This is often true, but is it always true? HALOT goes on to list examples of clauses with an antecedent referent but without any resumptive pronoun (e.g. Gen 2:8; Deut 1:39; Jud 21:5). It says, "so any word relating to a preceding noun may be omitted and the sense of the omitted element is inferred from the the context."
But you have such a word, whether you realize it or not. It is ?$r!
Furthermore, while the reference to an antecedent can be omitted, where is it omitted when it is the object of a preposition.
I agree that bly is acting as a preposition, but then it must take an object... In your reading, it is just ?$r where ?$r is the reference to an antecedent. But where do you find that ?$r can reference the antecedent like this?
While ?$r can be properly translated by 'which' in many cases, this is only where 'which' introduces a subordinate clause. It does not translate 'which' when 'which' references just the antecedent. There may be border cases where it seems a reading says, for example, 'by which such and such', but the Hebrew phrase really reads 'by { which such and such }' whereby ?$r and the entire clause is the object of 'by'.
Furthermore, bly cannot take the object suffix forms of the 3rd person, which is why in that case bl(dy would have to be used.
This is why the form that properly references the antecedent is bl(dyw, even if bl(dyw is not attested by itself in BH. It is a reconstruction, but a very reasonable one.
Finally, ?$r is not always found immediately after its referent.
Archive powered by MHonArc 2.6.24.