Learning some English Consonants

It is fascinating to watch small children learning to use language. My grandson speaks very clearly at three years six months old and has done for some time, but hasn’t yet mastered a couple of things in the English sound system. Especially at the beginning of words, he pronounces:

• s as d, for example dilly instead of silly
• f as b, for example bunny instead of funny
• gr as fr, for example fran and franfrad instead of gran and grandad.

These are not just random errors, they are systematic differences between his sound system and the sound system of an adult native speaker. The following table analyses the features of some consonants in adult English.

Coronal: sound made with tip of tongue near teeth
Voiced: sound made with strong vibration of vocal cords
Unvoiced: sound made with weak vibration of vocal cords
Fricative: sound in which air flow continues
Stop: sound in which air flow stops completely
Labiodental: sound made by closing mouth with lower lip and upper teeth
Bilabial: sound made by closing mouth with both lips

Substituting d for s

As the table shows, his first substitution (d for s) replaces an unvoiced fricative with a voiced version of the nearest corresponding stop. Some of the literature or children’s acquisition of language suggests that many children acquire fricatives later than stops. This seems to be going on in my grandson’s case: he is not yet able (or perhaps not yet willing) to produce a fricative, so he defaults to an otherwise similar stop.

Where does this feature of his sound system sit? If it sits within his perception system, he couldn’t identify fricatives. Or if it sits within his production system, he could identify fricatives, but wouldn’t produce them. I tested out which it was when he told me his toy was dick (meaning sick). When I asked whether his toy was dick, he rejected my question and wouldn’t accept it until I asked whether the toy was sick. So his perception system decodes s, but his production system won’t produce it, at least in some cases.

This result is sometimes called the ‘fis phenomenon’ because of an incident reported by J Berko and R Brown in 1960. A child referred to his toy fish as a fis. The child rejected adult utterances using the form fis and would only accept the adult form fish.

My grandson’s change from s to d doesn’t just turn the consonant into a stop: it also makes it voiced (d, not t). I am not sure what causes this change. Interestingly, though, one other change he commonly makes is to pronounce this as dit. At the start of this word, the voiced coronal fricative <th> changes into the voiced coronal stop d. But at the end of the word, the unvoiced coronal fricative s turns into the unvoiced coronal stop t. Why doesn’t he change that s into the voiced d (rather than unvoiced t)? I don’t know, but maybe it’s because this change occurs at the end of the word (or perhaps just at the end of a syllable), rather than at the beginning).

As an aside, I’m not entirely sure whether the word he produces as dit is this, that, sometimes this and sometimes that, or a single word that covers both this and that, without distinguishing them.

Substituting b for f

My grandson’s second substitution is of b for f. As the table above shows, this is another substitution of an unvoiced fricative (in this case labio-dental) with a voiced version of the nearest corresponding stop (in this case, bilabial).

Substituting fr for gr

My grandson’s third substitution is also interesting. At the beginning of words, he produces fr- instead of gr- .

Consonant clusters are notoriously difficult for young children to produce, so they often simplify them by various strategies, such as dropping one of the consonants, inserting a vowel or changing one of the consonants. My grandson uses the third strategy. In contrast, up until the age of about 3 his older brother routinely used the first strategy, simplifying the cluster gr- by replacing it with d (so dan and dandan instead of gran and grandad).

There is one very striking thing about my younger grandson’s third substitution: he produces the consonant f that he can’t (or won’t) produce in words like funny. So what is going on in this case? He retains the consonant r (a voiced, coronal approximant), but changes the preceding g (a voiced, velar stop) into f (an unvoiced, labiodental fricative). There seem to be two parts of this change:

• partial assimilation of g (stop) to r (approximant) to make g fricative. G is a stop: the airway closes totally. But r is an approximant: the airway doesn’t close totally, and in fact doesn’t even narrow enough to create much friction. Influenced by the following r, the first consonant in the cluster (g) becomes a fricative (f)—not completely closed, but producing more friction than an approximant does.
• assimilation of the velar feature of g towards the coronal feature of r. G is velar: the airway is closed at the back of the mouth by contact between the the body of the tongue and the velum (soft palate). R is an approximant—there is no closure, nor even significant friction, but the tip of the tongue is at the front of the mouth, just behind the teeth. So under the influence of the following r, the first consonant of the cluster (g) (which constricts the airway at the back of the mouth) becomes instead a consonant (f) that constricts the airway at the front of the mouth (in this case, lower lips and upper teeth).

These two components of this process both seem entirely natural and seem to produce a sequence of sounds that is easier to say. (I’m not sure why the first consonant doesn’t also assimilate to the voiced feature of the following r, though I don’t perceive that voiced feature as particularly strong anyway.)

Summary

In summary, at the beginning of words, I think my grandson is not just making random errors but making systematic changes showing an underlying structure, though that structure is not yet the structure adult speakers have. He converts:

• unvoiced coronal fricatives into voiced coronal stops
• unvoiced labial (labiodental) fricatives into voiced labial (bilabial) stops
• voiced velar stops into unvoiced labial (labiodental) fricatives, when they are followed by a voiced coronal approximant