Acoustic & Language models independent?

[jason]

I'm using Sphinx for an experiment to see how different models achieve results
and their accuracy. One area that I'm not sure about is whether the acoustic
model is dependent on the source for the language model.

For example, I'm planning on creating a n-gram based on an english language
wikipedia dump. Are the words inside that n-gram dependent on the acoustic
model, or are they independent of each other?

[Nickolay V. Shmyrev]

Are the words inside that n-gram dependent on the acoustic model, or are
they independent of each other?

Acoustic and language models are connected by means of a dictionary. If
dictionary is compatible with both it will work.