The notion of ‘zachte g’ (literally ‘soft g’) is often used in a confusing way in discussions
about Dutch pronunciation. It usually refers to a north-south divide, where Dutch-speaking Belgium
and the Dutch provinces Noord-Brabant and Limburg have the ‘zachte g’, and more northern regions
have not. This difference is in essence one of place of articulation: The south uses velar sounds
i.e. the friction occurs in the back of the mouth, but not really that far back. These sounds
can be influenced by adjacent vowels, to varying degrees depending on region, which can go
as far as making them palatal sounds, similar to the German so called Ich-Laut. But others keep
them closer to velar even in the vicinity of front vowels such as [i], [e] and [E].
Similar velar fricatives occur in modern Greek, written chi and gamma, when followed by a consonant, or a vowel a, o, or u (written ou).
More northern regions of the Netherlands use fricatives, or one fricative, with a place of articulation farther back in the mouth. These are guttural sounds (I’m not sure about this term, it may be a synonym of velar) or even uvular. They are, or it is, similar to sounds used in Arabic and modern Hebrew.
But the term ‘zachte g’ is also used for whether or not there is a clear distinction between the sound written ch (always voiceless) or written g in final position (where all voiced fricatives and plosive becomes voiceless in Dutch, like in German and Russian) on the one hand, and g in medial or initial position. Some make the latter sound fully voiced, and make it clearly different from the voiceless sound. With others, the sounds are very similar, or even completely the same. Those who talk like that often also don’t distinguish between s and z, and f and v. See also Twijfelfonemen and Labio-dentals in Dutch about this.
Now the confusion arises because many who do not or hardly distinguish between the voiced
and voiceless sounds, often also have the backer type of sounds regarding place of articulation.
So they do not have the ‘zachte g’ in both senses of the term. This leads many to believe
that this is a fixed rule, and they have difficulty understanding the two completely different
meanings the term ‘zachte g’ can have.
In fact, place of articulation and voicing difference do not always coincide, and all four combinations do occur. I think the combination of the ‘fronter’ kind of g/ch and no voice distinction occurs in Nijmegen, but I'm not entirely sure of that.
One thing I do know for sure is that the backer variety combined with a voice distinction does occur, because I am one of the people with such a pronunciation myself. Because this is difficult to make clear by just describing it in words, here is a sound sample, of the not-quite-minimal pair ‘waggel’ vs. ‘kachel‘. In this position between vowels, the gg in my pronunciation is voiced and lax, or at least more so than the fully voiceless and tense medial ch.
(Sound samples are in au-format, for Sun/Next computers, but also supported by most browser plug-ins for other platforms)
So this pronunciation stems from Rotterdam, where I was born (but I don’t quite talk
like people talk there in every detail any more, I think).
To show the difference, and to make clear that there really are four different sounds, I have tried to record these words also with the velar sounds uses in the sounds, and again with the voiced and voiceless sounds distinguished. The result is less than satisfactory, because for this type of pronunciation I’m not a ‘native’ speaker, and the w, k and l have to be a little different too. Anyway, I hope my imitation is good enough to illustrate a point.
Voiced and lax vs. voiceless and tense:
If the difference is not only in the presence or absence of voice, it means that it should be able to survive whispering - in whispered speech all speech sounds are voiceless by definition. And it does survive it in my speech. Listen to this (now using a different example):
vlaggen - lachen (gefluisterd)
As a further attempt, I have recorded the word ‘echt’ (=real(ly)) four times, while using these sounds for what is written ch:
I can imagine that readers may still find the above samples unconvincing, especially where the difference between southern Dutch [x] and northern [X] is concerned. My imitations are not good enough, and in a real word, the sounds are extremely short, almost too short to appreciate finer shades of difference. For this reason you’ll find below a artificial ‘word’ with four ultra-long fricative consonants. I hope this will make the differences clearer, and help recognise them in real-life situations.
Phonetically, the following samples are meant to be:
[C:: x:: X:: H::]
[C<vcd>:: Q:: g":: H<vcd>::]
Later addition, 29 April 2011, some more near minimal pairs:
Copyright © 1999-2005 by R.Harmsen.
Last updated April 1, 2005
Moved to its own subdirectory: 28 March 2011 (also changed " " to ‘ ’, and updated some hyperlinks)