Abstract: The subject of the study are homonyms and polysemous words. The problem of homonymy and polysemy always raises many questions, because sometimes there is no clear boundary where ambiguity ends and homonymy begins. Indeed, the boundary between homonymy and polysemy is unstable, permeable, it is constantly violated, and not only by polysemy, when, due to the loss of an "intermediate link", the connection between the individual meanings of a polysemantic word is lost, and it splits into two different homonymous words, but also (which is noted by only a few authors) by Homonymy is when, due to the action of various kinds of reasons, a semantic connection arises between two homonymous words and these words merge into one polysemantic word.The purpose of the article is to distinguish homonyms from polysemous words using an experiment between native speakers, by identifying homonymous words among respondents.
 In this article, special attention is paid to the differentiation of these phenomena by the psycholinguistic method. The author also uses a descriptive method when conducting research. The lexical richness of any language is words that are related to each other, but have different meanings and perform different functions. In colloquial speech, some words are close to each other in meaning, while others are opposite in their semantics. Some words, while not being close in content, can be similar to each other in appearance, shape, spelling and pronunciation. A group of such words in the lexical system of the language is commonly called homonyms.
The novelty of the research lies in the fact that the author for the first time made an attempt to distinguish homonymy and polysemous words in the Mari language using a psycholinguistic method.
The main conclusions of the article are that the conducted associative experiment is one of the most effective ways to distinguish between two lexical phenomena. Of course, the nature of associations is influenced by age, geographical conditions, a person's profession, as well as the level of proficiency in the Mari language, but, nevertheless, the opinions of respondents mostly coincide with lexicographic data.