I must decide [which English course to take].
Preliminary point: there's no such thing as a noun clause. The classification of finite subordinate clauses is based on their internal form rather than spurious analogies with the parts of speech.
In your example, the bracketed infinitival clause is a subordinate interrogative clause (embedded question) ...
10 books [to read], subject to change.
The bracketed expression is an infinitival relative clause modifying "books".
Such clauses have a modal meaning similar to that expressed in finite relatives by "can" or "should"
Your example is thus comparable to 10 book that can/should be read.
Subject to change is a set expression ...
In order to understand what Flesch is trying to say, you should first of all realize that grammar isn't his focus, or even interest. For example, on pp. 71-72, he rewrites
 American belles-lettres also give a much more faithful and adequate picture of the entire civilization to which they belong than literature abroad.
 American belles-lettres ...
In the simplest terms, the second “to” adds emphasis and, often, separation. That emphasis is used to enhance the context.
I want you to verb1 and to verb2 indicates two distinct actions. I want you to wash the car and to dig the garden. The context should indicate why the two actions are, or need to be, distinct.
I want you to verb1 and verb2 indicates ...
You heard me [sing].
"Hear" is a catenative verb, and "sing" is an infinitival subordinate clause functioning as its catenative complement.
"Me" is the syntactic object of "hear", and the semantic subject of the subordinate clause.
In a comment John Lawler wrote:
Why do you assume only one can be correct? That's usually the wrong way to bet, unless there's an ESL textbook in the woodpile. Both are in fact grammatical and colloquial, and both mean the same thing. The infinitive to reverse a string will be interpreted as a purpose infinitive, and the for with the gerund is the normal ...