Some doubts and problems in categorizing lexical data

U radu se, osim kratkoga pregleda cjelovitih prijevoda Biblije na hrvatskome u prošlomu i ovomu stoljeću, govori o važnosti biblijskih tekstova i za jezikoslovna istraživanja. Radom se želi pokazati kako u brojčanim istraživanjima pojedinih biblijskih tekstova, iako se to odnosi i na drugačije tekstove, nije uvijek jednostavno donijeti odluku što brojiti. Izbor nije uvijek jednoznačan zbog mogućih drugačijih teorijskih pristupa ili zato da bi se rezultati mogli usporediti s dotadašnjim istraživanjima. Na primjerima se, većinom iz Lukina evanđelja, pokazuje kako u odabiru građe za brojčanu raščlambu različite odluke mogu ili ne moraju bitno promijeniti rezultate istraživanja. Najviše se razlika pokazuje na primjeru vlastitih imenica u odnosu na rod, padeže i završni slog budući da vlastite imenice imaju osebujna obilježja. Pokazuje se koliko je postotni udio padeža ili drugih kategorija različit bez vlastitih imenica ili s njima te u čemu se drugačija i vrlo malena građa može usporediti sa znatno većom.After a short overview of Croatian whole Bible translations in the last and this century, this paper points to the importance of Biblical texts for linguistic research, esp. as they have different synchronic or diachronic versions. The paper also points to the fact that the quantification of vocabulary according to certain categories, and categorization of marginal members, may depend on the purpose of the quantitative research. Sometimes the purpose of Biblical language research does not have to do with the content at all. Gospels may be taken as examples of known and simple texts, which could be used to find out how many different word forms make a simple texts in order to predict the usefulness of a dictionary which consists of citation forms only. Sometimes the purpose may be to predict the frequency of some categories in a similar text, as when the percentages of different declension types in Matthew, Luke and John are used to predict those in the same Croatian translation of Mark. The choice may also depend on the linguistic discipline the research is dealing with. For example, common nouns written with capital letters may be interesting in orthography, semantic or stylistic research and considered variations of the same lemma, or even different lemmas, but for morphological purposes they may be viewed as tokens of the same, common noun. Sometimes it is difficult to make a decision about some categories due to different theoretical frameworks (as traditional categories could be more appropriate in comparing recent with previous findings). Sometimes it is difficult to decide due to ambiguous morphological categories, such as adjectives that were nominalised but still keep their adjectival flexion - the decision does not play an important role among nouns or verbs, but does affect adjectives as their numbers are much smaller. The example of proper nouns and the effect of their inclusion or exclusion on some of the researched categories are presented. It is shown that proper nouns have more nominatives, datives and vocatives, while general nouns have more accusative, genitive, locative and instrumental cases, which is not surprising as nominative is typically the Case of a subject or predicative and that accusative is typically the objective case. The distribution of gender is different: while general nouns have 44% masculine, 38% feminine and 18% neuter nouns, proper nouns have 76% masculine, 23 feminine and 2% neuter nouns