Orthography & Phonology in heŚakþe
▲
0▲ 0 ▼ 0
How to predict the pronunciation of heŚakþe words
This public article was written by [Deactivated User], and last updated on 19 Jan 2018, 22:29.
[comments] fuqorthographyalphabetphoneticsphonology
?FYI...
This article is a work in progress! Check back later in case any changes have occurred.
This article is a work in progress! Check back later in case any changes have occurred.
Below is a list of characters used in the heŚakþe alphabet; how to read heŚakþe phonetic guides; and the phonological rules that can be applied to predict heŚakþe pronunciation from spelling.
[top]The HeŚakþe Alphabet | HeŚakþek HeJoweźka
▼ A
▼ Æ
▼ B
▼ C
▼ Ć
▼ D
▼ Ð
▼ E
▼ F
▼ G
▼ H
▼ I
▼ J
▼ K
▼ L
▼ M
▼ N
▼ Ń
▼ Ŋ
▼ O
▼ P
▼ Q
▼ R
▼ S
▼ Ś
▼ T
▼ Þ
▼ U
▼ V
▼ W
▼ X
▼ Y
▼ Z
▼ Ź
[top]Phonetic Pronunciation Guides
For the assistance of those who are not yet familiar with the phonological rules of heŚakþe (seen below), there is a standard way to write pronunciation guides. It uses many of the same characters, but does not use L/l, Ń/ń, Q/q, or Y/y. Beyond this, it adds three nasal vowel characters: ă, ĭ, ŭ.
In order to read these, note that the capitalized syllable is given primary stress.
[top]Phonological Rules
There are a number of different phonological rules which are applied to heŚakþe words. All but the first can be applied automatically by CWS when it converts orthography to IPA, but they should still be known. Here they are presented in the order they are applied. Note that these are to get from orthography to phonetics, not necessarily from the underlying representation to phonetics, thus rules to handle eń dæhe and eń xuńa are included, while in speech nasal vowels are (arguably) phonemes.
The first rule applied is vowel deletion. It can apply in situations with the structure VCVCV, as long as the middle V is not stressed, and neither of the C's is an affricate. As long as that is the case, VCVCV -> VCCV. This can be clearly seen in the word "mahabes" (twelve), where in normal speech (though not necessarily in slowed speech /'mahabes/ -> /mahbes/ (which is affected further by a later rule, becoming ['ma:bes]. Also worthy of examination are "mabesaniś" (twenty-eight) and "mabesacaj" (twenty-one), pronounced [,mɑbes'niʃ] and [,mɑpsɑ't͡sɑj], where the preferred syllable for deletion (immediately before the primary stress) is deleted in "mabesaniś" but cannot be in "mabesacaj" and so the syllable before that is deleted instead. This rule is sometimes applied inconsistently, and its application can vary between speakers; but it does apply before all other phonological rules.
The second rule applied is the creation of voiceless sonorants. This happens in clusters of sonorant-h and h-sonorant, and leaves a phonetically long voiceless sonorant in place of both. Consider the word "mahaniś" (eighteen), where /mɑhɑ̃niʃ/ -> /mɑhniʃ/ -> ['mɑ̃n̥:iʃ]. Voiceless nasals still nasalize vowels (although that rule is encountered in the next rules).
The third and fourth rules govern how vowels nasalize, and are not ordered with respect to each other. Here I will order them for convenience: the third concerns the non-low vowels, while the fourth concerns the low vowels. Note that the given examples are nothing more than nonsense syllables. The third declares that the non-low vowels (e i o u) become high nasal vowels when they immediately precede a nasal; thus, "eń" and "iń" both become [ĩn], while "oń" and "uń" both become [ũn]. The fourth declares that low vowels (æ a) merge to the nasal back vowel when they precede a nasal, thus "æń" and "ań" both become [ɑ̃n]. It is because of these alternation patterns that many heŚakþe works call e, o, and æ the "weak vowels" and i, u, and a the "strong vowels", and combine these into one rule, in which all vowels become strong nasal vowels when they precede nasals.
The fifth rule deletes the /n/'s created by the letter 'n' (but not those created by 'ń') when they occur in a syllable coda and do not come before another nasal. Thus, "ran" (with) becomes [ɾɑ̃]. However, its two exceptions must be noted. Firstly, it does not delete initial n's; this is fairly obvious. But secondly, it does not delete medial n's, like in the word "cana" (to write). The word "cana" splits into the syllables "ca-" and "-na", and thus the 'n' cannot be deleted. Secondly, it does not delete when the sound after it is also nasal; this can be seen in the word "junma" (male), where the n does not delete, and instead assimilates to the 'm' (nasal place assimilation is later), giving the realization ['jum:a], rather than *['juma] (which is what would be predicted if the 'n' deleted).
The sixth rule is nasal place assimilation, and effects all nasal consonants. When there is a cluster of nasal-consonant (where the second consonant is not w, r, or h, not that an h could still be present there), the nasal takes on the place value of the consonant. This makes "ðimda" (last) become ['ðĩndɑ], as the /m/ becomes [n] to match the /d/. It is also seen in the earlier example, "junma", where /'jũnmɑ/ -> ['jũm:ɑ].
The seventh rule is obstruent voicing assimilation. It should be noted that this is not crucially ordered with any rule except the first. It affects clusters of two obstruents (noting that 'h' is not considered an obstruent for this purpose). Whenever there is a sequence of obstruents, the first takes on the voicing of the second. Thus, the word "rokda" (afraid) becomes [ɾogdɑ], as the 'k' takes the voicing of the 'd' and becomes a [g]. It is through this rule that the allophones [ɣ], [d͡z], and [d͡ʒ] can be created, from the phonemes /x/, /t͡s/, and /t͡ʃ/ preceding voiced obstruents.
The eighth rule concerns semivowels /j w/ in coda position. When they follow the vowel they correspond to (in the case of /j/, /i/; in the case of /w/, /u/), they lengthen that vowel and disappear. Thus, the word "þija" (woman) is simply ['θijɑ], as the /j/ is not in coda position, but the word "þijðarze" (lesbian) is ['θi:ðɑɾ,ze], as the /j/ is in coda position.
The nineth and final rule is vowel lengthening for cases with /h/ in coda positon. This is seen in the first example for the first rule, "mahabes". Its original form, /'mɑhɑbes/, deletes the medial vowel, leaving /mɑhbes/; this then has an /h/ in coda position (as the syllables split up as mah-bes) which is not deleted by later rules; thus, it lengthens the vowel in this final rule, leaving [mɑ:bes].
Comments
