Leksikon_001 Click on the image on the left to explore Sloleks, Slovene morphological lexicon with 100,000 words.


Sloleks is a lexicon of Slovene word forms. This means that it contains, in an XML database, basic information on the Slovene words, especially their word class and related features. For each word, all its word forms are provided. Since Slovene is morphologically very rich language, each word has many word forms. Declension is typical of nouns, adjectives, pronouns, numerals, verbs and adverbs. Word classes and their features are specified according to the specifications of the Linguistic Annotation of Slovene project (JOS), as explained next.


The Sloleks lexicon is conceptualized with two aims in mind: the use of the lexicon for (a) language technology applications and b) applications used by the Orthography Guide portal. By following these two aims, the contents of the lexicon have two contrasting tendencies: for language technology applications, the lexicon has to provide as exact information as possible on morphological characteristics of all the lexicon entries (words), found in authentic texts, including speech, and enable machine-readability of data. For the Manual of Style, it must contain the information on the norm, compared with contemporary reality of Slovene language use.

The ISO standard format Lexical Markup Framework was selected for structuring the information in the lexicon. The format is the latest result of a long series of standardization projects for recording European and other language data. The lexicon as part of language technology applications needs to be compatible with two other closely-related parts of the project: a training corpus and a tagger. Word classes (or categories) and their morphosyntactic features for these two resources are based on the JOS project specifications and are found in both the training corpus and the lexicon. Similarly, the same tagset is used by the tagger, making it possible to directly link the tags in the lexicon and the tags in the Gigafida and Kres corpora. Consequently, the information on the frequency of each word form in the corpus could be included in the lexicon database. In an online search engine of Sloleks both databases are connected, so we can see the concordances of any word form by clicking on the link in the “frequency” column.


Sloleks lexicon contains 100,784 entries or lemmas and 2,791,919 word forms with formalized grammatical descriptions.

Element Description Number Part of speech Number
<LexicalEntry> Element containing a lemma and its word forms 100,784 Nouns 54.254
<Lemma> Canonical or dictionary word form 100,784 Adjectives 26,597
<WordForm> Element with info about a word form 2.773.511 Verbs 10.242
<FormRepresentation> Element with info about a word form variant 2.791.919 Adverbs 6.906
<RelatedForm> Link to related lexical entries (in terms of word-formation) 65,951 Numerals 2,240
<Sense> Element with semantic information 74 Prononuns 169
<Definition> Element containing a definition 74 Prepositions 96
<ListOfComponents> Element contatining parts of a multi-word expression 3 Interjections 85
<MWELex> Element with a multi-word expression 2 Abbreviations 70
<MWENode> Part of the lexicon with multi-word expressions 1 Particles 68
<Lexicon> Root element 1 Conjunctions 54
Multi-word expressions 3
Total 100,784


Compilation of the lexicon: Peter Holozan, Simon Krek, Kaja Dobrovoljc, Miro Romih
Computational processing of the lexicon: Peter Holozan, Miha Arčan
Manual processing of the lexicon: Kaja Dobrovoljc
Integration of the lexicon in the Obeliks tagger: Miha Grčar, Matjaž Juršič



Simon Krek, Tomaž Erjavec, Peter Holozan (2008): Specifikacije za leksikon besednih oblik (kazalnik 3). Projekt Sporazumevanje v slovenskem jeziku.

Articles and books

Špela Arhar, Učni korpus SSJ in leksikon besednih oblik za slovenščino, Jezik in slovstvo 54/3–4, 2009, 43–56.