Is your feature request related to a problem? Please describe.
One has to set the language manually in several widgets, if the corpus is not in English. Having a smart default, that would propagate downstream and set the language automatically would be neat (in Preprocess Text, Sentiment Analysis, Document Embedding).

Describe the solution you'd like
Once #582 is implemented, we could save this as a Corpus property and if present, set defaults in certain widgets to that language. If several languages are present, we could find smart ways of handling this (Document Embedding could embed with a different model for each document, stopwords could be removed for each document separately, etc.).

Describe alternatives you've considered
Setting it manually. 😆

Additional context
Widgets that should support this:

• Preprocess Text
• Document Embedding
• Sentiment Analysis
• ?