In everyday language, quantities can be expressed with exact numbers or with quantifier words like “few” and “many”; non-numerical quantifiers are inherently more vague than numerical expressions. To investigate how this vagueness shapes the mental representation of non-numerical quantifiers, we administered two tasks to two independent groups of 35 participants. In the first task, participants viewed images depicting varying proportions of targets (animals) and non-targets (objects) and selected the most appropriate quantifier from a set of five. In the second task, using the same type of images, participants judged whether a displayed quantifier matched the scene. Across both tasks, the results revealed an ordered representation of quantifiers. However, in the second task, the category boundaries were much less clearly defined, with substantial overlap among the distributions associated with each word, indicating that their meanings lack sharp boundaries. Interestingly, in both tasks, the organization of quantifiers followed a logarithmically compressed scale: smaller quantifiers displayed more distinct and widely separated distributions, whereas larger quantifiers showed greater overlap. This pattern is consistent with how humans mentally represent non-symbolic quantities and Arabic numerals.