Abstract
<p>While humans perceive color as a nearly continuous spectrum, most languages categorize hues using a small set of basic color terms (BCTs). Several languages (e.g., Russian, Greek, Italian) lexicalize two or more basic ‘blue’ categories, distinguishing light from dark blue, a distinction English is usually said to lack. No basic split for English blue has been found, yet speakers may differentiate blue shades in everyday usage by attaching modifiers such as “light” and “dark” to blue more than to other color terms. Using the Corpus of Contemporary American English (COCA; ~560 million words, five genres), we extracted the 11 English BCTs with their most frequent preceding collocates and quantified chromatic specification, correcting for non-chromatic uses of the color words and their qualifiers. Although light blue and dark blue do not function as separate basic categories, blue attracts chromatic specifiers more often, and a far wider variety of unique ones (navy, cobalt, cerulean), than any other BCT, a lead we link to its communicative prominence and chromatic richness. Brown is the other heavily qualified term, but its qualification runs through common, largely achromatic modifiers rather than unique specifiers, a pattern we link to its status as a constitutively dark, lightness-defined color whose subcategories are largely lightness variants. The central contrast replicates on the current, larger COCA release (~993 million words, eight genres) using an independently constructed classification of qualifiers. We argue that color-term basicness is gradient rather than discrete, and that corpus structure offers a naturalistic window onto the vision–language interface.Keywords: color categorization; basic color terms; color names; blue; brown; corpus linguistics; COCA; vision-language interface</p>