Because both may make you cry, but when cooked and digested, are delicious and nutritious.
I always get extremely vexed when a lot of textbooks and lectures introduce concepts for the first time so formally that they end up as mere functions to the listener, (maybe an attempt to make the science appear like a very difficult subject matter or a faithful following of convention), so much so that the soul or true usefulness of the concept is lost. Take for example a lecture in industrial electronics. The students become familiar with the industrial electronic components such as the SCRs, UJTs, thyristors, diacs, triacs, you name it. And, of course, they'll learn how they function as manipulators of power delivery and how to design the triggering circuits and resistors (like for the misleadingly named Programmable Unijunction Transistor or PUT). Fast forward a couple years to the future when they are now working in the industry. Say, that there is a series of loads supplied by a low power supply. The client wants all the loads to continue working even if a load breaks down. Then, from the student's schema on basic circuit analysis, he will just simply offer to place the loads in parallel. Unfortunately, the client needs a stable supply of current on each load, and driving the loads in parallel risks current instability, because it is impossible for 2 loads to have the exact same resistance to have a stable current flowing through it by Kirchoffs laws (not to mention the horrifying case that may arise when the loads have varying impedances).
What solution do you think the student will implement? Of course, the student will experience a little confusion at first, what book or lecture taught him how to deal with fixing a load in series that has failed? Then comes the brainstorming part. When the load in series fails, a potential will occur at its terminals. This potential must serve as a trigger to activate a short circuit over the two points. But what field of low power electronics could implement this kind of logic? Ah, a logic circuit. Why not design a logic circuit with a zener diode at its feedpoint (to regulate towards the logic treshold voltages) that will shorten the 2 terminals when a voltage is sensed? This, of course, is a pretty expensive solution. So what is the student missing? The soul. The usefulness of what he/she has learned from high power industrial electronic components. What if the student learns that the logic circuit above can be replaced by the lowest rated SCR and a few resistors (which will compose the triggering circuit)? A big difference in cost, since the number of components needed is significantly reduced. How can it do it? Simple. When a potential is exerted on the terminals, current will cross through the resistors towards the gate terminal of the SCR. If the resistors have been chosen appropriately so that the treshold voltage is reached at the gate, the SCR will fire and thus create a short circuit. SCRs also happen to be the perfect candidate for this kind of switching since it dissipates minimal power. But going back to our student, how could he/she realize that an SCR could do a better job if his/her mind has filtered it out as a component for control of power delivery? OK, my example may not be perfect for the point I'm trying to make (its the best among the few examples I can think of), but I believe it will suffice to prove that a student can't just rely on the definition from a textbook and lecture. An understanding of the concept must be reached so much so that the concept begins to be defined by what does and doesn't exist. Then, we'll find out that the uses of the concept we have learned is infinite.
I'm not saying that this thing is imperative and unavoidable. Even I don't pretty much know how to think this way, but I'm doing my best to put it to practice. (I'm also not saying that conventions don't give you this benefit, its just that the soul is not always directly implied since the way one comprehends a concept is always unique)