Kant distinguishes between cognitions that are rational and those that are historical, and while this distinction does not feature much in the critical work, it is helpful to understand it to place Kant's critical thought into a larger picture. (Kant does discuss and make use of this distinction regularly in his logic.)
Since this distinction isn't spoken of much I don't need to worry so much about some inherited interpretation that must be challenged, however, the terms historical and rational have their own senses that may lead this distinction to be drawn with different instincts, so I will still need to take care to follow Kant.
(Note: Kant also calls rational cognition dogmatic.)
The Distinction
The distinction is made regarding cognitions taken generally, that is, all cognitions can be divided into either historical or rational cognitions.
When you consider a cognition and recognize that it's object is some particular objects, then you have a historical cognition. On the other hand, if the object of the cognition is not a particular, then you find a rational cognition.
The historical cognitions seemed to be called this because their objects are typically encountered by us directly, through memory. Rational cognitions, on the other hand, construct their object from concepts, and they do so only to the point required without placing the object into the world.
An illustration of an historical cognition is: "if we undermine the foundation of this house, then it will collapse." This example could be used to illustrate a rational cognition simply by replacing 'this house' with 'an house', giving us: "if we undermine the foundation of an house, then it will collapse."
Another example that illustrates a certain difficulty is the historical cognition described in the following: "that judgment is universal." Even though the object of the cognition is a universal judgment, the cognition about it still concerns some particular judgment.
One subject may take the description of a cognition as historical, while another may take it as rational. For example, "analysis supposes prior synthesis", may be taken as a statement about a particular analysis, but it could also be taken up as a statement about all analysis. This distinction isn't meant to favor either one of these interpretations, but only to discuss this particular laogical difference between them.
Phenomenological Demonstration
Find some object in the surrounding area and consider some of its characteristics. The relation to the object is as something that exists, and this would also be the case if you left the area so the object was no longer visible. If this object is material, then it also has mass and is under the influence of gravitation. This remains a historical cognition of the object, but if you consider that some object with mass will be influenced by gravitation you can notice the way this new object ('some object') relate to you.
When employing rational cognitions we don't have the same relationship posited between us and the object. We can consider all sort of objects that we stipulate, but which may or may not exist, and we can remain completely indifferent to their existence of non-existence.
(A writer or reader of fantasy may relate themselves to objects in their fantasy world historically, or treat of the system of science in that world and so also have rational cognitions about that world.)
Importance in Kant's Work
This distinction is not given much direct treatment in the critical works, but you can find it discussed and developed in his lectures on logic.
All cognition begins with historical cognition, and ultimately derives from it, but our ability to not deal with particulars allows for generalization into rational cognition. In order to apply rational cognition, one must once again return to historical cognition. This movement between historical and rational seems to provide some guidance in how to think about moving from experience to science and then back to engineering.