14 August 2011

Phonology -- whats and hows part II

Last time, I wrote about phonology and the necessity of physically training the tongue to produce new sounds.  However, as I pointed out, not all new phonemes require new physical skills.  Can we pick these up just by listening?  I think not, and I'd be happy to tell you how.

Meaningful sounds

The problem that I'm always trying to stress is that the brain is only interested in meaningful input -- if something has no meaning, the brain isn't interested.

This leads to some striking (and often unexpected) results. The BBC documentary Horizon showed this with colours in the programme Do You See What I See? (UK only). In the program, you see several Himba tribespeople trying to pick out different colours on a computer screen. The show two tests -- one with a very slightly different green, which is difficult for the viewer and fairly easy for the Himba, and one with an obviously different colour... well, obvious to us, but not to the Himba.

The distinctions that the Himba find easy are ones that they have names for, and the distinctions we find easy are the ones we have names for. It would appear that the act of naming something focuses the consciousness on it, so if you tell me that a French P has a puffy sound, I'm more likely to notice it, because I know what I'm looking for.

Consider the old face/vase optical illusion: the first time you look at it, you see either the faces or the vase, and your brain fixates on that single image. If someone else tells you about the other picture, you struggle to see it at first, because your brain already sees something meaningful in the image. But once your brain finally sees the second image, you can change your mental focus between the two meaningful images at will.

But that example doesn't say much about subjectivity and objectivity, because the two objects are fairly arbitrary. A better example would be one where you can predict what the viewer will see based on simple demographic information. Maybe adults vs children, like this painting, where adults immediately see a particular image and children see a different one. (View the picture, and then read the explanation on the page.  I saw the second picture without reading the explanation, but only because I could understand the French label on the bottle....)

So what is meaningful to us is normally a matter of past experience and expectation. When it comes to meaningful sounds, past experience and expectation all comes from the languages we already speak.  So it would follow that we need to consciously draw the student's attention to the differences, or they're just not likely to notice them.

What do we need to draw their attention to?

The phoneme is not the minimal unit of sound

The phoneme is often mistakenly considered the atomic unit of pronunciation in a language, but most languages build their phonemes out of a series of distinctions, in a fairly systematic manner.

In English, for example, we have voicing of consonants as a distinction, and it occurs pretty much wherever it can.  Voicing is the difference between P & B (at the front of the mouth), T & D (in the middle) and C/K & G (at the back).  We also have nasalisation, which takes those three pairs and gives us the sounds M, N and NG.  It's a stable and systematic structure.

There are other languages (EG Gaelic) where the distinction between P & B is not one of voicing, but aspiration.  The same distinction carries through for P&B and T&D.  In fact, it's hard to find any language that has a voicing distinction on one of those pairs, but makes a distinction in aspiration -- in general, the same distinction carries through.

Polish gives a great example of how regular these consonant distinctions can be.
In the diagram above, you can see a clear structure uniting 12 sounds in 3 distinctions (two 2-way distinctions, one 3-way distinction).  It's almost entirely systematic -- this cannot happen by accident, so we must assume that the native speaker's internal model of language acts on the level of these distinctions.

For this reason, I believe that it is not enough to draw the learner's attention to an individual phoneme, but that we must teach them the individual distinctions.

This doesn't have to be done in a dry "linguistics" way, though.

Teach once, then repeat

When teaching a phonemic distinction like voicing or aspiration, you don't need to start with the idea in the abstract.  Instead, you can start by teaching the pronunciation of one letter, then its contrast (eg P first, then B).  In teaching the contrast, you pick a word that describes it ("puffiness" or "breathiness" is more meaningful than "aspiration") or you just describe it.  Then when you move onto the next pair (T,D), you can refer back to the first pair, because it's the same difference.  And once you get to the final pair (K,G), it'll be very easy to do.

Of course, this means that you have to restrict the number of phonemes to start off with, but there are many people who are theoretically in favour of gradually introducing phonemes -- it's just the order of material that messes them up.

Teaching one thing at a time

Most teachers like to start with seemingly useful words and phrases.  Hello, how are you, goodbye -- that sort of thing.  This takes away the teacher's control over the phonemes -- teachers don't choose them, they just use whichever ones pop up.

Worse, quite a lot of teachers will introduce numbers early on, and in many languages you'll have encountered half of the phonemes of the language by the time you reach ten.  (This probably isn't an accident -- ambiguity in numbers would be a problem, so they naturally evolve to be fairly different.)

One commercial course points out this problem, and suggests that the way round it is to teach numbers one at a time, in a way which supports a progressive increase in the number of phonemes.  The example they used was 10 and 100 in Spanish: diez and cien.  These two words share all but one phoneme (C before I or E is pronounced the same as Z in Spanish), so if you teach one then the other, you're only introducing one phoneme the second time round. 
(I think I remember which course this was, but the blurb on the website no longer mentions this, so I'm not going to link to it.)

And after all, why should we teach numbers in numerical order in a second language?  When teaching children numbers in their first language, we're teaching both the concepts and the words, but in a second language you're only teaching the words, because they've already got the appropriate concepts to peg them to.  We can now selectively use any of those pegs we want to, in any order we want to.

Putting it together

So if we teach a couple of consonants well, and then we introduce new consonants one by one, we can use the earlier consonants as an anchor to show repeated distinctions.  It doesn't matter whether the student can consciously remember what those distinctions were -- a native speaker normally wouldn't have a clue.  What matters is that the model the student uses automatically for pronunciation implicitly respects the consistent rules of the language.

This will not happen if the student is left to listen, because one misheard phoneme can threaten the integrity of the entire structure -- pull any one of the sounds out of my neat little Polish diagram and dump it somewhere else and the whole thing will collapse.

Next time

Previously I spoke about sounds as new muscle movements, today I spoke about simply the meaning of sounds.  Next time, I'd like to demonstrate how almost all new sounds really are new physical movements anyway.

No comments: