Has there any research been done on how difficult certain languages are to learn for chatbots? For example, CleverBot knows a bit of Dutch, German, Finnish and French, so there are clearly chatbots that speak other languages than English. (English is still her best language, but that is because she speaks that most often)
I would imagine that a logical constructed language, like lobjan, would be easier to learn than a natural language, like English, for example.
It's not the language itself but rather the structure and for the language's ambiguity, for example in English: person a says "John and Bob (his fish)" person B says "He died!", posed question, to whom does person B refer to by he died. More than the language, but the application. You can write a chat bot in Assembly, C, C++, C#, Java or Python. The all work a bit differently but can accomplish the same result, but one language might have more pros or cons to the other. So it will boil down to not language but the understanding of what is being said, research of language in the brain has come to confirm we associate a meaning/feeling/and other inputs with a given language.
So to conclude: English is by far the most chaotic for a chat bot but Japanese is actually the best due to the way the language itself is written/spoken. There is more structure to it and less ambiguity.
I'm a Software Engineer and An AI Researcher for the past 7 years.
The easiest language to learn by a bot is the language that has the most training data available. Sure, language structure matters, but when there are multiple order of magnitude differences in the size of training data, then that overwhelms everything else.
One should expect English, French or Chinese to get much better results than any constructed language even if (especially if?) the system includes no language-specific tuning.
