Memory and Response Selection

The memory chain

Memory is three stores in series, each with its own capacity and clock. First the sensory stores, one per sense: they hold raw stimuli just long enough for relevance to be judged, which is their whole importance, for without that brief buffer anything not instantly attended to would be gone before the attention mechanism could select it. The two that matter in aviation: the echoic store for sound, the longest-lasting at 2 to 8 seconds and replayable (you can still count the clock's chimes after they stop), and the iconic store for vision at only 0.5 to 1 second — brief, even though some 80 percent of processed information arrives by eye.

What attention selects passes to working memory (the short-term store), which holds it for 10 to 20 seconds unless actively rehearsed, losing it to decay and to interference from newer or even older material. What rehearsal consolidates passes on to long-term memory, where storage is effectively unlimited in duration and the weakness moves to retrieval: the item is in there somewhere, and access is slow or fails. Acoustic material survives working memory better than visual, being easier to rehearse — one reason read-backs are spoken.

Working memory: the narrow shelf

The shelf is narrow: the syllabus figure for its capacity is 5 ± 2 separate items (the classic experiments put it at seven plus or minus two; the modern syllabus value is the more conservative one, and it is the exam answer). Exceed the limit and items are lost or transposed — the design constraint behind checklist length and RT message structure, and the reason a complicated departure clearance gets written down rather than trusted to the shelf. Working memory is also highly sensitive to interruption: a frequency received and then trodden on by an interruption is simply gone, and the pilot asks for it again. What lives on the shelf in flight is exactly that kind of cargo: frequencies, headings and altitudes before they are set, pressure settings, short ATC instructions, checklist responses awaiting execution.

Capacity can be stretched three ways. Chunking groups related items into one: a twelve-digit number becomes three chunks of four, and it works best where a mental model makes the material familiar. Association ties each item to vivid imagery. And mnemonics pack a whole drill into a word or phrase — HASELL and FREDA are working-memory prosthetics every pilot already uses.

Long-term memory and its three compartments

The long-term store divides in three, and the syllabus wants examples. Semantic memory holds general knowledge and language in a dense web of associations: what a stall is, what words mean. Once in, it is essentially never lost; failures are failures of finding, not of keeping, and it is the most accurate compartment. Episodic memory holds the events of a life: a particular flight, a particular meeting. Procedural memory holds the motor programmes — and skills are kept primarily in long-term memory, which is the syllabus point in one line. The tennis version: the rules are semantic, who served last is episodic, how to hit the lob is procedural.

Episodic memory is the unreliable witness of the three. It is bent by expectation (people remember what should have happened; an expert witness's recollection bends toward the expert's interpretation), by suggestion (ask how fast the car came around the bend and you get one speed; ask how fast it screeched around and you get another; students recalled books in an office that held none, because offices should), and by repetition, each retelling embedding its embellishments until narrator and facts have parted. Amnesia completes the picture by dissociating the compartments: it commonly strikes episodic memory only, so the sufferer who cannot say where they live can still take the third door on the left and sit in the blue chair, because words and meanings live in semantic memory.

The syllabus asks for the common problems and their counters, and the table is short. Working memory: too small, too brief, interference-prone and interruption-fragile, so chunk, rehearse, use mnemonics, and write anything that matters down. Long-term memory: slow or failing retrieval and episodic distortion — so overlearn (train past the minimum standard, which speeds recall and armours it against stress), revise, repeat, and anchor material with association.

How pilots learn

Three basic forms of learning carry the syllabus, each with a required example:

Classical and operant conditioning — the behaviouristic approach of Pavlov's school: responses trained into the body by stimulus, consequence and repetition. The pilot-shaped example is the experienced pilot's instant, unthinking reaction to a fire warning.
Learning by insight — the cognitive approach: the material is intellectually understood and retained as knowledge, as when a pilot works out and sets up the on-board navigation equipment.
Learning by imitating (modelling) — behaviour copied from a model: the student who follows through on the controls during an instructor's approach, then flies the same approach alone shortly after.

Skill learning blends the three: observation, much practice and corrective feedback. The quality of any learning rides on the syllabus list of intrinsic motivation, good mental health, rehearsal, consciousness, vigilance, and application in practical exercises, and the memorisation techniques are the working-memory tools scaled up: mnemonics for lists and drills, and mental training, the deliberate rehearsal of a procedure in imagination, which measurably improves the real performance and costs no fuel.

Skills, motor programmes and schemas

A skill is an organised, coordinated pattern of activity (physical, social, linguistic or intellectual) and its value is anticipation: a skilled performer is planning the next actions while executing the current ones. Skills are built in Anderson's three phases: the cognitive phase, thinking consciously about every action with no spare capacity at all; the associative phase, where the separate components fuse into sequences; and the autonomous phase (also called automatic), where the whole manoeuvre runs smoothly without conscious control. Stress and lack of practice can knock a performer temporarily back a phase — the emergency that makes an experienced pilot fly consciously again.

The finished product is the motor programme: a behavioural subroutine learned by practice and repetition, stored in long-term memory, and run without conscious thought. Its cousin the mental schema is the same idea for knowledge: a stored representation of a whole category (say, everything 'aerodrome' brings with it) that activates as a cluster. Their advantage is the whole economy of the single-channel mind: the programmes fly the aircraft while the central decision maker handles the radio, which is the only reason multi-tasking exists at all. Their disadvantages mirror it: they are non-declarative (the owner cannot fully explain the skill, teaching it is hard, and thinking about it mid-execution spoils it), they are hard to modify once grooved, and (as the schema-filled office showed) a schema happily fills in details the world never supplied, remembering books that were not there.

The Rasmussen model

Rasmussen's SRK model sorts a pilot's behaviour into three levels by how it is guided, and the syllabus wants each level's risks. Skill-based behaviour runs on stored motor programmes without conscious thought. Its risks are the slips of automaticity (the routine that runs when it should not, the action captured by a more familiar habit), and they belong, tellingly, to the experienced: a novice thinking about every action cannot commit them, and they strike hardest when preoccupied, tired or relaxed by easy conditions.

Rule-based behaviour — a learned procedure, consciously initiated: preparing an unfamiliar route, flying the approach chart, running an emergency drill. It is robust, and standardisation lets each crew member monitor the other. Its risks: the error of commission (misidentify the problem and run the wrong procedure entirely), applying an inappropriate rule to a correctly identified problem, and the unjustified departure from the rules — the GPWS warning judged safe to ignore.
Knowledge-based behaviour — no procedure exists; the pilot must evaluate, reason and design a plan from knowledge and experience through the central decision maker. It is the slowest and most fragile level, and its risks are the thinking traps: incomplete or inaccurate mental models, overconfidence, lost situation awareness, confirmation bias, frequency bias (reaching for the familiar scenario), and inference in accord with wishes — deciding the problem is the one you can solve because you hope it is.

The taxonomy of what goes wrong at each level (slips, lapses, mistakes and their anatomy) is the next lesson's whole subject; here it is enough to see that each level fails in its own characteristic way, so knowing which level you are operating at is itself a defence.

Motivation

Motivation is the drive that energises and directs behaviour toward a goal (goal-directed behaviour, in the syllabus phrase), and each person carries a blend of drives (achievement, mastery, reward, competition) mixed into a personal signature. Its relationship with learning is asymmetric: learning without motivation is possible, since attention is the true essential, but learning with it is vastly better, and high performance rarely arrives without it; reinforcing success feeds it.

The syllabus dwells on the failure mode: over-motivation, and especially the extreme need to achieve. The commander determined to land at the destination regardless of the weather is the canonical case: achievement motivation promoted above safety. Obstruct such a drive and the aggressive impulse follows: anger, frustration and stress released at the blockage. And any excessive motivation manufactures stress, which degrades the very performance the motivation was chasing — the curve the workload module will draw in full.