Zipf law plot (frequency as function of frequency rank) for various texts.

The languages, texts and the word frequency files are:

[[Arabic language|Arabic]]. The ''[[Quran]]'' (~650 CE). Based on the ''Unicode Quran'' document from the Sacred Texts site, maintained by John B. Hare, with several corrections. Arabic Unicode characters were mapped into [[ISO 8859-1|ISO latin-1]] characters in a vaguely phonetic way.  '''With''' vowel marks, hamza, madda but '''without''' sukuns.

* Whole text. Sample: ''<nowiki>bīsmī alllāhī alrrāµmānī alrrāµīymī alµāmdū līllāhī rābbī alæālāmīynā</nowiki>'' [...] ''<nowiki>tāttā©ī£ūwnā mīnhū sākāräa wārīzqäa µāsānäa a”īnnā fīy</nowiki>''. File arab/quv/tot.1/gud.wfr (original 77411 words, truncated/filtered to 35027 words, ''N'' = 10762 distinct).

The ''[[Quran]]'' (~650 CE). Based on the ''Unicode Quran'' document from the ''Sacred Texts'' site, maintained by John B. Hare, with several corrections. Arabic Unicode characters were mapped into [[ISO 8859-1|ISO latin-1]] characters in a vaguely phonetic way, with joined articles.  Version '''with''' short vowel marks, sukuns, hamza, and madda.

* Whole text. Sample: ''<nowiki>bīs°mī alllāhī alrrāµ°mānī alrrāµīymī al°µām°dū līllāhī rābbī</nowiki>'' [...] ''<nowiki>a!ān° attā©ī£īy mīn° al°jībāalī būyūwtäa</nowiki>''. File arab/quf/tot.1/gud.wfr (original 77394 words, truncated/filtered to 35027 words, ''N'' = 10935 distinct).

The ''[[Quran]]'' (~650 CE). Electronic phonetic transliteration from [[University of Southern California|USC]]'s Muslim Student Union, transliterated again into [[ISO 8859-1|ISO latin-1]] characters.  The result, particularly short vowel marks, is partly guessed. The subtler phonetic marks (sukuns, hamzas, maddas, alef-maksura, teh-marbuta etc.) were lost.

* Whole text. Sample: ''<nowiki>bīsmī allāhī alrrāµmānī alrrāµymī alµāmdū līllāhī rābbī alæālāmynā</nowiki>'' [...] ''<nowiki>wāmā</nowiki>''. File arab/qph/tot.1/gud.wfr (original 77845 words, truncated/filtered to 35027 words, ''N'' = 9434 distinct).

The ''[[Quran]]'' (~650 CE). Electronic version of unknown provenance, without short vowel marks or sukuns, in a vaguely phonetic transliteration into [[ISO 8859-1|ISO latin-1]] characters.

* Whole text. Sample: ''<nowiki>bsm allh alrµmn alrµym alµmd llh rb alæalmyn alrµmn alrµym malk ywm</nowiki>'' [...] ''<nowiki>nfwra nµn aælm bma ystmæwn bh a£ ystmæwn</nowiki>''. File arab/qcs/tot.1/gud.wfr (original 74212 words, truncated/filtered to 35027 words, ''N'' = 9025 distinct).

The ''[[Quran]]'' (~650 CE). Based on the ''Unicode Quran'' document from the ''Sacred Texts'' site, maintained by John B. Hare, with several corrections. Arabic Unicode characters were mapped into [[ISO 8859-1|ISO latin-1]] characters in a vaguely phonetic way, with joined articles.  Version '''without''' short vowel marks, sukuns, hamza, madda.

* Whole text. Sample: ''<nowiki>bsm alllh alrrµmn alrrµym alµmd lllh rbb alælmyn alrrµmn alrrµym mlk ywm</nowiki>'' [...] ''<nowiki>walllh anzl mn alssma' ma' faµya bh alarš bæd mwtha ann fy</nowiki>''. File arab/qud/tot.1/gud.wfr (original 77455 words, truncated/filtered to 35027 words, ''N'' = 8531 distinct).

The word frequency files '*/*/*/gud.wfr' are available at the [https://www.ic.unicamp.br/~stolfi/EXPORT/projects/voynich/Notes/tr-stats/dat/ UNICAMP website].  The original annotated full texts, before truncation/filtering, are in the companion files */*/org/main.src.  The truncated/filtered texts -- one word per line, without punctuation -- are in */*/*/gud.tlw.