I found this passage Sparks of Genius, a book on thinking tools, appealing. It describes the difficulty of abstraction.
Indeed, the essence of abstractions is that we say to ourselves, "My kid could do that." It is easy to forget that although we learn to use these abstractions, few of us could invent a new mathematics, discover a new law of nature, devise a new way of portraying perception, develop a new gestural language, or describe a fundamental truth about human feelings. Such triumphs are rare and difficult to achieve. Picasso commented repeatedly on how difficult it was from him to learn to draw simply and directly. He had to learn the process step by step. The notebooks of E.E. Cummings similarly show how hard he worked to achieve his simple effects. It is so much easier to see and convey the complexity and confusion of reality.
Indeed, abstracting is difficult for people in every discipline. Many famous novelists -- Mark Twain and Ernest Hemingway come to mind -- have written to their editors that they regretted the extreme length of their manuscripts; if they had had more time, the work would have been half as long. Winston Churchill is supposed to have said that he could talk for a day with five minutes' notice but needed a day to prepare if he had only five minutes in which to speak. The poet Edwin Arlington Robinson shifted from writing short verse to lengthy works as he got older, remarking, 'I am over sixty now and short poems require too much effort....' Similarly, when one reads a brilliantly simple paper in science, it is all too easy to pass it off as inconsequential. But as Harvard biologist George Wald once said to Nobel laureate Albert Szent-Gyorgyi after reading one of his typically lucid and simple papers, 'This paper of yours is so lightly written that you must have sweated terribly.'
I'd define it this way: abstraction is what is, essentially.