The Vernam Cipher key:
1. Must be random. 2. Must not be reused. 3. Must not be compromised. 4. Must be as long as the ciphertext.
Does plaintext injection into Vernam Cipher messages promise to reveal any important information about the key? If an adversary can get you to put certain words into your message, then does that present a problem? Does the collector gain any information about your key?
1. The ciphertext is captured complete
2. The language of the message is known (which reveals frequency of digraphs and trigraphs and grammar)
3. The message is not padded
4. The format of the message is known (along with headers and Russian Copulation)
5. The plaintext injection is certain to have occurred repeatedly over many messages
6. The amount of presumed key code is large
None of these factors will help break the key as long as the key is truly random.
But if the key is weak such as being a passage from a book, then the weak key can become readily apparent, and any of the factors listed above will just accelerate the exposure.