Where is the boundary between written and spoken language?

The theme of my research in the master's program is the enhancement of reading in Japanese junior high school EFL learners.

The high school entrance exams, which most of those learners take, contains some reading tasks. They are sometimes an short, easy essay and sometimes a speech script. They usually seem to be easy (i.e. simple and plain in lexicon and syntax) for fluent English speakers to process automatically, or fast with little effort, no matter how they are told, through text or speech. After all, some of them (like speech scripts) are on the border of spoken language and written language of the easiest kind. In short, they are at the level of daily spoken language.

Now, how are they processed? Or, how should they be processed by the developing ESL readers?

I, based on the argument by Kadota on this book, am seeking for the possibility of utilizing phonological processing mechanism as a way to enhance reading performance.

I speculate that, if the text is as easy as spoken words that may well be uttered in daily conversation (as in the case of Japanese high school entrance exams), they may be easier for "comprehenders (readers or listeners)" to understand when spoken/read aloud by someone else, because the load from the decoding process is no longer there. If this should be the case, RWL (reading-while listening) is a god way to help learners to read texts faster. In addition, if they learn to read aloud fluently and use inner speech to read aloud to themselves, they can perform "silent RWL" for themselves, achieving greater reading speed in independent reading tasks.

I need to admit, however, there is an partly contradictory arugument that meaning units does not always matches phonological units--the mismatch of lexical and phonological chunks.

At a conference few days ago, I learned from listening researchers that 7 is the key number of syllables for sychronization of phonetic and lexical chunks. That is, lexical chunks longer than 7 syllables tend to be cut in the middle, probably for breathing of the speaker.

The latter position seem to counter-argue the first, saying RWL (and other phonological strategies/techniques for reading) has a limited benefit on syntactic processing of reading materials.

The two possible explanations for this is that

1) Phonological aid is available only for easy texts, most of whose components consists of short chunks of less than 7 syllables
2) Phonological aid helps only the decoding process.

However, where to put the chunk boundaries depends on the reader. Some may make short chunks.