Negative Space — The Shadow Words Almost Written

Word Connotation

When a writer chooses a word, they pass through other words on the way to it. “She walked to the door” — but she could have strode, strolled, trudged, shuffled. The reader feels those alternatives without naming them. The chosen word carries the weight of what was almost chosen instead.

This is the decision space around every word: not synonyms, but plausible replacements, words that could have occupied the same position in the same sentence. Choosing “walk” over “stride” reveals neutrality over confidence. Choosing “house” over “home” reveals structure over belonging. The shadow exposes the choice.

What the tool does

Paste any text and the analyzer reads each word in context using Claude, Anthropic’s AI. Rather than matching against a fixed synonym list, it considers the specific sentence — cold in cold water and cold in cold stare produce different shadows. A local table of roughly 660 common entries handles instant rendering while the API call completes; Claude’s results overlay on top.

Hover any underlined word to see its shadows, the alternatives that were passed through. The underline weight reflects shadow depth: more alternatives means a more contested decision point. A heatmap across the top shows sentence-by-sentence where the decision space is densest.

Compare Drafts 

lets you paste two versions of the same passage side by side. The Shadow Shift Index measures what percentage of the mapped vocabulary changed between drafts, which decision points were added, which were abandoned, and which held steady through revision.

What it doesn’t do

The table is a curated sample, not a complete map. Every word in natural language sits in a decision space far larger than any static table can represent. Coverage will vary, literary and personal writing tends to score higher because the table is built around literary vocabulary. Technical or specialized writing will show lower coverage, not because it lacks depth, but because it draws from a vocabulary the tool doesn’t map against. The coverage percentage is always visible so you know exactly what the tool is and isn’t catching.

Negative Space Analyzer
The words that were almost chosen
0 / 500
0 / 500
Context-aware analysis powered by Claude. Each word's alternatives are chosen for that specific sentence — "cold stare" and "cold water" produce different shadow words. Not synonyms: decision-space neighbors.
Reading Decision Space
Analyzing context
Negative Space
Revised
Shallow Mid Deep