A chosen plaintext attack is an attack where the cryptanalyst is able to define his own plaintext, feed it into the cipher, and analyze the resulting ciphertext.
Mounting a chosen plaintext attack requires the cryptanalyst to be able to send data of his choice into the device which is doing the encryption, and it requires the cryptanalyst to be able to view the output from the device. Because of these requirements, a chosen plaintext attack is in some cases impossible to attempt.
A good example is the attacks on the German Enigma machine — the Allies could not mount a chosen plaintext attack on the Enigma until we were able to capture one!
Got Something To Say: