According to Al-Ghazali, concepts are initially acquired through the apprehension of an object in the sensible realm. In fact, it is this direct apprehension which is regarded as the highest form of knowledge, presumably because it is the least motivated by subjective factors. This initial required contact with experience probably implies the impossibility of pure reasoning a priori, and any reasoning performed after this apprehension must therefore be motivated by previous experiences. This claim gives rise to serious implications if it is to be taken seriously, not only for epistemology but for the entire enterprise of rational philosophy.
According to Al-Ghazali, prior philosophers have been propelled towards their conclusions through the necessity of reason. They begin with certain premises, which, as noted, are likely to be influenced by subjective factors, and use the dictates of reason to disguise their subjective motivations. Al-Ghazali would thus propose that any efforts of reason, if they are to be sufficiently justified, should be supported by direct apprehension between the subject and the object to be known.
I realize that Avicenna might reply by arguing that it is precisely reason which allows us to go beyond experience to reach novel principles which can form the basis of metaphysics. The prior causes which lie out of reach dictate how we structure our ontology. These causes cannot be apprehended by a subject, and yet it is the capacity of rational beings which enable them to be deduced through a process of inference.
For example, we can deduce that the universe is beginingless. All existents must have a prior cause, and if the necessary first cause had a beginning, then it must have been necessitated by something prior, but this is absurd for then it would no longer be the first cause. We can therefore deduce from this reasoning that the necessary existent is eternal and therefore beginingless. From this we can also deduce that the universe is eternal, because if the universe had a beginning, a determining factor in the eternal must have necessitated its coming to be at one time rather than another. But positing either a change or an obstruction to the will of the necessary existent at one point in time would be impossible in a state of eternal recurrence. Thus, we have deduced that the universe is eternal, and further innovations in the sciences can now be formed on the basis of this conclusion.
Al-Ghazali could object to Avicenna, however. It may be noted that Avicenna can be commended for his rational enquiry, for it does indeed seem apparent that the eternity of the universe follows from the presupposition of the eternity of the necessary first cause. Your process of reason is infallible, and yet it is direct and obvious sense perception which makes your reasoning untenable. The very eternity of the world can be denied through direct observation of change: the fact that things are generated and destroyed implies change in the totality. But change is impossible in a universe that is eternal, unless you posit an infinite series of generation and destruction. But this too can be countered through observation. Consider the fact that we can observe that the sphere of the Sun completes its rotation in one year, and that the sphere of Saturn completes its rotation in thirty years. If the universe is eternal, then all rotations of the spheres are infinite, but we can directly apprehend that Saturn completes its rotation thirty times less than the Sun. This counters your premise since eternity would imply an infinite number of rotations throughout the totality of the universe, which has just been ruled out.
In conclusion, any efforts of reason should be pursued only insofar as the premises on which they are based can be verified through strict empirical means. We should reflect carefully before making our assent, keeping in the back of our minds the thought that it is likely to be influenced by our subjective inclinations, resulting in a process where even reason will not fail to act as a disguise for the passions.