The whole programme is worth watching and appears to be available here.
The essential point is that colour names are merely names we apply to perceptions; and that our linguistic naming of colours actually determines what colours we see. In English, for example, we distinguish pink-purple-red-orange but have only one “green” despite a similar range:
Put another way, I remember arriving in South Africa as a 10-year-old and my first thought as we landed at Johannesburg, genuinely, was “The grass is yellow!”
This was a first indication that colours are literally different in Southern Africa, and it is more important to be able to distinguish between yellows and greens that, say, between greys and blues. Southern African languages would name their spectrum accordingly – with different words for various shades of green but quite possibly nothing at all for blue and relatively little for red – and this in turn means speakers of those languages literally distinguish those colours more easily in sight, but effectively only see one red and do not see blue at all (as evidenced in the linked article above).
Indo-European languages (such as English, Irish, French, Greek and Hindi) all derive from a single language, probably spoken in Ukraine around 5,000 years ago. To this day, there are some common markers derived from it, such as -r for family relationships (father/mother/brother; athair/mathair/braithair; pere/mere/frere and so on). Another point of interest, as noted in the article, it is impossible to reconstruct reliably a word for “blue”. This may give us some clues as to climate, location, geography and so on – in the same way we can reconstruct words for some types of tree and foliage, but not others.
Ultimately, what linguistics is doing here is given us a clear source of evidence about where we come from and how we perceived and perceive the world.
Give up languages in primary schools, you say? Crazy!