Indo-Aryan is spoken by ~74% of India, with Dravidian spoken by ~24% of India. So in practice the real-life language diversity is very small.
As for 'managing diversity', India has historically 'solved' this through a caste system, not a solution I would propose.
Source: I'm trained as a general linguist, but are not specialized in ethno- and sociolinguistics. Check out https://www.ethnologue.com for more information.
True, but old-school linguists tend to revel in this diversity, whereas more modern linguists try to find commonalities between these very diverse languages.
I remember auditing a linguistics class in college largely because I was interested in NLP. Sadly, I got the old-school professors who were more interested in students memorizing hundreds of variants of sounds than understanding the links between them. I dropped the course. Maybe if I stuck around long enough they would get to more global insights, but I was turned off by their entire approach.
That is clearly not adequate if you want to take a look at languages and culture (in that language). Classifications of languages as languages are partly political, sometimes even highly political, and mutual comprehensibility is not a working criterion.
I don't recall the details, but the classification used by ethnologue.com are a fairly reasonable set of soft criteria. The alternative would be not to speak of languages and dialects at all, and instead only about varieties of various language families with a certain degree of mutual comprehensibility to other varieties, but that would be even more counter-intuitive to laymen.
As to your personal experience: That's sad, you merely seem to have ended up in the wrong department / with the wrong professors. Find some general linguists and computational linguists and you should get your global insights. Phonology was actually at the forefront of general structural descriptions, sometimes even earlier than in syntax. For example, they invented optimality theory. It's worth giving it another try!
It was a while ago (graduated in 2005), but I think linguistics was under the English or anthropology department (I do remember the building was in the liberal sciences area). While I saw a clear connection between linguistics and NLP, I got the sense that the focus was to understand culture through language, and less about understanding language and communication itself.
> It's worth giving it another try!
I'd like to, but in the ~15 years since I've graduated it feels like the entire computational linguistics field has grown so dramatically that many of the problems I was originally interested in have been more-or-less developed (e.g., segmenting words into their phonemes for machine learning models, building grammar trees). Today, I'm more likely to grab one of the many libraries that do all this magic under the hood while I remain ignorant.
It's a bit of a tricky business, because what counts as a language is often also determined by political decisions. Mutual comprehensibility is not a reliable criterion. However, modern linguistics uses pretty good sets of "soft" criteria for the classifications.
Is this like saying Italian, French, Spanish are pretty much the same because they are Romamtic languages? Or English and German are interchangable, because English is a Germanic language?
Or, is much closer like British and American English- just mostly difference because of accents and dialects?
Trying to understand how different Indian subgroups of languages are.
It really varies. Some languages, like Hindi, Gujarati, and Punjabi, are about as similar as Italian, Portuguese, and Spanish. In other cases the common languages can be very different. Telugu and Tamil, for instance, are probably much more like trying to go between English and German. Telugu and Kannada are kind of odd ducks. They're basically Dravidian languages that borrowed tons of their vocabulary from an Indo-European language (Sanskrit). That actually makes it a very good analogue for English, which is a Germanic language that spent a lot of time trying desperately to be a Romance language (French).