This chapter discusses the general restrictions on possible arrangements of sounds in Skerre. The more specific restrictions (and the fixes to all these restrictions) are discussed in the next chapter. Because the phonotactic restrictions in Skerre are quite transparent, most transcriptions in this chapter will be phonemic; the rules given in the next chapter make the phonetic realizations easy to recover.
The topics of this section include:
Sequences of alternating consonants and vowels — long and short — commonly appear in Skerre words and there are few restrictions on their phonetic content or their location within words. Generally, words begin with consonant (though see below). Words can end with short vowel, a long vowel, or a short vowel plus a consonant. However, words cannot end in a long vowel plus a consonant. Examples of the possible sequences are given below.
|CV:CV||ke:ni||take care of||CV:CVC||ke:ter||bind|
|CV:CV:||se:ra:||stump (of a tree)|
|*Note: function words are the only CV words.|
The words in the last row are not examples of how large Skerre words can get (there is, in fact, no maximality constraint), but they are some of the larger exemplars.
Vowels without a preceding consonant are permitted word-initially. However, these are phonetically realized with either an initial epenthetic glottal stop (as described by this fix-up rule), with an initial epethentic semivowel, or use the preceding consonant for an onset, making them no different phonetically to the strings above.
In addition to these consonant-vowel-(consonant) sequences, Skerre also allows consonant clusters. However, there are restrictions on consonant clusters (and also word-final consonants), so I will now discuss these restrictions at each of the beginning, middle, and ends of words.
At the beginning of words, certain two-segment consonant clusters are allowed. All of these initial clusters are morphologically-conditioned; that is, they are all instances of a prefix comprised of a single consonant attached to a stem-initial consonant. The two prefixes involved are the agentive nominalizer s- and the instrumental nominalizer ts-. These prefixes cannot create geminates (as part of a larger pattern in the language), and also cannot appear with some sounds phonetically similar to them. Given below are the possible clusters (phonemically), sorted by first segment.
|Possible sC clusters||Possible tsC clusters||st
For examples, see the noun formation page. All other combinations of s + C and ts + C must be fixed up. The sequences /sj/ and /tsj/ are subject to a particular further alternation. Finally, all the clusters above obey the further restrictions on clusters in Skerre.
Medial clusters are also possible in Skerre and, unlike the initial clusters, they are not morphologically-conditioned. However, the possible medial clusters are still quite restricted. They are given in the table below:
As the above table shows, the initial segments in these clusters are all continuants and the final segment in these clusters is almost invariably an obstruent.
In the actual phonetic realization of the NC clusters, the nasal must be articulated at the same place as the following consonant (per this) and the consonant must be voiced like the N (per this) Some examples of NC clusters.
|ranha||[ɾanɦa]||yell, call out||([4anh\a])|
In their phonetic realization, SC and RC clusters must also agree in voicing (again per this constraint). Some examples:
Consonant clusters are not allowed in word-final position. Furthermore, the range of singleton consonants allowed in word-final position is more restricted than other positions. The permitted word-final consonants include all stops, fricatives, nasals, and rhotics (t, ts, k, ?, s, n, r), but not the approximants [h], [j], [w], or the approximant-final [kw]. Some examples of words with final consonants:
The additional following sequences of segments are not permitted in Skerre:
Thus far, this section has talked just about restrictions on phonotactics within the word. Beyond the individual word, the restrictions are considerably more free: aside from a few alternations which apply across word boundaries, any two sounds can appear next to each other. These same word-level phonotactics also apply in compounds.