The paper deals with word forms occurring in the SYN corpus (1.2 billion words) only once, twice or three times, the so-called hapax legomena, which provide a basis for the study of potentiality in language. As the material was very large, only 20 samples were chosen, each containing 3 000 forms, i. e. 60 000 forms overall. Approximately 50 % of word forms were various mistakes, especially typing errors, or words from other languages. Therefore only the remaining 30 000 word forms were selected as the basis for this study. The analysis showed that the most relevant suffixes for hapax legomena are -ovský (e. g. jimmyreedovský), -ák (e. g. medvěďák), -ista (e. g. havlista), -ing/-ink (e. g. gardening, dancink), -ovitý (e. g. kladivovitý), type po vojensku, diminutives derived from abstracts (e.g. minulůstka) and names of women professions (e. g. meduprodavačka). Moreover, compound words with the first parts dlouho- (e. g. dlouhorožec), gala- (e. g. galamenu), jino- (e. g. jinomluva), kino- (e. g. kinofajnšmekr), mega- (e. g. megakatastrofa), nízko- (e. g. nízkohlučný), polo- (e. g. poločitelný), video- (e. g. videokomentář) were typical for new words.
Along with derivation, composition represents the second most important word-formative process in Czech, primarily with certain names (such as professional terms). The paper deals with two specific word-formative types of deverbative names of persons, traditionally referred to as nouns of agents (nomina agentis) -compounds with suffixes -tel and -č. These compound names, excerpted from the Czech National Corpus (SYN2010) and confronted with Czech dictionaries (including neologisms), are compared with parallel derived-names, namely in terms of onomasiological and semantic functions of their constituent parts. Their systemic and empirical (textual) productivity (based on corpora) is further considered. Presented analysis is a part of larger research of Czech compounds conducted currently by the author.