Because most people are somewhere in the middle in almost any way you define it. Most people need both time with other people and time alone. Most get tired when being with people all the time and their mental health goes down when being isolated.
Whether you derive energy from other people or whether they are draining is also dependent a lot on who those other people are and whether you have have social skills currently (e.g. how much effort it takes to you to figure out what to say).
There is no scientific foundation for it. Jung dreamt it up with an extremely literary approach. Then the swindlers Briggs and Myers popularised it with their 20th century astrology.
The introversion-extraversion personality axis is the only part of the MBTI that modern psychometrics accepts basically without reservation. It lives on as the introversion-extraversion factor of the Big 5 Personality Model.
I find it interesting that the rest of the MBTI is so thoroughly rejected -- in particular, I read in work that accepts the MBTI (uh-oh, red flag!) that males tend to split 75% T / 25% F and females do the reverse, 75% F / 25% T.
It seems to me that this very large difference in responses should be sufficient to show that the T / F axis is capturing something. If the T / F distinction were meaningless, it wouldn't be able to predict anything, but it clearly can predict the respondent's sex.
To add to watwut's comment - replacing a dynamically changing behavior with a permanent personality trait is detrimental to any attempts at changing things.
Whether you derive energy from other people or whether they are draining is also dependent a lot on who those other people are and whether you have have social skills currently (e.g. how much effort it takes to you to figure out what to say).