Corpus SYN2000

Name SYN2000
Positions Number of positions (tokens) 120 908 724
Number of positions (tokens) without punctuation 100 061 381
Number of word forms (words) 1 763 813
Number of lemmas 891 713
Structural attributes Number of documents (not opera) 233 797
Number of sentences 7 639 321
Further information Reference YES
Representative YES
Publication date 2000

The corpus SYN2000, published in October 2000, contains 100 million words and is composed of complete texts only. The criteria for selecting texts were based on researches of written language: they were to cover the widest possible genre stratification of the Czech language. The SYN2000 is a synchronic corpus, which means that it covers contemporary Czech. Therefore it contains primarily texts that were created in 1990–1999. However, also important works of Czech literature were included in the corpus (i.e. Karel Čapek's Krakatit or Josef Škvorecký's Zbabělci (The Cowards)). As to older texts, there has been a rule that authors had to be born after 1880 for the text to be included in this corpus.

An inspiration for the SYN2000 corpus was the British National Corpus, however work on the BNC ceased in 1994.

The SYN2000 corpus is lemmatized and morphologically tagged. That means that for each word (that is the occurrence of the word in the text) its morphological tag, which shows its grammatical categories (the part of speech, number, case etc.) and so-called lemma, which is the basic form of the word (for instance, in case of nouns, it is the nominative singular, for verbs it is the infinitive) can be viewed. Besides these, you can view the code, which identifies the text, in which the searched word occurred.

The FSC2000 corpus is a modified version of the SYN2000 with enhanced lemmatization, which was used as a source for the Frequency Dictionary of Czech.

Changes in the SYN series corpora

Please take note of the vital changes in the composition and processing between the SYN2000 and SYN2005 corpora (and also SYN2000 and SYN2010), which are summarized on the pages relating to the corpus SYN2005. A consequence of these changes is a difference in the data relating to frequency. Changes in the concept of the corpus’s representativeness, resulting in vital differences in the composition compared to other corpora from the SYN series, can be observed in the following table which compares the text-type composition of the SYN2000 a SYN2005 corpora.

SYN2005 SYN2000
fiction 40 % 15 %
non-fiction 27 % 25 %
journalism 33 % 60 %

Composition of the SYN2000 corpus

Structure of corpus SYN2010: 60 % journalism, 25 % technical literature, 15 % fiction
Structure of technical and other specialized literature according to thematic orientation (no. of words in mil.)
Structure of journalism according to the year of issue (no. of words in mil.)
Structure of journalism according to the newspaper title (no. of words in mil.)

Structure of the SYN2000 corpus

Among the structural units used in this corpus are <doc> and <s>; document and sentence – followed by each individual position. They can be displayed using the menu item View options. The newer corpora of the SYN series have an additional, higher structure labelled <opus> (this difference is essential for example when searching with the help of the within condition).

Structural units in the SYN2000 corpus.

How to cite SYN2000

Čermák, F. – Blatná, R. – Hlaváčová, J. – Klímová, J. – Kocek, J. – Kopřivová, M. – Křen, M. – Petkevič, V. – Schmiedtová, V. – Šulc, M.: SYN2000: žánrově vyvážený korpus psané češtiny. Ústav Českého národního korpusu FF UK, Praha 2000. Available on-line: http://www.korpus.cz