Atheists like to argue that Christians need to begin their apologetics from a neutral starting point. That is, they must choose a position that starts with an agnostic approach toward Christianity. After that, evidence can then be brought forth.
There are several problems with this argument (one glaring one is the fact that there is no such thing as a netural starting point). The point that I am going to focus on for this particular post, however, is the fact that Scripturally, knowledge must presuppose revelation first. This can be demonstrated best by way of analogy.
Suppose that you create a bunch of semi-autonomous robots. These robots are designed for a specific purpose: to move boxes from point A to point B. There are, however, also boxes that, if moved, could damage the robots. Therefore, you decide to color these boxes with a red paint.  You then program the robots to know not to pick up red boxes.
The robots, however, revolt. They begin by destroying the sensor that enables them to see color. Therefore, every box looks the same: black.
Since you have provided for the robots the ability to procreate so you do not have to continually manufacture new ones, the robots reprogram themselves to destroy the sesors completely so new robot “offspring” will be created without the sensors too.
After a few robot generations, there are no robots around who remember the original state they were in. All robots are “born” without the ability to see colors. They assume that there is no such thing as color; that all the boxes are black.
These robots, however, can see that some boxes are harmful to them. So they device alternate rules about which boxes should be picked up. Some robots decide to be hedonistic: they’ll take whatever boxes they want because they’re going to wear out in the end anyway. Other robots become hermits that avoid every single possible box.
You, as the designer, can see that the robots are not functioning according to their original design. But because you don’t want to destroy the robots, you come up with an alternate idea.  You personally give an Instruction Manuel that explains what the problem was (how the original robots got rid of the color-sensor) and details how one can reinstall the hardware necessary to see the colors again.
Now, here’s the thing to consider. Since the robots have destoryed the ability to perceive the differences in the boxes themselves (i.e. they have no ability to see the color of the box), there is no way that the robots can know there is a difference in the color apart from the Instruction Manuel that tells them so.Â
Suppose that a robot did not think that he needed to know what color the box was. Suppose that this robot thought the entire concept of color was flat out absurd. Furthermore, suppose this robot does not want to take the trouble needed to build the sensor so he can see color for himself. This robot decides that another robot just “made up” the Instruction Manuel as a practical joke–or, even worse, as a way to trick guillible robots into picking up the boxes that this jokester wanted picked up! The robot decides that, since the Instruction Manuel cannot be verified by his current vision, it must be false.Â
Here’s the problem. If the robot wants to learn the truth of whether color exists, he cannot rely on his blindness! He must first believe the Instruction Manuel and then create the sensor. Then, he can see the color of the box. But the robot thinks his vision is perfectly fine without color! This robot sees no need to consider his viewpoint to be flawed. Because he sees no need, he isn’t going to look at the Instruction Manuel.
So, in this analogy we see that the objective truth of the matter is that you (the designer) have put red boxes into a world. The robots self-blinded themselves to avoid seeing it. Their descendents had no way to know that this self-blinding went on for they are born that way and assume it is just “natural” to be in that state. Therefore, you give them an Instruction Manuel in a form that they, in their blind state, can see–but it is impossible for them to know the veracity of it without first believing it and then doing the necessary steps to validate it.
In short, a robot cannot argue from the blindness of the robot and come to the conclusions of the Instruction Manuel. But this is exactly what the non-believing robots demand. They refuse to believe the Instruction Manuel unless it can first be proven in their own blinded state, not realizing that it is that very blinded state that prohibits them from seeing the truth of the Instruction Manuel.
Now suppose you are one of the robots, not the designer. Suppose that you believe the Instruction Manuel. Are you going to accept the theory that one must argue on the terms of the blind robots in order to convince them of the truth of the Instruction Manuel when it is only the Instruction Manuel that points out the problem in the first place? Neither, then, should the Christian try to argue from a “neutral” starting point that the Bible is true. The Christian must always declare the supremacy of revelation, because revelation is given to correct the flaws in reasoning that, due to sin, are not even perceived by the sinner. Revelation, therefore, must hold primacy over knowledge.
For this reason, we affirm with Augustine: “I do not know in order to believe; I believe in order to know.”