Skoyles is proposing a procedure which can make an attempt at translating letters to sound. Any such procedure must be learned by the child from examples of print words and their pronunciation. Such learning requires a degree of knowledge about print words and letters which the child is unlikely to have without experience in reading. The relation between phonemic awareness and reading skill is clear but I'm not sure that the direction of influence has been decided. I favour the idea that the two skills feed off each other, as opposed to a straight causal relationship.
2. Skoyles is proposing a procedure which can make an attempt at translating letters to sound. I claim that any such procedure must be learned by the child from examples of print words and their pronunciations (or possibly from letters and their names/sounds). Any such learning requires a degree of knowledge about print words and letters which the child is unlikely to have without at least a little experience in reading print words. As an example, the child needs to know that words are made up of letters and that the order of the letters is important before any associations between letters and sounds can be made. The same applies for any orthographic unit one might choose, in fact, larger units might require more knowledge from the learner.
3. I agree that an effective letter-to-sound (and I use that term rather than grapheme-to-phoneme on purpose) procedure need not get all of the words correct. As a supplement to other cues, some phonetic information would be useful in word identification. So a simple phonological reading route can begin to be useful long before it reaches an acceptable level of accuracy.
4. Now to Skoyles's discussion of word frequencies. The statistics Skoyles quotes on the number of words with significant frequency are not very far removed from those for adult texts (from memory). The corpus cited (Joam and Share, 1983) would have around 600 words occurring more than five times and presumably around 100 occurring more than 10 times. Skoyles suggests that this is not enough to allow the child to read these words logographically: what logographic reading mechanism would Skoyles propose that requires so many exposures? Since logographic reading presumably involves storing a visual representation of the word, it seems reasonable to assume that the child could store such a representation after just a few exposures, although it might need to be reinforced to become permanent. (Actually this is one of the areas I'm a little vague on, there doesn't seem to be much information available on this question -- pointers would be appreciated.)
5. My own simulation of a logographic (I call it visual) lexicon (Cassidy, 1990) is only able to cope with a small number of words (a few hundred) without producing a large number of errors. This is because it is organised in a way that does not take advantage of any features of printed words -- because at this stage the child hasn't had the chance to learn them. A later revision of the visual lexicon might be more efficient, incorporating things learned from visual and perhaps phonological reading. The important point is that the first stage shouldn't have to incorporate any special knowledge about words before the child has had a chance to learn them.
6. Skoyles suggests that the use of a phonological procedure would provide readers with a reading lexicon as large as their oral one. Well one of the most telling facts about early reading is that children do not have such a lexicon. Reading errors, at the first stage, tend to be words drawn from a reading lexicon -- words that have been experienced in print before (Seymour & Elder, 1986). If a phonological procedure were in place, such an effect would be very unlikely, barring perhaps a strong episodic constraint on responses to print words. Later, children start to produce word responses from outside their reading experience; this is usually seen as an indicator of the use of a phonologically based naming procedure.
7. Skoyles's final point on phonemic awareness is something I have been considering recently. The relation between phonemic awareness and reading skill is clear but I'm not sure that the direction of influence has been decided. I tend to favour the idea that the two skills feed off each other, as opposed to a straight causal relationship. My own theory (PhD Thesis, in preparation) suggests that further development of the logographic lexicon depends on the successful development of a phonological route, which in turn relies on phonemic segmentation skills.
Cassidy, S. (1990) Substitution Errors in a Computer Model of Early Reading. Paper presented at the First Conference of the Australasian Society for Cognitive Science, Sydney, 1990.
Cassidy, S. (1992) Bootstrapping the child into reading: Is the first reading process phonological or visual? PSYCOLOQUY 126.96.36.199.
Joam, A. F. & Share, D. L. (1983). Phonological recoding and reading acquisition. Applied Psycholinguisitics, 4, 103-147.
Seymour, P.H.K. and Elder L. (1986) Beginning reading without phonology. Cognitive Neuropsychology, 3, 1, 1-36
Skoyles, J.R. (1991) Connectionism, Reading and the Limits of Cognition. PSYCOLOQUY 2.8.4.
Skoyles, J. R. (1992) Not All Phonological Reading Need Use Accurate Letter-Sound Rules: Reply to Cassidy. PSYCOLOQUY 188.8.131.52.