Human Memory

2.1 Core concepts

Memory is one of the most central topics in cognitive psychology — a system that allows information to be retained and recalled over a period of time. Memory comprises three independent but interrelated stages — encoding, storage and retrieval (NCERT §Nature of Memory, p. 96). Encoding is the first process by which an external stimulus (an event, object, sound or image) is registered by the sensory system and converted into a form that the memory system can use; storage is the process through which encoded information is retained over time so that it remains available for future access; and retrieval is the process by which stored information is brought back into conscious awareness to perform a cognitive task such as recall, recognition or problem-solving (NCERT §Nature of Memory, pp. 96-97). Without successful encoding there can be no storage, and without retrieval the contents of storage remain functionally useless — students should remember the strict ordering encode → store → retrieve.

The first systematic experimental work on memory was undertaken by Hermann Ebbinghaus (1885), who used lists of meaningless CVC (consonant-vowel-consonant) nonsense syllables so that prior associations would not contaminate the results; his data produced the now-famous Curve of Forgetting showing that forgetting is steepest in the first hour after learning and then levels off (NCERT §Introduction, p. 96; §Nature and Causes of Forgetting, p. 103). Building on this experimental foundation, Richard Atkinson and Richard Shiffrin (1968) proposed the Stage Model of memory, drawing an analogy between human memory and a computer with both temporary RAM-like and permanent disk-like stores (NCERT §Information Processing Approach, p. 97). According to the Stage Model, information passes sequentially through three memory stores. Sensory Memory has very large capacity but holds information for less than one second; it contains modality-specific registers called the iconic register (for visual stimuli) and the echoic register (for auditory stimuli) that store an exact replica of the stimulus (NCERT §Sensory Memory, p. 98). Short-term Memory (STM) can hold only a small amount of information for approximately 30 seconds and is primarily encoded acoustically (in terms of sound); unless rehearsed continuously it is lost (NCERT §Short-term Memory, p. 98). Long-term Memory (LTM) has practically unlimited capacity, is encoded semantically (in terms of meaning), and information once entered is theoretically never truly forgotten — apparent forgetting from LTM is in fact a failure of retrieval rather than a loss of the trace (NCERT §Long-term Memory, p. 98).

Several control processes govern the flow of information between these stores. Selective attention transfers information from the sensory register into STM by allowing only the attended stimuli to enter; maintenance rehearsal keeps information active in STM through silent or vocal repetition; chunking expands STM's famous 7 ± 2 capacity by grouping smaller units into larger meaningful chunks; and elaborative rehearsal transfers information from STM into LTM by linking new content to already-stored associations so that the new information acquires deeper meaning (NCERT pp. 98-99). NCERT emphasises that elaborative rehearsal (not mere repetition) is the route into LTM.

There are major alternatives and modifications to the Stage Model. Alan Baddeley (1986) reconceived STM as Working Memory — an active mental workbench rather than a passive store — with three components: the phonological loop (which stores sound-based information and decays in roughly 2 seconds), the visuospatial sketchpad (which stores visual and spatial information) and the central executive which organises information flowing between the loop, the sketchpad and LTM and allocates attentional resources (NCERT Box 6.1, p. 98). Strong evidence against a strict serial Stage Model comes from the KF case reported by Shallice and Warrington (1970): damage to the left cerebral hemisphere left this patient's LTM intact but severely impaired his STM — a result that contradicts the strict Atkinson-Shiffrin claim that information must pass through STM to reach LTM (NCERT §Levels of Processing, p. 100). Building on similar critiques, Fergus Craik and Robert Lockhart (1972) proposed the Levels of Processing view, according to which retention depends less on the store information passes through and more on the depth at which it is processed. They identified three levels — structural (shallowest, e.g., noticing the shape of letters), phonetic (intermediate, processing the sound) and semantic (deepest, processing the meaning) — with semantic encoding producing the most durable memory traces (NCERT §Levels of Processing, p. 100).

LTM itself is not a single store. Declarative memory holds facts, names, dates and other content that can be described verbally, while procedural memory stores skills such as riding a bicycle or tying a knot that are difficult to put into words. Endel Tulving further divided declarative memory into episodic memory (autobiographical, dated, emotionally coloured memories of personal life events) and semantic memory (the storehouse of general knowledge, concepts and rules that is affect-neutral and not tied to a particular time) (NCERT §Types of Long-term Memory, pp. 100-102). Box 6.2 introduces three additional varieties: flashbulb memories — vivid, photograph-like memories of arousing or surprising events; autobiographical memory including childhood amnesia (the typical inability to recall events from the first four to five years of life); and implicit memory that influences behaviour outside conscious awareness (NCERT Box 6.2, p. 101). Each of these is assessed using a different experimental task — free recall and recognition for facts and episodes, the sentence verification task for semantic memory and priming for implicit memory (NCERT Box 6.3, p. 103).

The remainder explains why we forget. Ebbinghaus's curve shows forgetting is maximal in the first nine hours, especially the first hour (NCERT Fig. 6.2, p. 103). Trace Decay (disuse) theory holds that memory traces simply fade with non-use but is refuted by the finding that learners who sleep after encoding forget less than those who stay awake (NCERT p. 104). Interference theory distinguishes proactive interference (old learning blocks recall of new) from retroactive interference (new learning blocks recall of old) (NCERT pp. 104-105; Table 6.1). Tulving's retrieval-failure theory locates forgetting in the absence of appropriate retrieval cues (NCERT p. 105). Box 6.4 covers Sigmund Freud's notion of repressed memories — traumatic experiences pushed into the unconscious — and the fugue state in which severe repression leads the person to assume a new identity (NCERT Box 6.4, p. 105). Memory can be improved through mnemonics: image-based strategies include the keyword method (learning a foreign word via a similar-sounding English keyword) and the method of loci (placing items at familiar locations in sequence) (NCERT pp. 105-106), while organisation-based strategies include chunking, the first-letter technique (e.g., VIBGYOR) and Thomas and Robinson's PQRST study method (Preview, Question, Read, Self-recitation, Test) (NCERT pp. 106-107).

2.2 Definitions to memorise

2.3 Diagrams / processes to remember

• Fig. 6.1 – Stage Model of Memory (p. 97): Information → Sensory Memory (iconic/echoic, capacity large, duration < 1 s) → (Attention) → Short-term Memory (small capacity, < 30 s) → (Elaborative Rehearsal) → Long-term Memory (unlimited capacity, up to lifetime). The arrows show the strict serial passage that the Stage Model claims is necessary; the KF case (p. 100) is the famous counter-example.
• Fig. 6.2 – Ebbinghaus's Curve of Forgetting (p. 103): A steeply descending curve in the first hour (with marked points at 20 minutes, 1 hour and 8.8 hours), followed by a long, slow plateau — visual proof that most forgetting happens immediately after learning and that very little additional material is lost once the first day has passed.
• Table 6.1 – Designs for Retroactive vs Proactive Interference (p. 104): Retroactive — Experimental: Learns A, Learns B, Recalls A; Control: Learns A, Rests, Recalls A. Proactive — Experimental: Learns A, Learns B, Recalls B; Control: Rests, Learns B, Recalls B. Remember: in retroactive designs A is recalled (new B hurts old A); in proactive designs B is recalled (old A hurts new B).
• Working Memory model (Box 6.1, p. 98): Baddeley's three-component diagram — Central Executive at the top organising and allocating attention, with two slave systems below: the Phonological Loop (sounds, ~2 s) on one side and the Visuospatial Sketchpad (visual/spatial) on the other, both interacting with LTM via the executive.

2.4 Common confusions / NTA trap points

• STM is encoded acoustically, LTM semantically — but later research shows STM can also encode semantically and LTM acoustically (p. 99). NTA likes to flip this.
• Maintenance vs Elaborative rehearsal: maintenance only repeats (keeps info in STM); elaborative links to existing LTM (transfers info to LTM). Don't swap.
• Proactive (forward) vs Retroactive (backward) interference: proactive = old hurts new; retroactive = new hurts old. Easy to invert under pressure.
• Episodic vs Semantic: episodic = personal events, emotional, dated; semantic = general facts, affect-neutral, not dated.
• Forgetting in LTM is retrieval failure, not loss — information in LTM is "never forgotten" (p. 98); a trap when paired with trace-decay statements.
• Levels of Processing was proposed by Craik and Lockhart (1972), not Atkinson-Shiffrin or Baddeley.
• KF case (Shallice & Warrington, 1970) is the counter-evidence to the Stage Model — STM impaired but LTM intact.
• Iconic = visual; Echoic = auditory — students often reverse these because "echo" is auditory but the modality names are easy to swap under pressure.
• Capacity numbers: Sensory = very large, STM = 7 ± 2 chunks, LTM = unlimited; duration: Sensory < 1 s, STM ~ 30 s, LTM up to lifetime.
• Method of loci ≠ keyword method: loci uses imagined spatial locations; keyword uses sound-similarity to a known word.
• PQRST = Preview, Question, Read, Self-recitation, Test (Thomas & Robinson); do not confuse with the SQ3R study scheme.
• Freud's repressed memories lead to the fugue state when extreme — not to amnesia from physical brain damage.

2.5 Thinkers and theories at a glance