This is a guest post by Randall Buth. I thank him for permission to relay it to the readers of this blog. Randall interacts with this post in particular. The spur for his comments came from similar ones made by another Hebrew teacher, Sharon Alley, who happens to be his daughter. The situation reminds me of a quip which Markus Barth, son of Karl Barth, repeated on more than one occasion: "I and the father are not one."
Reading unvocalized texts is a great way for a student to discover what level of internalization they may already have. At the Biblical Language Center, we include some unvocalized texts for students to read to see 'where they are' in our first year materials. But we have come to conclude that it is a poor learning technique. We now think that the modern Hebrew Israeli ulpans are making a big mistake by using primarily unvocalized texts with early levels of Hebrew learners. We have noticed many adult learners reach fairly high communicable levels of modern Hebrew with a rather sloppy articulation of word shapes and vocalizations.
Modern ulpan teachers like to point out that Israeli children all learn from unvocalized texts. By and large that is true. But the comparison hides a problem. Israeli children already know how to correctly pronounce words they see written without vowels. They have already internalized the basic language and basic few thousand words of vocabulary. Not so the foreign learner. They typically try to read unvocalized texts with a background of only a few hundred words and find themselves spending long hours searching through dictionaries to get the vocalization correct -- or they just ignore the correct vocalization. It they can communicate with enough mistakes in a sentence they will not be corrected by fluent speakers. Who wants to correct three separate mis-vocalizations in one sentence? Those learners end up with a sloppy or skewed language, and it is attributed to their being 'foreigners' rather than poor pedagogy.
My strong recommendation is that beginning Hebrew students work with vocalized texts until they have internalized the basic morphological structures of the language. Even at that level, they would be greatly helped by having any new vocabulary vocalized. One possible approach would be to produce materials that assume a basic 2000-3000 word vocabulary, all vocalized at lower levels. Advanced material would only vocalize words that are not predictable by morphological pattern and not included in the first 2000 words.
I agree on the importance of being comfortable in reading unvocalized texts. If a student wants to learn biblical Hebrew, then they should not stop actively studying/ learning until they can comfortably read both a biblical Hebrew text, AND commentaries written on those texts in unvocalized Hebrew. Until they reach that level their language competence does not really pass muster. It falls short of a professional standard that would be expected in most any other literary field.
 Naturally, 'comfortably,' means texts that do not have inherent diachronological obscurities. There are many places where no one knows exactly what the Hebrew text is saying/said.
 I've met many a PhD whose language skills mis-produce and mis-predict biblical Hebrew, all the while claiming to 'control the grammar'. the biggest danger is when the Dr. is unaware of the mis-firings. A couple of thousand hours of use of the language would have provided this background, but our current pedagogies do not provide that for PhD students at any level. Their only practical alternative is to became fluent in modern Hebrew and conversant with the whole history of the language. Theoretically, this is also necessary because the biblical language is so limited (7000+ words) that fluency and natural use would be rather difficult to provide all the way to advanced levels without leaving the realm of BH-only.