📚 Study Notes – Synonyms & Antonyms
Program: Maharashtra Police Constable (MPC)
Subject: English Language
Topic: Synonyms & Antonyms
These notes are designed to give you a deep, MIT‑level understanding of synonyms and antonyms, starting from first principles and moving through formal semantics, edge‑cases, and a challenging problem to test your mastery.
1. Overview & Motivation
| Concept | Why It Matters | Typical Use in Police Work |
|---|---|---|
| Synonyms | Expands lexical repertoire; allows nuanced reporting | Describing a crime scene with varied vocabulary |
| Antonyms | Enables comparison, contrast, and precise negation | Differentiating between suspect and accomplice |
| Semantic Precision | Essential to avoid mis‑interpretation | Clarity in statements, affidavits, or interrogations |
2. From First Principles
2.1 What is a Word?
- Lexeme: Abstract unit of meaning (e.g., run, sprint).
- Token: Concrete instance in text (e.g., the word run in a report).
- Sense: The specific meaning assigned by a particular usage.
2.2 Defining Synonyms
Formal Definition
Two lexemes A and B are synonyms if for any context C, the substitution of A for B preserves the truth‑value of the proposition expressed, i.e.,
[ \forall C,, \text{Proposition}(C, A) \leftrightarrow \text{Proposition}(C, B) ]
- Partial Synonymy: The above holds only for a subset of contexts.
- Near‑Synonymy: Very close semantics but slight differences in connotation, register, or collocation.
2.3 Defining Antonyms
Formal Definition
Two lexemes A and B are antonyms if for any context C, the substitution of A for B flips the truth‑value of the proposition, i.e.,
[ \forall C,, \text{Proposition}(C, A) \leftrightarrow \lnot \text{Proposition}(C, B) ]
- Gradual Antonyms (gradable): hot ↔ cold (degree scale).
- Complementary Antonyms: alive ↔ dead (no overlap).
- Subantonyms: blonde ↔ brunette (subset of “hair color”).
3. Lexical Semantics – Theoretical Foundations
| Theory | Key Ideas | Relevance to Synonyms / Antonyms |
|---|---|---|
| Semantic Network (WSD) | Words as nodes connected by semantic relations | Helps visualize synonym clusters and antonym pairs |
| Distributional Semantics | “You shall know a word by the company it keeps” (Firth) | Quantifies similarity via vector spaces (e.g., cosine similarity) |
| Prototype Theory | Concepts represented by a prototypical instance | Explains why some synonyms are more “core” than others |
| Reconstructionist Approach | Semantics derived from usage patterns | Useful for computational extraction of synonym/antonym pairs |
4. Formal Semantics & Computational Models
4.1 Montague Grammar
- Treats natural language as a formal logic system.
- Synonymy becomes α‑equivalence of expressions:
[ \forall x,, \text{Meaning}(A, x) = \text{Meaning}(B, x) ] - Antonymy corresponds to logical negation or complement operation.
4.2 Vector Space Models
- Word Embeddings (Word2Vec, GloVe):
[ \text{cosine}(v_A, v_B) \approx 1 \implies A \text{ and } B \text{ are synonyms} ] - Antonym Detection:
[ \text{cosine}(v_A, v_{\neg B}) \approx 1 ] where ( v_{\neg B} ) is the vector for the antonym.
4.3 Probabilistic Models
- Latent Dirichlet Allocation (LDA):
Synonyms often co‑occur in the same topics; antonyms appear in mutually exclusive topics. - Bayesian Networks:
Capture conditional probabilities of synonym/antonym usage given context.
5. Edge Cases & Nuances
| Edge Case | Explanation | Example |
|---|---|---|
| Polysemy | Same form, multiple senses | lead (metal) vs lead (to guide) |
| Homonymy | Different origins, same form | bass (fish) vs bass (low sound) |
| Register | Formal vs informal | purchase vs buy |
| Collocations | Words that rarely appear together | make + mistake (not do + mistake) |
| Idiomaticity | Meaning not deducible from parts | kick the bucket (die) |
| Semantic Drift | Meaning change over time | broadcast (radio) → broadcast (social media) |
| Partial Synonymy | Synonym in one context, not another | assist vs help (legal context) |
| Complementarity | No overlap between antonyms | alive ↔ dead |
6. Practical Examples (Police‑Specific)
| Synonym Pair | Usage Context | Antonym Pair | Usage Context | |--------------|---------------|--------------|