Across the Dene languages, the sounds are represented with similar symbols in their orthographies. The following is a general version of a Dene orthography. The top row are the sounds under discsussion in this section, these sounds are represented by a voiced symbol in some of the Dene languages these are voiceless sounds.
Dene Sųłine voiced d
Figure 1
North Slavey voiceless d
Figure 2
The Dene languages vary in whether the ’voiced’ stop or affricate in the series ( d g dz dl j ) are actually voiced sounds. In some Dene langauges they are voiced [d] , in some they are unaspirated and voiceless [t].
In Figure 1 is a d [d] phoneme from Dene Sųłine; this sound is fully voiced, as we can see in the spectrogram and waveform. There voicing all the way through the closure period of the d sound. This pattern can be seen in all the voiced stops in the series: g, dz, j and dl. In Dene Sųłine the voiced stops are voiced. Note also that the release burst is followed very closely by the voicing in the vowels. Dogrib is also a voiced stop language; as we can see in these examples: d, g, dz, dl, j and in several others examples in the dataset.
In North Slavey, the stops are voiceless and unaspirated. In Figure 2 is an example of a word with two voiceless d [t] sounds in North Slavey. There is no voicing during the closure period, in the IPA this is represented as a t. Note there is no voicing present during the closure, even when the d [t] sound is between two vowels. North Slavey and Navajo are a voiceless stop languages.
Variability in voicing: When a stop is voiced, a voicing bar is present during the closure. In a fully voiced stop, it is voiced all the way through the closure, as in Figure 1. This is actually uncommon. Many voiced stops, even within the same language, only show partial voicing. You can see this by contrasting the g [g] in Dene Sųłine with the d in Figure 1. This g [g] is only partially voiced. The reasons for this partial voicing are physiological; to produce full voicing through a stop or affricate closure requires extra effort to maintain the vibration of the vocal folds, due to the mechanics of airflow and air pressure.
Dogrib voiced d and dz
Figure 3
Another example of variability in the voicing of voiced stops can be seen in Dogrib in Figure 3. The d sound in initial position does not appear to be voiced; the voiced affricate dz in medial position is voiced through the closure portion of the affricate. Languages with voiced stops, like Dene Sųłine and Dogrib, often have a lot of variability in the voicing of their stops, sometimes they do not appear to be voiced at all,as we see n initial position in the example in Figure 3. Voicing is often conditioned by position in an utterance, intervocalic stops are often more voiced than word initial ones.
In languages with voiceless stops, like North Slavey and Navajo, there is less variability. The stops are voiceless in these languages, even intervocalically the voicelessness is maintained (as in Figure 2).
A second issue that comes up is the relationship of the sounds written as voiced sounds ( b d g dz j dl ) and the sounds written as voiceless sounds ( t k ts ch tl ) in the series. While there is variability among the Dene languages in whether or not the ‘voiced’ sound in the series is a voiced or voiceless unaspirated, this is not the case with the sounds written as voiceless t; these are all described in very similar ways and are realized in similar ways across the Dene languages, independent of whether or not the d sound is voiced [d] or voiceless [t].