Advanced Study Notes on Analogy
Reasoning Ability – Program: Haryana Forest Guard
These notes are written for students who already have a solid foundation in mathematics, logic, and cognitive science.
We will start from first principles, build a formal theory of analogy, explore its applications in reasoning tests, and finish with a challenging problem that tests both understanding and creativity.
1. What is Analogy?
| Aspect | Definition |
|---|---|
| Intuitive | “Seeing the same pattern in two different situations.” |
| Philosophical | A relational inference: from known facts about A to new facts about B by preserving a set of relations. |
| Cognitive | A mental transformation that maps the structure of one domain onto another. |
Analogy = Structure‑preserving mapping between two sets of elements (the source and the target).
2. Formal Foundations
2.1 Set‑theoretic Representation
Let
( S = {s_1, s_2, …, s_n} ) be the source domain
( T = {t_1, t_2, …, t_n} ) be the target domain
Define a bijection ( f: S \rightarrow T ).
For a collection of relations ( R = {R_1, R_2, …, R_k} ) defined on (S), an analogy holds iff for every relation (R_i),
[ \forall x,y \in S,; R_i(x,y) \iff R_i(f(x), f(y)) ]
This is the structure‑preserving condition.
2.2 Relational Algebra
If ( R_i ) are binary relations, we can write:
[ R_i = {(x,y) \mid P_i(x,y)} ]
The analogy requirement becomes:
[ \forall (x,y) \in R_i,; (f(x), f(y)) \in R_i ] and conversely.
2.3 Category‑theoretic View
Let ( \mathcal{C} ) be the category whose objects are domains and whose morphisms are structure‑preserving maps.
An analogy is simply a functor ( F: \mathcal{C}_S \rightarrow \mathcal{C}_T ) that preserves all relevant morphisms.
3. Cognitive Aspects of Analogy
| Component | Description |
|---|---|
| Feature Extraction | Identifying salient properties of source and target. |
| Mapping | Establishing correspondences between features. |
| Transfer | Using the mapping to predict new relations in the target. |
| Evaluation | Checking consistency and plausibility. |
3.1 Levels of Abstraction
| Level | Example |
|---|---|
| 1st | Concrete objects (e.g., map ↔ brain). |
| 2nd | Functional roles (e.g., engine ↔ heart). |
| 3rd | Structural patterns (e.g., DNA ↔ computer code). |
3.2 Mechanisms
- Structural Alignment – the algorithm of Gentner (1978) that aligns substructures.
- Pattern‑Matching Networks – neural models that learn analogical mapping via attention.
- Rule‑Based Systems – e.g., Cognitive-Computing systems that encode analogies as production rules.
4. Types of Analogical Reasoning
| Type | Definition | Typical Test Question |
|---|---|---|
| Conceptual | Mapping concepts to concepts. | “Which pair best completes the analogy: Tree : Forest :: Cloud :?” |
| Relational | Mapping relations to relations. | “If A is to B as C is to D, what is D?” |
| Structural | Mapping structures (e.g., nested sets). | “Which figure is analogous to the following diagram?” |
| Causal | Mapping causal chains. | “What would happen if X is removed from the system?” |
| Functional | Mapping functions or processes. | “Which of the following is a function of X?” |
5. Edge Cases & Pitfalls
-
Over‑extension – assuming a mapping where one relation is weakly preserved.
Example: Fire ↔ Heat (both burn) but mapping fire to water is wrong. -
Context‑dependency – mapping may change with additional constraints.
Example: In a legal context, law ↔ rule. -
Multiple Mappings – a source may map to several targets with different relations.
Solution: Use partial mappings and evaluate each separately. -
Non‑bijective Mappings – many-to-one or one-to-many correspondences.
Solution: Allow surjective or injective functions and adjust the structure‑preservation condition accordingly. -
Cultural Bias – analogies may rely on shared cultural knowledge.
Solution: Use universal domains (mathematical, physical) for objective testing.
6. Advanced Theorems
Theorem 1 (Isomorphism Invariance)
If two domains (S) and (T) are isomorphic (i.e., there exists a bijection preserving all relations), then any property expressed solely in terms of these relations is identical in both domains.
Proof Sketch
Let (f: S \rightarrow T) be an isomorphism.
For any first‑order property (P) defined on relations,
(P(S)) holds iff (P(T)) holds because (f) maps each tuple preserving truth values.
Theorem 2 (Minimal Mapping Uniqueness)
*Given a finite set of relations (R), there exists a unique minimal mapping (in