I'm wondering if this should be a single tag. However, if we do keep them as separate tags, how do we disambiguate?
It looks to me like they are both used for similar questions, and based on the current Tag Info for the two tags they don't really appear to be different either. So, based on current tag usage, I'd argue that they should be combined into a single tag (which, in my opinion, should be
natural-language-processing because that's the full term that everyone in the field uses in my experience).
I suppose that, in theory,
natural-language could refer to something else than NLP... like, it could be for questions about language itself, rather than questions about processing (generating and/or understanding) language. I have a very difficult time imagining any such questions would actually be on-topic for AI though.
If I put quotes around the two terms and do a web search, the documents in the search results are very different, and the definitions of the terms don't match either. Natural language is mostly used as part of the social sciences, and NLP is mostly used in computer science and programming.
Natural language by itself (that is, without considering computation-related aspects) has little to do with AI. In AI, we want to do NLP, which can be based on natural language, but the tag natural-language-processing or nlp should also include these related discussions or questions. So, the tag natural-language should not really exist.
There is a key difference between the two terms. Whether this finer level of granularity is useful in the tags, I have no opinion.
Natural language is concerned with the general idea of conveying ideas via vocalization and the comprehension of the idea by a listener.
Natural language processing sounds more well defined, but it is actually poorly defined and the definitions in the literature are scattered between these two extremes:
- Parsing text into linguistic structures.
- Linguistic processing components in chat-bots designed to replace human experts.
What is included in NLP?
- Generating text?
- Generating linguistic associations?
- Parsing text?
- Topic detection?
- Story invocation? — See Schank.
Depends on who is teaching and when. I don't even see any consistency between the same person's view of NLP over time.
If addressing the terms literally, natural language is simply the field of linguistics minus the addition of formal languages. NLP would then become the action that occurs when some system deals with natural language at either its input, its output, or both.
I saw those two tags earlier today. I don't have a recommendation as to whether to combine them or leave them alone.