An economy-based versus richness-based account of non-overtness of subjects in Arabic subject relatives

In many languages, called pro-drop languages (e.g. Modern Standard Arabic, Irish, and Hebrew), the pre-verbal or post-verbal subject position is not filled. In these languages, the subject position within the relative clauses is left phonologically vacant. I assume that pro-drop phenomenon in matrix clauses and the non-overtness of the subject position within relative clauses are related and they can be given a unified account. The main argument for leaving the subject position unfilled was explained along the lines of agreement and richness, or in other terms, licensing and identification. I reject the accounts of null subjects within relatives based on these arguments, and argue that an economy-based approach to null subjects can provide better explanation of Modern Standard Arabic data