Lexical Search
Also called Term-based Search
Search that matches literal words and their forms, not underlying meaning.
Think of it like
A dictionary lookup: it finds the word you typed, not the idea you had.
Example
A legal search for the exact phrase "force majeure" returns only documents containing those words.
How it actually works
Lexical search compares the tokens in your query to tokens in documents, usually with stemming and stopword handling, then ranks by term frequency and rarity. It is transparent, fast, and precise for known terms, which is why it powered search for decades. Its blind spot is language variation — different words for the same concept simply do not match.
For product teams
Predictable and explainable, ideal when users know the exact term; frustrating when they paraphrase.
For engineers
Token-level matching over an inverted index with stemming/analysis; no semantic generalization.
Related
- Sparse Retrieval — Implemented with sparse retrieval methods.
- Semantic Search — Often blended with semantic search.
Read anything AI without the jargon
Look up any term in plain English, or save terms as you read with the free Chrome extension.
Open DecoderAdd to Chrome