Do native speakers of languages using logographic writing internally vocalize while reading?

Question

I study Chinese as a second language. Listening is my strong point.

When I read, I go through this process:

see character  imagine myself speaking the pronunciation  think of meaning (from the sound)

• The second step is perhaps described as speaking without actually opening the mouth to form the words and requires about the same amount of time as actually speaking aloud.
• The English translation only comes in when I encounter newly learned words.
• It is as if I am relying on my listening ability to read.

The second step seems really inefficient. With practice, I could change to this, but I am not sure it is ideal to skip the middle step:

see character  think of meaning (from the shapes of the characters)

I wonder which is a more natural approach that a native speaker will take to read.

• Do adult native speakers of a language that uses a logographic writing system, such as Chinese, similarly internally "speak" the pronunciation or do they skip directly to the meaning? Or use some other way?
• Might this 3-step process be a normal stage a child acquiring this as their L1 would initially use, but later develop away from?

Answer 1

The paper shared by snailboat is a great survey of the issues. But all it does is really show how complex the issue is.

There is no one clear meaning to vocalization. Also, there is not just one type of reader. There will be a difference between fluent experienced readers, beginning readers, struggling readers, reluctant readers, levels of education, etc. Also, there is no one 'Chinese'. So there are multiple questions to answer here.

1. Do readers of logographic languages like Mandarin Chinese (which is pretty much the only such language available for investigation) subvocalize at the level of the phonological representation in a way that readers of languages with alphabetic scripts are apt to do?

2. Do readers of Mandarin vocalize the words at all or do they just form direct mental representations of what they see written? In other words, is reading a Chinese character more like seeing a symbol (e.g. toilet sign with man or woman, traffic light or arrow) or is it more like reading a word.

3. What is the variation between different types of readers depending on how they acquired the language. What other dimensions of variation are there?

There are no definitive answers available. I would look at some of these for further investigation.

• There is likely to be a mix of cognitive processes involved in the reading of Chinese characters. (Just like there is in reading English.)

• The first thing to take into account is the fact that all people who have learned to read Chinese in the last 60 years or so will have done it through the medium of Pinyin which relies on deconstructing the words into their phonological representations. (Taiwan uses Bopomofo - a similar non-Latin based phonological system.)

• Many people do the majority of their character writing on computers and phones - and therefore through Pinyin so the phonology of words is re-emphasized. In fact, Victor Mair just published an interesting blogpost on how many educated people are forgetting how to write even simple characters.

• However, despite the strong linkage of characters to sound, there are at least some aspects of writing that do not correspond to any vocalization. E.g. 他 = he 她 = she but tā = he/she. So the reader sees the person is a male or female but the hearer will not hear that distinction. And there are many visual word/character puns in Chinese as well that cannot be vocalized.

• There is some research that seems to hint that there is less incidence of dyslexia in Chinese (dyslexia is a disorder of phonological processing) or rather the problems a dyslexic brain encounters with a Latin-based orthography don't become apparent in Chinese - but that may be more related to issues of identification.

• Many readers of the Chinese script do not actually speak Mandarin so may vocalize the characters differently - although I suspect that this issue is a lot more complicated than just saying the characters make communication between different languages possible.