Eliminating subvocalisation

[hedwards]

@Demonic, because there's no evidence to suggest that one should be reading out loud. Subvocalizing is a very, very bad habit that comes from reading out loud as a means of gaining literacy and the only reason we do it in English is because it makes it easier to tell if the person learning to read is actually reading the correct words on the page. Unfortunately, once you get in that habit you have to do a ton of reading very quickly to break the habit and very few people are willing to relearn how to read later on in order to break that habt.

 

From what I saw in China, they learn by writing the characters out a dozen times, looking at the definition and then repeating until it sticks. The connection with the written form is to a meaning and that meaning is ascribed to both the written and spoken forms.

[Demonic_Duck]

That sounds like an argument from ignorance to me: "because there's no evidence that reading aloud is beneficial, it must be harmful".

 

I'd argue that, even if subvocalisation is a net loss for one's L1 due to reading speed, it isn't necessarily a net loss for other languages, as it can help strengthen the ties between pronunciation and meaning. I'd also argue that this is especially important for a language such as Chinese, in which the relationship between written form and pronunciation is so weak, and as such subvocalisation results in a substantial net gain (again, even if it results in some loss in reading speed).

[hedwards]

@Demonic, sounds to me like you're arguing to tradition, because we've learned to read this way it's the right way of doing it.

 

Subvocalization is a net loss for basically all languages. It is literally impossible to read phoneme by phoneme in any language and produce spoken language that adheres to the typical conventions of intonation and stress. The reason for that is that these are frequently at the sentence level and are influenced by the neighboring phonemes. Chinese has its tone change rules, but all languages do things similar to that, there's probably a half dozen different ways to produce a t sound or a d sound in English and which one you use depends upon what's happening near by.

 

And there isn't a lack of evidence here, there's a lack of research here. Those are not the same thing. Educational research in general is typically of low quality and difficult to impossible to replicate. But we know that readers read is based on the shape of the words and the clusters of words. Fluent readers are taking in chunks of a sentence at a time, the question is how they use those chunks. If they take them as meaning, then you can much more quickly understand the sentence. With short sentences you can read them entirely in one chunk and move on to the next sentence. With longer sentences you generally read by the clause.

 

Subvocalizing requires you to read the sentence and then hold onto that information long enough to subvocalize it before reading the next chunk. This is a very inefficient process.

 

So, yes, it's possible that more research may change this, but the evidence is sufficient, IMHO, to just admit that there's a high degree of certainty that subvocalization is a problem. And this doesn't even factor in for the ambiguities that come from substituting homophones that aren't homonyms.

[realmayo]

hedwards, when you refer to "subvocalisation", are you referring to the sound "in your head" of the words you're reading? Because I'm struggling to see where is the "high degree of certainty that subvocalization is a problem".

 

Where is the evidence or even just reasoning behind your claim that "Subvocalizing requires you to read the sentence and then hold onto that information long enough to subvocalize it before reading the next chunk"?

 

Actually, I'm pretty sure you can't mean the same as what other people mean by the word. See http://en.wikipedia.org/wiki/Subvocalization -- is this the kind of thing you mean?

[hedwards]

@Realmayo, there's degrees of subvocalization. The most egregious cases involve the lips actually moving as one reads, and the most subtle forms are where the larynx is moving without producing any visible motion. But in all forms it involves taking written communication and processing it for sound and then processing the sound for meaning.

As for evidence, there is no other way of subvocalizing text. Your eyes work on entire words and phrases but the process of vocalization and subvocalization only works at the phoneme level. Which means that there has to be storage going on for each word and usually multiple words. Whether or not you subvocalize, you wind up having to read those chunks the same way, the difference is whether you slow yourself down by engaging in that pointless conversion process. When reading the correct way, you don't have to spend all that time waiting and burn through those glycogen stores, as soon as the brain has all the information you move on. It's how I can read so fast. Last I checked I have to go to nearly 700 wpm before I start missing words. That's with complete comprehension of what I'm reading.

 

It's not a matter of whether or not there's evidence, it's a matter of that's how reading works. Except with new words and characters that one has never encountered, you look at the entire word or character and decide the meaning. It's why beginning readers frequently read so much more slowly than more advanced readers. Advanced readers have already memorized the shape of the word and don't need to look that closely.

 

That's what I mean by subvocalization, but I'd strongly suggest against taking any of the information in that article seriously. It's based upon a lot of completely unsubstantiated ideas that fly in the face of observable reality. A word is just the shortest unit of meaning that stands on its own. If a word were a sound, then deaf people wouldn't have language at all as neither sign nor writing would have meaning to them. Sign languages wouldn't spontaneously generate in situations where people can't hear each other due to noise.

 

 

That's just my 2 cents, but I didn't spend a full year in graduate school getting my teaching degree to not understand how reading works. Things like this don't change much from language to language. Chinese is a language that's even more assertively anti-subvocalization. Why anybody would go through the work of learning to read characters when Pinyin would do just as well is beyond me.

[Rufus]

I have to agree with most of the comments here. Basically, reading out loud slows you down. Subvocalization also slows you down, but not as much as reading out loud. The truth is that most of us subvocalize in our native language and turning that off is the key to speed reading. 
 

As for reading out loud, this very aspect has been researched with no real conclusive empirical evidence. However, here is what the experts recommend...

 
Thanks for calling me out on this Demonic Duck. I say this because there is no overwhelming EMPIRICAL evidence, meaning that research studies in this area have been unable to back up a recommendation with hard data. That may be due to a variety of reasons (research design, difficulty of measurement, etc) but aside from that the general consensus among academics in the extensive reading community is to advocate silent reading. This is also known as "sustained silent reading" of which Stephen Krashen is a huge advocate. The recommendation is based off of decades of experience. If you'd like to read up a bit more, I'd invite you to check out the Guide to Extensive Reading put out by the ER Foundation.