structures, some of them internal to the human actors, some external, involved in the performance of some invariant task”. The account is “somewhat stylized”, from one airline’s MD-80 manual, informed by more than 100 observer flights, simulator recordings and interviews. Three descriptions of one activity: the procedure, the representations outside the pilots, and what those leave the pilots to do inside.
The retrieval. The speed-card booklet holds speeds for each weight interval, unchanged by crew action. The gross weight, displayed on the fuel quantity indicator, “is used as a filter on this written memory”, and since weight and speeds share a card, “selecting the correct weight can’t help but select the correct speeds.” The card posted where both pilots see it allows cross-checking against the displayed weight.
The transformation. The speeds are spoken, read back, and set as bugs on both airspeed indicators. “What was accomplished without speed bugs by remembering speed values, reading the ASI needle values, and comparing the two values is accomplished with the use of speed bugs by judgments of spatial proximity. Individual pilot memory has not been enhanced; rather, the memory function has now become a property of a larger system in which the individual engages in a different sort of cognitive behavior.” What the pilots must hold is the meaning of each region of the dial, not the speeds.
Across time. “Speed bugs permit a shift in the distribution of cognitive effort across time. They enable the crew to calculate correspondences between speeds and configurations during a low workload phase of flight, and save the results of that computation for later use.” Landing data are prepared “about 25 to 30 minutes prior to landing” and used “ten to fifteen minutes later”, and the bugs are “relatively insensitive to the interruptions, the distractions, and the delays that may disrupt internal memories.”
What this does not say
- That the claims are measured. No timings, error rates or comparison of crews or procedures. The system without bugs is imagined, and the pilots’ internal memories are inferred: “Lacking additional evidence, we cannot know the duration or quality of these memories.”
- That the bugs give a better outcome. The bugged system “achieves the same results” as one without; the gain claimed is lower demand on scarce resources at the moment of use and robustness to interruption.
- That pilot memory plays no part. The memory process “emerges from the activity of the pilots”; the claim is that the cockpit’s memory “is not made primarily of pilot memory”, not that pilots remember nothing.
- That it covers unanticipated retrieval or search. The task is mandated, done “on every flight”, with four values needed in a fixed order and known in advance. Nothing here bears on finding one item among many on a request under a deadline (CASE-085’s situation).
- That such systems do not fail. Only success is analysed: “In all of my observations, never have I seen a crew fail to compute and set the approach speeds.” Missed speeds are mentioned, not examined.
- That it generalises beyond this cockpit. One aircraft type, one task, one airline’s manual; the paper says similar procedures exist on every airliner, without showing them.
Connections
- QUESTION-101 asks, for a stated class of tasks, when availability is a property of the system and not of a member, with a case where changing the arrangement alone changes it. This is that case argued in principle, for anticipated, invariant tasks prepared in advance: the system remembers when the task’s structure is built into artifacts, the cue (displayed weight, the needle nearing a bug) supplied by the environment.
- CLAIM-296 holds that deployability is a relation, not a property of one element. This corroborates it, the nearest case the record holds, without the claim resting on it, and adds a factor the claim does not list: when the work of retrieval is done relative to the deadline.
- QUESTION-095 starts from this case: a result computed once at low load and read back by perception. THEORY-331 and THEORY-342 price the machine-learning versions of moving computation before the request.
Note of 2026-10-29: what holds it back
Read in full (NOTE-867), but its promote_when is unmet. The record holds no measured comparison of crews with and without speed bugs under the same load that records timing, errors and the pilots’ recall. The paper is a stylized task analysis of successful flights, as the title’s “argued …, not measured” already says. An Active “Hutchins argues” reading on THEORY-1110’s pattern would free nothing, now that CLAIM-296 cites it as corroboration.