Czech phonology #

This file defines the phonemes of Czech as segments, the natural classes of its velar alternations, the second palatalization of velars and the reading of Czech spelling as phonemes, after Short's description of the language. Czech has five vowels, each short and long, and twenty-five consonantal phonemes. The second palatalization replaces k by c, h and g by z, and ch by š. Short files h among the laryngeals, and in the features of the PHOIBLE chart it is not dorsal, so the four segments the rule changes are not the dorsals. They are a natural class all the same, the obstruents that are neither labial nor coronal.

The spelling is close to the phonemes. A palatal ď, ť or ň is written on a following ě or i, as in dě and di, a labial before ě is followed by /j/ and m by /ň/, and i and y write one vowel.

Main definitions #

Czech.Phonology.a, Czech.Phonology.ch and the like: the phonemes, named by the letters Short writes them with, and Czech.Phonology.inventory the set of them.
Czech.Phonology.dorsals, Czech.Phonology.palatalizing: the velars k, g and ch, and these with h.
Czech.Phonology.secondPalatalization: the second palatalization of velars.
Czech.Phonology.ofString?, Czech.Phonology.letter: the phonemes a word writes, and the letter that writes a phoneme.

Main results #

Czech.Phonology.exists_mem_ces, Czech.Phonology.ofPhoneme_mem_or: the phonemes are those of PHOIBLE's Czech inventory.
Czech.Phonology.naturalClass_palatalizing: the segments the second palatalization changes are the obstruents that are neither labial nor coronal.
Czech.Phonology.voiced_mem_iff, Czech.Phonology.devoiced_mem_iff: the obstruents paired in [voice], all but c, č, ch and h.
Czech.Phonology.secondPalatalization_g, Czech.Phonology.not_exists_merge_eq: the reflex of g is that of k voiced and de-affricated, so no one feature change gives the rule.
Czech.Phonology.mem_inventory_of_mem_ofString?: the spelling writes phonemes of the inventory.

Implementation notes #

The inventory is PHOIBLE 1140, whose consonants are Short's twenty-five and whose vowels are his five short and five long vowels and three diphthongs. PHOIBLE marks /oː/ and the diphthongs /au/ and /eu/ marginal, the three that Short confines to loan-words (p. 456). PHOIBLE's other Czech inventories, 2274 and 2480, add the glottal stop and the voiced affricate [dž], which Short counts as not phonemic (p. 457).

Hayes's features have no length, so a long vowel is two segments of its quality, and a diphthong, which the chart omits as contour-valued, is its two vowels. A vowel keeps the features that distinguish the five, since the chart gives the short high vowel as the lax [ɪ] and the long one as the tense [iː]. The chart's t, d and n are alveolar where Short calls them dental, a difference in [distributed] on which no Czech contrast depends.

ofString? reads the phonemes and not their realization: the final devoicing and the voicing assimilation of obstruents (p. 458) are left out, so that oběd 'lunch' is /objed/ where Short writes /objet/. It follows Short's rules for the palatals without exception, so a loan-word in which di, ti or ni writes a plain dental is misread, and it reads x as [ks], one of Short's two readings.

References #

Phonemes #

The phonemes are Short's (pp. 456–457), each named by the letter he writes it with, so that ť is [c], c is [ts] and h is [ɦ].

Czech phonology #

Main definitions #

Main results #

Implementation notes #

References #

Phonemes #

Natural classes #

Voicing pairs #

The second palatalization #

Spelling #