Tolkien was livid. He made his opinion clear in a related letter to his British publisher, where he said, "I have many Jewish friends, and should regret giving any colour to the notion that I subscribed to the wholly pernicious and unscientific race-doctrine."1
In his reply to the German publisher, he feigns innocence at first. "I regret that I am not clear as to what you intend by arisch. I am not of Aryan extraction: that is Indo-iranian; as far as I am aware none of my ancestors spoke Hindustani, Persian, Gypsy, or any related dialects."2
Tolkien had no truck at all with the Aryan Fallacy, but many of his contemporaries embraced it enthusiastically, including several other fantasy writers active at a similar time — very notably Robert E. Howard. So how did it come about?
The word Aryan correctly refers to a group of nomadic tribes that moved into northern India about three thousand years ago. It's also used for the linguistic family descended from Sanskrit, deriving ultimately from the language(s) spoken by these tribes. Lastly, and with extreme caution, it can be applied to the people who speak these modern languages (most of the languages of Pakistan, Bangladesh, Nepal, northern and central India, and the majority language of Sri Lanka). As long as it's understood that this in no way represents a racial identity.
The family, as Tolkien indicated, also includes Romani ("Gypsy"), its original speakers having migrated from north-west India. Ironically, in Hitler's time (i.e. before the mass migrations from the sub-continent after World War Two) the only substantial ethnic group in Europe who could legitimately call themselves Aryan were the Romani — who were persecuted by the Nazis.
The most obviously related group of languages is the Iranian family, together forming the Indo-Iranian group (Iran clearly derives from the same root as Aryan). Today, the Iranian group covers most languages of Iran and Afghanistan, together with Kurdish, but in earlier historical periods Iranian-speaking nomads lived on the steppes north of the Black and Caspian Seas: Scythians, Sarmatians, Alans etc.
In the late 18th century, linguists began to recognise a clear relationship between the earliest forms of Sanskrit, Greek and Latin, and postulated that they might all have derived from a common source. With this leverage to build on, the Germanic, Celtic, Slavonic and other language groups were added to the growing super-family, nowadays known as Indo-European. An extinct family of Indo-European languages, Tocharian, was even spoken in north-west China during the Han era.
In the 19th century, it was proposed that the whole family should be called Aryan, in the erroneous belief that this was the earliest name of a people speaking an Indo-European language and therefore the most likely to be the original. Though mistaken, this was originally innocuous, but a tide of racial supremacism gradually saw the label become more and more equated with the "German Race". Max Müller, one of its originators, was later scathing about colleagues who confused linguistic and racial characteristics, suggesting that "Aryan race, Aryan blood, Aryan eyes and hair" were as absurd as "a dolichocephalic dictionary or a brachycephalic grammar".3
The tide was against him, though, and the concept of the Aryan Master Race became more and more widespread, eventually finding its lunatic home in the warped mind of Adolf Hitler. In fact, there have been suggestions that the characteristic differences between the Germanic languages (including English) and all other Indo-European languages may have been the result of an unrelated people abandoning their own language and taking up a broken form of Indo-European. In which case, the Germans have even less right to the name Aryan than most other Indo-European speakers.
The clear pattern of divergence among the languages through various periods has always suggested that it should be possible to trace them all back to a time when a single language, from which they are all descended, was spoken by a community in a relatively small area. The favourite explanation nowadays is the steppes of south-eastern Europe somewhere between 4500 and 2500 BC, but other propositions have ranged from the eccentric (such as the North Pole) to the more plausible (such as Anatolia or the north-western European plains). The image to the left shows one proposed model of expansion from the homeland.
It shouldn't be assumed, though, that these original Indo-Europeans represented a race, or even a "people" as we'd understand it. Archaeological evidence suggests that the communities who probably spoke Indo-European were actually made up of several distinct groups — a people who'd come down from the north, another who'd come up from the Mediterranean, and possibly others from Central Asia or the Caucasus — who can all be seen from their remains to have been very different physical types.
Nor are they likely to have had much in common, including a political or tribal structure, except the language which allowed ideas and customs to spread. They constituted a culture area, but no more.
In terms of the later spread of Indo-European, too, we can't assume any racial connection. It's not unusual for peoples to adopt the language of either the dominant or the "cool" culture, and this often means communities speaking a language aren't racially connected with the those who spoke the language's distant ancestor. If we didn't accept this, we'd have a hard time trying to explain how an English-speaking African American could have derived from the language's source in north-western Europe.
The subsequent evolution of the languages is likewise anything but straightforward. Linguists use the family tree model, (eg Latin is the "parent" of French, Spanish, Italian etc) and this is useful, and a gorgeously artistic interpretation of which by Minna Sundberg is shown right. But the influence of other, often unrelated languages is important too. This can be down to extensive borrowing of vocabulary for social or political reasons, such as how English, fundamentally a Western Germanic language, has a vocabulary heavily derived from Latin.4 It can also be explained, though, by the wave theory of language change.
The wave theory is a model which suggests that specific changes, whether a sound-change such as a final t changing to s or a grammatical change such as the development of grammatical gender, spreads from an epicentre and affects both related and unrelated languages. The next major change will have a different epicentre, and/or the speakers will have moved, so it won't be the same set of languages affected.
Ultimately, this will create a patchwork, in which it can be difficult to reconcile a language's position on its family tree with apparent similarities to more distantly related (or unrelated) ones. It's inconvenient, but we're talking about human behaviour. What do you expect?
So where did the original Indo-European language come from? If it was being spoken a mere five or six thousand years ago, it obviously can't have sprung from nowhere. The reason it's recognised as the "original" is that it marks the latest point that all Indo-European languages can be traced back to, but its history must have gone back a long way up an unknown family tree.
Various propositions have been made as to what Indo-European might ultimately be related to, the most widespread being a super-family known as Nostratic. At its most ambitious, this hypothesis includes Uralic, Altaic, Kartvelian, Afroasiatic, Elamo-Dravidian and Eskimo-Aleut, along with a few others, though some more cautious proponents restrict it to the first three plus Indo-European.5
There's very little evidence for hypotheses such as this, and what has been put forward is strongly disputed. The problem is that it becomes progressively harder to be sure of connections between languages the further back the proposed connection is. It's probable that some, at least, of these connections are correct, but nothing's likely to ever go beyond speculation.
To speculate, though, how far might this process actually go? Recent evidence suggests an origin for language considerably further back than was believed even a decade or two ago, certainly when our ancestors were living in a fairly small area of Africa.6 Maybe the invention of human language really was a single event, and we're all speaking variants of the same language.
The Aryan Fallacy should have died in the bunker with the Führer, but unfortunately bigots aren't famous for their intellectual rigour, and there are still white supremacist morons who use it as a keystone for their disgusting creeds.
So take a leaf out of Tolkien's book. Next time a white supremacist proudly claims to be Aryan, point out that they're actually claiming to be Indian. Or even Iranian. You won't change their views, but see how they like it.
1 The Letters of J.R.R. Tolkien, ed. Humphrey Carpenter, 1981, p.37
2 Ibid. p.37
3 Quoted in In Search of the Indo-Europeans, J.P. Mallory, 1989, p.269
4 That the roots of English are Germanic can be illustrated by taking a passage of average English and identifying where each word came from. It's likely that more would be Latin-derived than Germanic-derived; but if you were to redo the exercise counting each word each time it's used, you'd find the count overwhelmingly Germanic. This is because the most common, infrastructure words (the, and, to, for and the like) are direct descendants of the words that came over with the Anglo-Saxons.
5 Uralic includes languages like Finnish and Hungarian; Altaic includes the Turkic languages, Mongolian, probably Korean and possibly Japanese; Kartvelian is a small group whose main member is Georgian; Afroasiatic is a huge group, ranging from Hebrew and Arabic to Hausa, the main language of northern Nigeria, and taking in ancient Egyptian along the way; Dravidian was the main language group in India before the coming of the Aryans, and is still dominant in the south; Eskimo-Aleut — yes, I know Eskimo is non gratis, but there's actually no other word for the whole linguistic group, the Inuit actually being just the largest ethnic group.
6 Remains have indicated that Neanderthals shared the same deformity of the larynx that allows us to manipulate complex sounds, suggesting that the mutation took place at least 150,000 years ago. On the "use it or lose it" principle of evolution, it's difficult to interpret this any way other than our ancestors already using language from that point.
Images reproduced under Creative Commons licence:
Dancers of the Shuvani Romani Kumpania: James Niland
Indo-European expansion according to the Kurgan hypothesis: Dbachmann
Minna Sundberg Language Tree: Tom Wigley