When consonants become vowels: A view from Danish and Dutch
Rasmus Puggaard-Rode discusses a few instances where the seemingly obvious boundary between vowels and consonants blurs, and one of the most stable distinctions in phonology dissolves.
I don’t remember learning much about phonology in primary school. It sometimes feels like teachers will go to great lengths to avoid talking about it. An example of this is when I started learning Dutch. That was when I was first introduced to the term soft ketchup. This turned out to be a weird but useful mnemonic teachers use to cover an important phonological generalization, without actually teaching students anything about phonology. It is used as follows: the past tense of a verb is formed by adding either -de or -te to the stem – you choose te if the last letter is one of the letters in soft ketchup. For example, the past tense of the stem werk (to work) is werkte (k is a soft ketchup consonant), and the past tense of the stem bel (to call) is belde (l isn’t a soft ketchup consonant). As for the important phonological generalization it makes; the sounds written as s f t k ch p are a clearly defined class of sounds, namely, they’re all voiceless.
In fact, the only phonological distinction I remember hearing about in primary school in Denmark was that between vowels and consonants. You know, the red letters and the blue letters. This is apparently the one major distinction between classes of sounds that we are comfortable with teaching our 6-year-olds.
Vowels vs. consonants
The distinction between vowels and consonants is relatively easy to grasp. You’ll usually find vowels at the center of syllables. When uttering a vowel, your mouth is open and you’re allowing outgoing air from the lungs to flow freely. And vowels are voiced. So if you hold your hand to your throat while holding an ‘aaah’ sound you should feel your whole head vibrating.
Consonants, on the other hand, are at the edges of syllables. You are generally using your tongue or lips to block outgoing air in some way. And they might not be voiced. (Recall soft ketchup: say these words without the vowels, sft ktchp, and you won’t feel your skull vibrating at all.) Stops are the prototypical consonants. These are the sounds usually represented with the letters b d g p t k. You use your tongue or lips to block outgoing air completely, and they’re usually found way at the edges of syllables. Some are voiced, some aren’t. You could say that voiceless stops are the least vowel-like sounds of all.
When consonants become vowels: goed, bage, bagte
Yet, consider the Dutch word goed (good). The final sound in goed is a voiceless stop, pronounced by raising your tongue tip to the fleshy area behind the upper teeth and blocking outgoing air. In words like goeiemiddag (good afternoon), that voiceless stop is replaced with a vowel-like i-sound. It’s obviously the same “word”, but the final sound alternates between the most consonant-y of consonants and something that looks suspiciously like a vowel.
More extreme versions of this phenomenon are common in Danish. Consider the infinitive form of the verb bage (bake), which is not pronounced how you’d expect based on the spelling. The second syllable, spelled ge, is pronounced as the vowel [ɪ] (roughly the same vowel found in English words like kick, big, lift, etc.) Like all vowels, it’s voiced, and is pronounced by raising the front of the tongue a bit in the direction of the roof of the mouth, although not in a manner that in any way blocks outgoing air. The past tense form of the verb is bagte, and this time around the pronunciation is much closer to the spelling. That <g> is pronounced as a stop. It’s voiceless and pronounced with the back of the tongue lifted to the soft palate in front of the uvula, where it blocks outgoing air completely.
How does it work?
It’s often assumed that these two very different sounds – the stop and the vowel – must at some cognitive level be one and the same. Mentally, the root of the word must be stored in some abstract form which encompasses both pronunciations. When a speaker says the word out loud, she retrieves the word from storage, and a complex system of rules are applied to the abstract form of the word. This eventually returns the desired pronunciation. But the process must be quick and precise; you don’t ever hear speakers of Danish stumbling and hesitating while computing the right pronunciation, nor do they ever pick the wrong one by mistake.
What is my take on the issue? I don’t think we need to overcomplicate the story. We know from historical sources that the vowel used to be a voiced stop, and that in certain contexts, such as in-between other vowels, it gradually traveled from stop to vowel through a number of intermediate steps. Over centuries of producing this sound, speakers’ tongues gradually moved further from the roof of the mouth, and the highest point of the tongue gradually moved forward. These intermediate steps can all be explained; the resulting sound either became easier to pronounce or to perceive. The particulars are too technical to get into here, but I heartily recommend my forthcoming dissertation, or this brand-new article I co-wrote with colleagues from Denmark, for more details.