10 Intelligence and Aptitude · 10.1 Intelligence and aptitude
10.1.1 Theories of intelligence, emotional and artificial intelligence
Intelligence is the global capacity to understand the world, think rationally and use resources well when faced with challenges; psychometric theories describe it as one or more factors found by factor analysis (Spearman's g, Thurstone's 7 PMAs, Guilford's cube, Cattell's Gf–Gc), information-processing theories describe the processes an intelligent person uses (Sternberg, PASS), and the idea has widened to multiple, emotional and artificial intelligence.
- APPSC Paper I
- Standard level
- 16 min read
- Free
Core facts
- Definitions (Binet, Terman, Stern, Wechsler, Thorndike's three kinds) and the B.Ed.-style theory list: see EDU-ID-01. This note adds the BA-level points the Paper II asks.
- NCERT's working definition: the global capacity to understand the world, think rationally and use available resources effectively when faced with challenges.
- Indicators of intelligence: judging well, reasoning well, understanding, solving problems, adapting to the environment, learning from experience. "Listening well" is not an indicator (2023 key).
- Approaches (three are named in Indian BA texts, and the 2023 key accepted all three):
- Psychometric: intelligence as an aggregate of abilities, shown as one score (IQ); found by factor analysis of test scores.
- Structural: the psychometric theories seen as maps of the structure of abilities (g at the top, group and specific factors below, or Guilford's cube).
- Information-processing: the processes people use in reasoning and problem solving; its focus is how an intelligent person acts (2023 A–R item).
Theorist–theory table
| Theorist | Theory / model | Year | Core of it |
|---|---|---|---|
| Alfred Binet | Uni-factor | 1905 | one general ability ("good sense") |
| Charles Spearman | Two-factor | 1904 | g (general) + s (specific) |
| E.L. Thorndike | Multi-factor | 1920s | many separate bonds; abstract, social, mechanical |
| Godfrey Thomson | Sampling | 1916 | each test samples many small abilities |
| L.L. Thurstone | Primary mental abilities | 1938 | 7 group factors, no single g |
| P.E. Vernon (after Burt) | Hierarchical | 1950 | g → v:ed, k:m → minor → specific |
| Arthur Jensen | Hierarchical, two levels | 1969 | Level I associative, Level II cognitive |
| J.P. Guilford | Structure of intellect | 1967 | operations × contents × products = 120 |
| R.B. Cattell (with J. Horn) | Fluid and crystallised | 1940s–60s | Gf reasoning, Gc knowledge |
| Howard Gardner | Multiple intelligences | 1983 | 7, later 8 independent intelligences |
| Robert Sternberg | Triarchic | 1985 | componential, experiential, contextual |
| J.P. Das, Naglieri, Kirby | PASS | 1994 | planning, attention, simultaneous, successive |
- Thurstone's 7 PMAs: verbal comprehension, word fluency, number, space, memory, perceptual speed, reasoning.
- Gardner's eight, each in one line:
- Linguistic: using language fluently and flexibly.
- Logical-mathematical: abstract reasoning, symbols, number, logic.
- Spatial: forming, using and transforming mental images (2023 key).
- Musical: sensitivity to pitch, rhythm and tone.
- Bodily-kinaesthetic: using the whole body or hands skilfully.
- Interpersonal: understanding other people's motives and feelings.
- Intrapersonal: knowing one's own feelings, strengths and identity.
- Naturalist: recognising plants, animals and patterns in nature (added later, 1990s).
- Sternberg's three: componential = analytical, experiential = creative, contextual = practical ("street smart").
- PASS rests on A.R. Luria's three functional units of the brain; tested by the Cognitive Assessment System (CAS).
- Emotional intelligence (EI): Salovey and Mayer (1990) first defined it; Goleman (1995) popularised it with five components. Models, branches and tests: see MAB-EI-01.
- Artificial intelligence (AI): term coined by John McCarthy for the Dartmouth conference, 1956. Alan Turing's test of machine intelligence: 1950.
- Connectionism = explaining the mind by artificial neural networks of simple connected units; parallel distributed processing (PDP) is its main model (Rumelhart and McClelland, 1986) (2023 key).
Explained
1. Psychometric vs information-processing, side by side
| Psychometric | Information-processing | |
|---|---|---|
| Asks | how much, and which abilities | how a person thinks |
| Method | factor analysis of test scores | analysing steps of a task |
| Output | IQ, factor scores | process profile (e.g. PASS) |
| Theories | Spearman, Thurstone, Guilford, Cattell | Sternberg, Das (PASS) |
Gardner's theory is usually taught with the modern "multiple abilities" views; it rests on evidence from brain damage, prodigies and savants, not on factor analysis.
2. From one factor to many. Spearman noticed that people who do well on one mental test tend to do well on others (positive correlations) and called the common part g; the leftover skill for each test is s. Thurstone's own factor analysis of many tests gave seven fairly separate factors instead of one g, and he built the PMA test on them. Thomson said a test score samples many small independent abilities, so correlations arise from overlap, not from one g. The hierarchical models reconcile the two: Vernon put g at the top, then two major group factors, v:ed (verbal–educational) and k:m (practical–mechanical–spatial), then minor group factors and specific factors.
3. Jensen's two levels (the 2023 match item called him "Jansen")
- Level I, associative learning: the output is much like the input; rote memory, digit span.
- Level II, cognitive competence: the input is transformed; reasoning, abstraction, problem solving.
Jensen argued all groups are alike in Level I, and that Level II tracks g; his claims about group differences were strongly disputed.
4. Guilford's cube. Three faces: operations (what the mind does: cognition, memory, divergent production, convergent production, evaluation), contents (what it works on: figural, symbolic, semantic, behavioural) and products (the form of the result: units, classes, relations, systems, transformations, implications). 5 × 4 × 6 = 120 abilities (1967). Splitting figural content into visual and auditory gives 5 × 5 × 6 = 150, the figure NCERT gives; a later split of memory gives 180. Full table: EDU-ID-01. Divergent production is the root of creativity (see EDU-ID-03).
5. Cattell's fluid and crystallised intelligence
- Fluid (Gf): solving new problems, seeing relations, culture-light; tied to the nervous system; rises to the 20s, then slowly falls.
- Crystallised (Gc): vocabulary, facts and skills from schooling and culture; keeps growing into late adulthood.
Culture-fair tests (Raven's, Cattell's own Culture Fair Test) try to tap Gf. The later Cattell–Horn–Carroll (CHC) model joins this with Carroll's three-stratum model (1993).
6. Sternberg's triarchic theory, in detail
| Sub-theory | Intelligence | What it does | Example |
|---|---|---|---|
| Componential | Analytical | analyse, compare, judge | topper in exams |
| Experiential | Creative | deal with new tasks; make old ones automatic | inventor, artist |
| Contextual | Practical | adapt to, shape or select the environment | successful trader |
Components are of three kinds: metacomponents (plan, monitor, evaluate), performance components (carry out the plan) and knowledge-acquisition components (learn new things). Sternberg later called the balance of all three successful intelligence.
7. PASS model
| Process | What it does | Luria's unit / brain area |
|---|---|---|
| Attention-arousal | stay alert, attend to the right stimuli | unit 1: brain stem, reticular system |
| Simultaneous | combine parts into a whole (patterns, relations) | unit 2: occipital-parietal |
| Successive | handle items in serial order (sequences, speech) | unit 2: fronto-temporal |
| Planning | choose, use and check strategies | unit 3: frontal lobes |
Too much or too little arousal hurts performance. Planning is the central, "executive" process. Das, born in Odisha, did most of this work at the University of Alberta, Canada; the CAS is used to diagnose learning difficulties.
8. Heredity and environment. Twin and adoption studies (as reported by NCERT): IQ correlation is about 0.90 for identical twins reared together, 0.72 for identical twins reared apart, about 0.60 for fraternal twins reared together, 0.50 for siblings reared together and 0.25 for siblings reared apart. Adopted children grow closer to their biological parents' IQ with age. Yet deprivation lowers IQ and enrichment (nutrition, schooling, stimulating homes) raises it: heredity sets a range, environment decides where in it a person falls.
9. Culture and intelligence. Western tests stress speed, analysis and abstract reasoning ("technological intelligence"). NCERT describes an Indian integral view (buddhi) that joins knowledge with self-reflection and social ties, under four competencies: cognitive capacity, social competence, emotional competence, entrepreneurial competence. Sternberg calls intelligence contextual: what counts as intelligent depends on the culture.
10. Emotional intelligence (short; full note MAB-EI-01). EI = the ability to perceive, understand, use and manage one's own and others' emotions. The emotionally intelligent person reads feelings from body language and voice, links emotions to thoughts, and controls strong emotions. Applications: counselling and probation work (managing anger, empathy with offenders and victims), leadership and teamwork at work, emotional-literacy and life-skills programmes in schools. Unlike IQ, EI can be trained.
11. Artificial intelligence: two traditions
| Symbolic AI | Connectionism | |
|---|---|---|
| Model of mind | rules acting on symbols | network of simple units |
| Knowledge stored in | explicit rules, facts | strengths (weights) of connections |
| Processing | serial, step by step | parallel, distributed |
| Examples | Logic Theorist, GPS, expert systems | perceptron, PDP models, deep learning |
- PDP: many simple units work at the same time; a memory or concept is a pattern spread across many units, not stored in one place; the network learns by adjusting weights from experience (e.g. back-propagation, 1986). Such networks recognise patterns, generalise, and degrade gracefully when damaged, like the brain.
- Milestones: McCulloch and Pitts' model neuron (1943); Turing test (1950); Dartmouth, McCarthy (1956); Newell and Simon's Logic Theorist (1956) and General Problem Solver (1957); Rosenblatt's perceptron (1958); ELIZA (Weizenbaum, 1966), a program that imitated a non-directive therapist; expert systems such as MYCIN (1970s, medical diagnosis); PDP volumes (1986); today's deep learning and large language models.
- Applications for psychology: computer simulation to test theories of thinking and memory; expert systems and computer-based (adaptive) testing; tutoring systems; screening and support chatbots in mental-health care; robots and aids for the disabled.
Common traps
- ✗ "The information-processing approach measures how much intelligence a person has." ✓ It studies how an intelligent person thinks and acts; "how much" is the psychometric approach.
- ✗ "Jensen gave the triarchic theory." ✓ Jensen = hierarchical, Level I and Level II; Sternberg = triarchic.
- ✗ "Vernon's model has primary mental abilities." ✓ Vernon = hierarchical (g, v:ed, k:m); Thurstone = PMAs.
- ✗ "Binet gave a multi-factor theory." ✓ Binet's view is uni-factor (one general ability); multi-factor = Thorndike.
- ✗ "Fluid intelligence grows with education into old age." ✓ That is crystallised; fluid peaks in early adulthood.
- ✗ "Spatial intelligence is skill with the body." ✓ Spatial = forming, using and transforming mental images; body = bodily-kinaesthetic.
- ✗ "PASS stands for perception, attention, speed, sequence." ✓ Planning, Attention-arousal, Simultaneous, Successive (J.P. Das).
- ✗ "PDP belongs to neural networking / associationism as its paradigm." ✓ The paradigm is connectionism.
- ✗ "Goleman coined emotional intelligence." ✓ Salovey and Mayer (1990); Goleman (1995) popularised it.
- ✗ "Turing coined the term artificial intelligence." ✓ John McCarthy, 1956; Turing proposed the test (1950).
Quick revision
- Approaches: psychometric (how much; factors), structural, information-processing (how one thinks and acts).
- Binet uni; Spearman 1904 g + s; Thorndike multi; Thomson sampling; Thurstone 1938 seven PMAs.
- Hierarchical: Vernon 1950 (g, v:ed, k:m); Jensen 1969 Level I associative, Level II cognitive.
- Guilford 1967: 5 operations × 4 contents × 6 products = 120 (NCERT 150).
- Cattell: fluid (new problems, falls with age) vs crystallised (knowledge, keeps rising).
- Gardner 1983: 7 + naturalist = 8; spatial = mental images.
- Sternberg 1985: analytical (componential), creative (experiential), practical (contextual).
- PASS (Das, Naglieri, Kirby; Luria): planning, attention, simultaneous, successive; CAS.
- EI: Salovey and Mayer 1990, Goleman 1995 (MAB-EI-01).
- AI: McCarthy 1956, Turing test 1950; connectionism = neural networks; PDP = Rumelhart and McClelland 1986.
Go deeper: Deeper (beyond what APPSC Paper I usually asks)
Deeper Advanced
1. Strong vs weak AI. Weak (narrow) AI: machines simulate intelligent behaviour in limited tasks (chess, translation). Strong AI: a suitably programmed machine would truly understand and have a mind. John Searle's Chinese room (1980) argues that following symbol rules is not understanding, so passing the Turing test is not enough. Psychology's interest: AI models are tested against human errors and reaction times, not only against success.
2. Why the "factor" debate never ended. Factor analysis gives different answers depending on the tests chosen and the rotation used: Spearman's method favoured one g, Thurstone's rotation favoured several factors. When Thurstone's PMAs were measured in ordinary samples they correlated with one another, which brought back a second-order g. The hierarchical models (Vernon, Carroll, CHC) are the compromise most test makers use today.
3. Critiques of Gardner and Sternberg. Critics say several of Gardner's "intelligences" are talents or personality traits, and that they are hard to measure separately. Sternberg's practical intelligence (tacit knowledge tests) predicts job success partly beyond IQ, but critics find it overlaps with g. Both theories still shaped teaching: one idea taught in several ways, and assessment beyond paper tests.
4. Associationism → connectionism. Thorndike's learning theory was also called "connectionism" (stimulus–response bonds). The 1980s connectionism of PDP is different: it uses networks of units and weights, and claims rules can emerge from learning without being programmed. Both share the old associationist idea that the mind builds knowledge from connections, which is why "associationism" is a tempting distractor.
Practise this topic. Questions on Nature and theories of intelligence; emotional and artificial intelligence, with answers explained.
All Intelligence and aptitude notes · APPSC Paper I syllabus · APPSC Paper I Mock 1 with answers · Start free
Medha AI writes its own notes and checks them against the sources listed. Medha AI is an independent preparation service. It is not affiliated with, endorsed by or connected to APPSC, TSPSC, MPSC, UPSC or any other commission or government body.