This is interesting to me, and I've thought about it a little bit, but I have no special knowledge or training in this area.
I've discussed this with one other person, and that person argues that thought is a linguistic operation -- linguistic in the broad sense, meaning the manipulation of symbols to which meaning has been ascribed. So in this view, the act of ascribing meaning is not thought, but the use of those symbols so endowed to express ideas, ask questions, draw conclusions, etc., would be thought.
This fits pretty tidily with my world view, but it is still difficult to completely accept, simply because I don't have any other sort of experience to test it against. I can't remember a time I was without language. One thing I have noticed is how readily a new way of thinking can push out an old one: When I was young and had just learned to read well, I had the experience, for a time, of all my thoughts appearing in my head as words. I visualized these words as black text on a white background. When this was happening, I couldn't imagine that thinking had ever been different, even though I knew, logically, that only a few months before I wouldn't have understood the words. So I wonder if this equating of thought with linguistic phenomena is similar: something people can accept because they can't imagine anything else.
In order to investigate this further, it might be useful to examine the reasoning ability of people with various types of aphasia, and of people who never learned to communicate with language (e.g., Hellen Keller types, or feral children). Certain drugs (e.g., LSD) can render someone temporarily aphasic.