To send an email I have to have money in my stamp box. I get that money by paying the post office for e-postage. They give me a code which I enter into my email program.
When I send email each item takes a few cents from the stamp box, calls the post office computer to get a "stamp", they agree on the ammount and the stamp is put in the header of my mail.
I fire up my email client, it gets my mail, extracts the "stamps" verifies them with the computer at the post office which "cancels" them. Any mail with no stamp goes in a bucket for a few days (User preference) and is then either trashed or sent back to the sender with "NO STAMP -Not Read by recipient", The email would explain how to get a client for the new system. If the email is sent back it may be reflected by the recieving system- this is detected and the returned email is just trashed to avoid loops
Mail sent with a fake stamp would be fraud, counterfieting but would be treated as mail with no stamp by the mail client
If an email does have a stamp - half the value goes into my stamp box, the other half goes to the post office for the use of their computer. (yes the post office makes a bit to sell the stamp and makes a bit when I read the mail)
I could also have a whitelist - when email is recieved from someone on my whitelist it is placed in my inbox and the post office is not bothered. The whole value of the stamp is (A- returned to sender or B- retained in my stamp box which do you think would be safer for fraud?)
For the tinfoil chapeu crowd - If the sender is on my white list the Post office is not notified at all and the email is as secure as email is today. If not well I suppose the post office could figgure out from canceled stamps who is sending email to whom.