Human Memory

This is the sixth chapter of NCERT's Class 11 Psychology textbook (pages 95–108). It describes memory as encoding, storage and retrieval; the stage model (sensory, short-term and long-term memory) and its control processes; working memory; the levels-of-processing view; the types of long-term memory; methods of measuring memory; the causes of forgetting, including repression; and ways to improve memory. It has four boxes, two figures, one table, four activities, seven review questions and one project idea. (Beyond NCERT: the rationalised edition has dropped earlier material on memory as a constructive process — Bartlett's serial reproduction and the reconstruction of memories — and on how knowledge is represented, though two leftover key terms, ‘memory making’ and ‘serial reproduction’, still refer to it. Those topics appear here in a Beyond-NCERT section.)

• Builds on Sensory, Attentional and Perceptual Processes (selective attention is the first control process; Miller's 7 ± 2) and Learning (Ebbinghaus's nonsense syllables, free recall, Bousfield's clustering, and the total-time principle — verbal learning and memory are two sides of one coin). Leads into Chapter 7 (thinking — concepts and schemas stored in semantic memory).
• Pairs with live chapters: Class 11 psychology What is Psychology? (Freud's unconscious, the cognitive approach and the computer analogy) and Human Development (infant memory, ageing); Class 12 psychology Meeting Life Challenges (stress — Box 6.4's examination anxiety).
• UPSC: the 2023 Essay ‘Education is what remains after one has forgotten what one has learned in school’ is this chapter's clearest use: Ebbinghaus's rapid forgetting of rote material, deep (semantic) processing, procedural memory and skills, and semantic memory that is ‘not dated’. In GS4 and GS2 it supports points on eyewitness reliability and fair investigation, institutional memory and record-keeping, and learning that relies on understanding rather than rote (an aim of the National Education Policy 2020). No Prelims question tests this chapter.

• Memory = encoding → storage → retrieval; failure can occur at any stage. First systematic study: Ebbinghaus (1885), on himself, with nonsense syllables.
• Stage model (Atkinson and Shiffrin, 1968): sensory (large, <1 s; iconic, echoic) → attention → STM (small, 7 ± 2, ≤30 s, acoustic) → elaborative rehearsal → LTM (unlimited, semantic).
• Control processes: attention, maintenance rehearsal, chunking (1947-1949-2004), elaborative rehearsal. Working memory (Baddeley): phonological loop, visuospatial sketchpad, central executive.
• Levels of processing (Craik and Lockhart, 1972): structural → phonetic → semantic (deepest, best retained). Patient KF (Shallice and Warrington, 1970) challenged the stage model.
• LTM: declarative (facts) vs procedural (skills); declarative = episodic vs semantic (Tulving). Box 6.2: flashbulb, autobiographical (childhood amnesia), implicit memory; Box 6.3: recall, recognition, sentence verification, priming.
• Forgetting: Ebbinghaus curve (sharp early drop); trace decay, interference (proactive, retroactive), retrieval failure (Tulving); repression (Freud, Box 6.4). Enhancing: keyword, loci, chunking, first letter (VIBGYOR), deep processing, PQRST (Thomas and Robinson).

Memory means retaining and recalling information over a period of time, as the task requires — an unfamiliar telephone number for a few seconds until you dial it, or the methods of addition and subtraction for many years. It is conceptualised as a process of three independent though interrelated stages, through which all information must pass. Encoding records and registers information for the first time so the memory system can use it: a stimulus produces neural impulses, received in different brain areas, where it is given some meaning and represented for further processing. Storage retains and holds encoded information over time. Retrieval brings stored information back to awareness for tasks such as problem solving and decision-making. Memory can fail at any stage — information may not have been encoded properly, storage may have been weak, or it may not be accessible when needed.

Memory was once pictured as a vast storehouse. With the computer came a new analogy: like a computer with temporary memory (RAM) and permanent memory (a hard disk), people register, store and manipulate information according to the task — retrieving arithmetic operations to solve a sum, for instance. This information-processing approach led to the first model of memory, the stage model of Atkinson and Shiffrin (1968), with three memory systems.

How does information travel between stores? Atkinson and Shiffrin proposed control processes that monitor the flow. Not everything the senses receive is registered — imagine the pressure if it were. Selective attention is the first control process: only attended information enters STM; unattended impressions fade quickly. In STM, maintenance rehearsal keeps information for as long as needed through silent or vocal repetition; when repetition stops, the information is lost. Chunking expands STM's capacity of 7 ± 2: the twelve digits 194719492004 exceed it, but as 1947, 1949 and 2004 — the years of independence, the adoption of the Constitution and the tsunami that hit India and South-East Asia — they are easy.

Information enters LTM through elaborative rehearsal, which connects the ‘to be retained’ information to what is already in LTM. The meaning of ‘humanity’ is easier to remember if ‘compassion’, ‘truth’ and ‘benevolence’ are already in place; the more associations created around new information, the more permanent it becomes. Elaboration means analysing information for the associations it arouses and organising it in as many ways as possible — in a logical framework, by linking it to similar memories, or by creating a mental image.

Mixed evidence. Experiments testing the stage model gave mixed results. Some show STM and LTM to be separate stores; others question it — for instance, information once thought to be coded only acoustically in STM and semantically in LTM has been shown to be coded semantically in STM and acoustically in LTM too.

Psychologists have since suggested that STM is not a single store. NCERT attributes the multi-component view to Baddeley (1986): STM is not a passive storehouse but a work bench where material is constantly handled, manipulated and transformed during cognitive tasks — working memory. Its first component, the phonological loop, holds a limited number of sounds, which decay within about 2 seconds unless rehearsed. The visuospatial sketchpad holds visual and spatial information, also with limited capacity. The central executive organises information from the loop, the sketchpad and long-term memory and, like a true executive, allocates attention and monitors, plans and controls behaviour.

The case of KF. Shallice and Warrington (1970) described a man, KF, whose accident damaged part of the left cerebral hemisphere: his LTM was intact but his STM was seriously impaired. If information reaches LTM only through STM, how could his LTM be normal? Other studies found memory processes to be similar whether information is kept for seconds or years, and that memory can be understood without positing separate stores. This led to a second model — levels of processing.

The levels of processing view, proposed by Craik and Lockhart in 1972, holds that how new information is perceived, analysed and understood determines how well it is retained. Information can be analysed at more than one level. At the shallowest, structural level one attends to physical features — the shape of the letters in ‘cat’, capital or small, or the colour of the ink. At an intermediate, phonetic level one attends to the sounds of the letters, so the shapes become a word. Both produce memory that is fragile and decays quickly. At the deepest, semantic level one analyses meaning — a cat is a furry, four-legged, tailed mammal; one may picture a cat and connect it with one's experience. Semantic encoding gives memory that resists forgetting considerably. The implication for study: elaborate the meaning of a lesson and relate it to other facts, concepts and your life, rather than depending on rote memorisation.

Just as STM is now seen as several components, LTM is seen as several types. One major classification is declarative versus procedural (sometimes nondeclarative) memory. Declarative memory holds facts, names and dates — a rickshaw has three wheels; India became independent on 15 August 1947; a frog is an amphibian; you and your friend share a name. Procedural memory holds procedures for tasks and skills — riding a bicycle, making tea, playing basketball. Declarative contents can be put into words; procedural contents cannot easily be described: you can explain how cricket is played but not how you ride a bicycle.

Tulving divided declarative memory into episodic and semantic. Episodic memory holds the biographical details of our lives — how you felt when you stood first, what your friend said when you broke a promise — and so is generally emotional; yet we do not remember every event, and painful experiences are not remembered in as much detail as pleasant ones. Semantic memory holds general knowledge — concepts, ideas, rules of logic: the meaning of ‘non-violence’, that 2 + 6 = 8, that New Delhi's STD code is 011, that ‘elaphant’ is misspelt. Unlike episodic memory it is not dated — you cannot say when you learnt that Bangalore is Karnataka's capital — and, NCERT says, being affect-neutral it is not susceptible to forgetting.

Why do we forget? Because information in LTM is lost, or was not memorised well enough, or was encoded wrongly, or was distorted or misplaced in storage? The first systematic attempt to answer was Ebbinghaus's. He memorised lists of nonsense syllables (consonant-vowel-consonant trigrams such as NOK or SEP) and measured how many trials he needed to relearn the same list after various intervals. Forgetting followed a regular pattern: it was fastest in the first nine hours, especially the first hour, then slowed, and little more was lost even after many days. His experiments were simple, but it is now accepted almost unanimously that there is always a sharp drop in memory followed by a very gradual decline.

Trace decay (or disuse) theory, the earliest, holds that memory changes the central nervous system — memory traces — which fade when unused. It proved inadequate: if disuse caused forgetting, people who sleep after memorising (when traces cannot be used) should forget more than those who stay awake; in fact those who stay awake forget more. So decay gave way to interference theory, perhaps the most influential: learning forms associations that remain intact and independent, but at retrieval different sets compete. Learn list A, then list B, then try to recall A: intruding B items show interference. Proactive (forward) interference: earlier learning interferes with recall of later learning — knowing English makes French harder to learn. Retroactive (backward) interference: later learning interferes with recall of earlier learning — you cannot recall the English equivalents of the French words you are memorising now.

Retrieval failure. Forgetting can also occur because, at recall, retrieval cues — aids to recovering stored information — are absent or inappropriate. Tulving and his associates showed this experimentally. Memorise words such as hut, wasp, cottage, gold, bronze and ant drawn from six categories (places of living, insects, metals …); asked later, you may recall only a few, but given the category names on a second attempt, recall is near total. The categories act as retrieval cues; so does the physical context in which one learnt.

Memory as reconstruction (Beyond NCERT). The chapter's key-terms list keeps ‘serial reproduction’ and ‘memory making’ from a section the rationalised edition dropped. Frederic Bartlett, in Remembering (1932), had English participants retell a North American folk tale, ‘The War of the Ghosts’, repeatedly or pass it from person to person (serial reproduction). The story grew shorter, unfamiliar details dropped out or were rationalised, and it drifted towards the participants' own cultural schemas. Bartlett concluded that remembering is ‘reconstructive’, not reproductive. Decades later Elizabeth Loftus showed how later information reshapes memory: in Loftus and John Palmer's study (1974), people who saw a film of a car collision estimated higher speeds when asked how fast the cars were going when they ‘smashed’ rather than ‘hit’ each other, and a week later were more likely to report broken glass that the film never showed — the misinformation effect.

Mnemonics (pronounced ‘ni-mo-nicks’) are strategies for improving memory. Some use images; others impose organisation. Mnemonics using images create vivid, interacting images of the material. In the keyword method, used for foreign-language words, an English word that sounds like the foreign word becomes the keyword, and keyword and meaning are imagined interacting: for the Spanish ‘pato’ (duck), take ‘pot’ and imagine a duck in a pot of water — NCERT calls this much superior to rote memorisation. In the method of loci, items are placed as visual images in a familiar physical space, in sequence — useful for serial order: picture bread and eggs in your kitchen, tomatoes on a table and soap in the bathroom, then in the market take a mental walk from kitchen to bathroom.

Mnemonics using organisation impose order on material; the framework makes retrieval easy. Chunking combines small units into larger chunks by finding an organising principle — a control process for STM that also improves memory. The first letter technique takes the first letter of each item to form a word or sentence: VIBGYOR for violet, indigo, blue, green, yellow, orange and red. NCERT then cautions that mnemonics are simplistic and underestimate the complexity of memory tasks, and gives a more comprehensive approach based on memory processes: (a) engage in deep-level processing — ask questions, consider meaning, relate new information to what you know (Craik and Lockhart); (b) minimise interference — do not study similar subjects one after another, and if you must, distribute practice with rest periods; (c) give yourself enough retrieval cues — find cues in the material and link parts of it to them. Thomas and Robinson developed PQRST. Finally, no method brings overnight improvement: health, interest and motivation, familiarity with the subject and fitting the strategy to the task all matter.

The chapter has four activities and one project idea. Short outlines follow.

The seven review questions at the end of the chapter (page 108). Try each first, then open the outline.

Numbers

One-liners

• Memory is encoding, storage and retrieval; failure can occur at any stage.
• The stage model came from the computer analogy: sensory, short-term and long-term stores.
• Attention moves information into STM; elaborative rehearsal moves it into LTM.
• Maintenance rehearsal keeps; elaborative rehearsal transfers.
• Chunking expands the 7 ± 2 capacity of STM using meaning from LTM.
• Working memory is a work bench, not a waiting room.
• The deeper — more semantic — the processing, the longer the retention.
• Declarative memory can be put in words; procedural memory cannot easily be.
• Episodic memory is dated and emotional; semantic memory is undated general knowledge.
• Most forgetting happens soon after learning; then the decline is gradual.
• Sleep after learning reduces forgetting — evidence against trace decay.
• Interference is competition at retrieval; retrieval failure is a missing cue.
• Mnemonics work by images or by organisation.

Traps to avoid

• Sensory memory has a LARGE capacity but lasts under a second.
• STM coding is mainly acoustic; LTM coding semantic — but later evidence found both in both.
• Selective attention is in Chapter 4, not ‘Chapter 5’ as NCERT says.
• Working memory was first proposed by Baddeley and Hitch (1974); NCERT gives Baddeley (1986).
• KF had impaired STM and intact LTM — not the reverse.
• Structural is the shallowest level, semantic the deepest; phonetic is intermediate.
• Semantic memory is not dated; episodic memory is.
• Procedural memory is also called nondeclarative.
• Ebbinghaus measured savings in relearning, not just recall.
• Those who sleep after learning forget less, not more.
• Proactive = earlier learning interferes with later; retroactive = later with earlier.
• NCERT's ‘Box 7.2’ and ‘Activity 7.3’ references are stale; read Box 6.2.
• A fugue is not ‘two personalities’; that is dissociative identity disorder.
• Flashbulb memories are vivid and confident, but research finds them often inaccurate.
• PQRST is Thomas and Robinson's; the ‘S’ is self-recitation.

Prelims — what UPSC asks

• No Prelims question in the archive tests this chapter. ‘Memory’ appears only in other senses — computer memory (1988, 2000), memory loss as an AIDS symptom (2003) and UNESCO's Memory of the World Register (2008, Rig-Veda manuscripts) — none of which this chapter teaches.
• If psychology appears, expect pairs: iconic/echoic, maintenance/elaborative rehearsal, proactive/retroactive interference, episodic/semantic, and names with models (Atkinson–Shiffrin, Craik–Lockhart, Baddeley, Tulving, Ebbinghaus).

Mains — where it appears and how to answer

• Essay: ‘Education is what remains after one has forgotten what one has learned in school’ (2023). Build on the forgetting curve (rote detail fades within days), levels of processing (understanding persists), procedural and semantic memory (skills and concepts remain, undated), and schemas (frameworks for rebuilding knowledge); add NEP 2020's stated shift from rote learning to conceptual understanding, and the counter-view that facts still matter as material for thought.
• GS4 themes this chapter grounds (no direct question linked): honesty versus accuracy — witnesses and officers can be sincere yet wrong; leading questions in inquiries; contemporaneous records and file notings as safeguards against reconstructive memory; institutional memory when officers are transferred; emotional intelligence and stress (Box 6.4's examination blank).
• GS2: police investigation and eyewitness identification (fair line-ups, recording first statements), and learning-centred schooling — use the attributed tables, without a verdict on the recovered-memory debate.
• Answer frame: name the memory process (encoding, storage or retrieval) → the model or study (attributed, dated) → an Indian administrative or classroom example → the contested view in one line → a practical safeguard.

Tap an option. NCERT's own exercise questions come first; the rest test the figures and common confusions.

Open any question to attempt it in the Question bank — your attempt is saved to your progress.