* I think that it is very likely that the Korean language family and the Japanese language family are related, even if it is challenging to find "smoking gun" evidence of it today. Japanese may have also have some Manchurian linguistic influence. The broader Altaic hypothesis has less strong support, but there may be something to it.
* I think that it is very likely that the Dravidian language family was influenced by an African language family, with the vectors of that transmission probably being people from the Horn of Africa who also brought some key African Sahel domesticates to Southern India around the time of the South Indian Neolithic ca. 2500 BCE.
* The Harappan language is almost surely not Indo-European, not Dravidian, and not Munda as a language family. It could conceivably have some connection to language isolates in the general region known as Indo-Pacific languages, or it might not. It is probably the main substrate influence on Sanskrit and through Sanskrit on the other Indo-European languages of India. The script associate with it was probably a proto-script, like a set of emojis or trademarks, and not a full written language. The same is true of the early Vinca script used in the Neolithic Balkans.
* I think that it is very likely that Indo-Aryans (Sanskrit speaking derived people) conquered almost all of India sometime in pre-history and imposing their language and the Hindu religion (although not as faithfully to some of its tenants like vegetarianism), except a small last stronghold, more or less in the vicinity of the modern city of Visakhapatnam, which then reconquered territory from the Indo-Aryans, restoring their dialect of the Dravidian language, but not effectively displacing the Hindu religion that the Indo-Aryan conquerors brought with them. This is why the Dravidian language family seems younger than it really is; it's historic linguistic diversity was wiped out at this point with most of its variants extinguished at this time. As I noted in a post at Wash Park Prophet:
[A]reas that are linguistically Indo-Aryan are more likely to be vegetarian than areas that are linguistically Dravidian, Munda or Tibeto-Burmese. Meat eating may reflect a thinner Indo-Aryan influence even in places that experienced a language shift to Indo-Aryan languages. Vegetarianism may alternatively reflect a stronger influence from the pre-Indo-Aryan Harappan society.
* Brahui, a Dravidian language pocket found far from the geographic range of the other Dravidian language, probably was not within the historic range of the Dravidian languages. Instead, it is probably a result of language shift through elite dominance around 1000 CE or so, by some foreign Dravidian warlord or king.
* Sometime around the Copper Age (a.k.a. the Eneolithic) in Anatolia, people from the eastern highlands brought the Hattic language (which preceded the Hittite language) to Anatolia. It is related to Kassite, other Iranian highland languages, and more remotely to most of the Caucasian languages (which are related to each other even if the connections are hard to establish), to Sumerian, and probably to Elamite. It is also probably related to Minoan. One of the litmus tests of all of these languages is that they were ergative.
Hattic probably replaced the Neolithic language(s) of Anatolia, including the Western Neolithic language which spread across Europe in two main branches, the Linear Pottery culture (LBK) to through the rivers of the north, and the Cardial Pottery culture to more or less along the Mediterranean coast, which was very different from Hattic. The Western Anatolian Neolithic languages were the substrate languages for the Indo-European language in most of Europe, but not in Anatolia where the Hattic language was the substrate. Hattic substrate influence is the reason that Anatolian Indo-European languages like Hittite seem so diverged from other Indo-European languages, because the Hattic society was much healthier when the Indo-Europeans arrived than in other places where the Indo-Europeans conquered Neolithic societies in a state of collapse. The most basil branch of Indo-European was probably that spoken in the Tarim Basin, which was on a frontier with almost no substrate influence.
* It is very likely that the languages of the European hunter-gatherers are completely lost. The Uralic languages arrived much later. In the Americas and Japan and Australia, we know that indigenous hunter-gather language substrates had very little impact on the food producing conquerer languages, even when indigenous peoples made a large genetic contribution to the people speaking the food producer languages.
* Basque, therefore, is very unlikely to be an indigenous European hunter-gatherer language. It could be the last survivor of the language family of the first farmers of Europe rooted in Western Anatolia frmo around 6000 BCE to 4000 BCE, or it could reflect a very distant outpost of a Copper Age language probably in the same language family as Hattic and Minoan. I probably lean towards the Neolithic hypothesis, as the corpus of Hattic (which remained a written liturgical language for a thousand years after the Hittites took over) and of Basque are both large enough that a connection would have been established by linguists by now if it was present, even though both are ergative languages, but the rarity of ergative languages outside the West Asian highlands, ancient Mesopotamia, and places to the east of that, favor a copper age origin for it. The Paleo-Hispanic languages may have all been a coherent group and Tartessos in Southwest Iberia was metal rich and a strong candidate for the source for Plato's Atlantis story. The "Tartessian culture was born around the 9th century B.C. as a result of hybridization between the Phoenician settlers and the local inhabitants. Scholars refer to the Tartessian culture as "a hybrid archaeological culture".
* We know the Etruscan, Raetic, and Lemnian (together called the Tyrsenian languages, an areal designation, since while the connection of Etruscan and Raetic is pretty solid, the linguistic family connection to Lemnian is not, and possibly Camunic as well, although it could also be related to Celtic) are also not Indo-European languages and pre-date Indo-European:
- Etruscan: 13,000 inscriptions, the overwhelming majority of which have been found in Italy; the oldest Etruscan inscription dates back to the 8th century BC, and the most recent one is dated to the 1st century AD.
- Raetic: 300 inscriptions, the overwhelming majority of which have been found in the Central Alps; the oldest Raetic inscription dates back to the 6th century BC.
- Lemnian: 2 inscriptions plus a small number of extremely fragmentary inscriptions; the oldest Lemnian inscription dates back to the late 6th century BC.
- Camunic: may be related to Raetic; about 170 inscriptions found in the Central Alps; the oldest Camunic inscriptions dates back to the 5th century BC.
The ergative substrate influence probably explains its presence in Indo-European Pashto, Kurdish languages and Indo-Aryan languages, which was shared with Basque and is absent from most Indo-European languages. It suggest that Harappan was probably ergative. The Tyresnian languages apparently non-ergative character suggests that they aren't part of the same language family as Basque, and tends to favor a Copper Age origin for Basque rather than a Neolithic origin for it.
But we haven't deciphered them very well since the corpus of those writings has mostly been lost, and what we have left is mostly monolingual and short. We can't even say with completge confidence that they were all in the same language family, although ancient Rhaetic spoken to the north of Etruscan (not linguistically related to the similarly named modern Indo-European minority language of Switzerland) was probably in the same language family with Etruscan. somewhat conflicting historical evidence suggests that Lemnians were migrants from the Alps and/or northern Italy, probably during the Greek dark ages after Bronze Age collapse had run its course.
We also don't know much about the substrate language that influenced Mycenaean Greek.