Sanskrit and Hindi are both part of the Indo-European language family and share, if one looks closely enough, many of its distinctive features. In the case of Hindi, the language has been so tossed about in the seas of history, that many shared features have been torn from its deck and swallowed by the flood of change or weathered into a nearly unrecognisable state. Concerning such shared features, many of the articles in this blog will throw some light on the matter. The interested reader may turn to: Father in Greek and Sanskrit: looking is not tastingVowel gradation 101Wolf and wolves and foot and feet or Father sky in ancient Greek and Sanskrit. But the present article is about the one feature of Sanskrit and Hindi, and indeed all modern North Indian languages sharing a common ancestor in Sanskrit, which stands out as being characteristically non-Indo-European: the retroflex phonemes ट् (ṭ), ठ् (ṭh), ड् (ḍ) and ढ् (ḍh), produced by touching the roof of the mouth with the lower part of the tip of the tongue.

Using comparative linguistics no such phoneme can be reconstructed for the mother-language of all Indo-European languages and none of the Sanskrit roots, the reflexes of which are also found across the board in the root inventory of the Indo-European languages, contain such a phoneme. In other instances Sanskrit has, by the same comparative technique, been shown to have retained an original feature of the mother-tongue, which has been simplified, by the loss of one of its articulatory features, in all other languages. The group of sounds concerned are the voiced aspirate stops घ् (gh), ध् (dh) and भ् (bh), which are, just as the retroflex phonemes, so distinctively Indian. By a rigorous comparison of the sound inventories of the Indo-European languages, these sounds have been proven to have been part of the original sound inventory of the Indo-European mother-tongue. How is this possible? By comparison of cognates, i.e. words having the same origin. One example will suffice to show how by such a comparison it is relatively easy to deduce which language has retained the most original form of a sound. The widely attested Indo-European verb for to carry will here serve as an example. It is भरामि (bharāmi) in Sanskrit, φερω (phérō) in Greek, beran in Old English (c.f. modern English to bear) and ferō in Latin. What we are concerned with here, is the initial consonant. On one point all languages agree: the initial consonant in a labial, i.e. a sound produced with the lips. But they disagree about the exact mode of articulation: Is it a stop (produced by total occlusion of the mouth) or a fricative (caused by partial occlusion of the mouth)? Latin is the only language that has a fricative, all other languages have a stop, so it is very likely that the original form of the sound was a stop of some sort and that Latin has independently developed it into a fricative by lessening the tension in the articulatory muscles, as it is not plausible that all other languages independently agreed to change the fricative to a stop. If it was a stop, then what kind of stop was it? In the examples above there are three different kinds of stops. In Sanskrit it is aspirated (pronounced while breathing out) and voiced (pronounced with active vocal cords), in Greek too, it is aspirated, but it lacks voice, in Old English, and all the Germanic languages, it has voice but lacks aspiration. What do we make of this? How do we pick the most original sound? Three of the languages indicate that the original sound must have featured aspiration: Sanskrit, Greek and Latin. The Germanic form alone lacks the aspiration. This is easily explained by the change of one of two articulatory features, a common thing in the development languages across the globe. This is a preliminary indication that aspiration is probably an original feature, as it seems to be dominant. If we take aspiration as a feature developed by alteration of an original sound, then we will again be faced with the difficulty of needing to explain how three languages independently developed the exact same feature. What about voice? Was the sound originally voiced or not? Sanskrit and Old English speak for voice, while Greek and Latin voice their opinion against this. If Sanskrit independently developed voice, then the original form might have been unvoiced, but the Germanic languages unanimously speak for voice, agreeing with Sanskrit, making it less likely that voice was a non-essential feature of the original sound. This line of argument would also require the Germanic languages totally to change the nature of the sound by altering both of its articulatory parameters from voiceless-aspirated to voiced-unaspirated, not the most likely of developments if one considers that the development of sounds usually takes, as all things in nature, the path of least resistance. Taking the voiced aspirate of Sanskrit as the original sound all the other sounds can be reached with as little resistance as possible, while at the same time taking into account that both aspiration and voice seem to have been dominant features of the original sound. This is how the initial sound bh was reconstructed for the Indo-European verb to carry. By the same comparative technique no retroflex of Sanskrit has ever been able to be proven to be an original sound that has later merged with the corresponding dental phoneme containing the same parameters of articulation.

Since there is no trace that this type of sound is part of the original phonetical stock of the Indo-European languages, it must therefore be part of the stock of another language, with which Sanskrit came into close enough contact slowly to assimilate some of the features carried in its sap. Without doubt this stock must be the Dravidian language family, in which the phoneme is a traceably original feature and for which there is ample evidence that the presence of its various representatives in the Indian subcontinent precedes that of Sanskrit. The main intention of this article is to point out that this process, by which a migrating language is grafted on the stock of a native language and thus may adopt some of its features, is by no means uncommon and finds parallels throughout the history of civilisation, which is essentially a story of us being on the move.

A considerable number of the world’s languages are today spoken on and, in our mind, linked by deep-reaching roots to a soil from which they ultimately have not sprung. English is the prime example. What is the native country of English? Well, England, if names mean anything at all. Many more countries, where English is a so called native language, will readily spring to mind. But, historically speaking, the language is neither native to the British Isles, a fact often forgotten because the Germanic migrations, which brought it there from the north of Germany, go very far back in time, nor, as we all know from our history classes, is it native to the American continent, where it was first introduced by intrepid Elizabethans looking for an exciting new life in a new world, or, for that matter, to Australia, New Zealand and South Africa. In all these countries the so familiar language English has grown on stocks belonging, for the most part, to entirely unrelated and now often nearly forgotten linguistic species, such as Maori in New Zealand, the multitude of Aboriginal languages of Australia, the Native American languages of North America and the Celtic languages of the British Isles. The influence of these languages on English is very difficult to measure, primarily because English itself has had such a dramatic history of successive colonisations, during which it has probably undergone more change than any other language on earth, but also because for the most part, apart from the initial Germanic colonisation of England, the contact happened at such a late stage in the development of English that it was no longer prone to extensive influence reaching beyond a word here and there.

English is not alone in being a language that grew up in exile. Spanish and French have, just as English, grown on originally celtic soil. And, surprisingly, even all of the great classical languages have deep cultural links with a country in which they originally were immigrants. Latin has displaced an earlier, non-Indo-European language called Etruscan, about which very little is known. Greek speaking people and their characteristic culture only appeared in Greece after 2000 BC. Before their arrival, Greece was inhabited by another, also non-Indo-European speaking people, whose only traces are fragments of pottery and a distinctive non-Indo-European suffix –nthos adopted by Greek as part of equally foreign words. This suffix features especially in names for plants and animals, e.g. sminthos, the mouse, or hyakinthos, hyacinth, but also in the names of towns, e.g. Korinthos, Corinth, and even in the epic word for bathtub: asaminthos. Finally, Hittite, the oldest of all Indo-European languages as far as its attestation in writing goes, contains a large number of non-Indo-European words in its vocabulary, which must have been growing on the soil of Anatolia long before the mighty chariot-riding immigrants first set foot on it and slowly started assimilating them into their own very different language. There is no reason to doubt that Sanskrit adopted the non-Indo-European retroflex phonemes in exactly the same way as Hittite adopted many non-Indo-European words, or as Greek adopted a distinctively non-Indo-European suffix attached to distinctively non-Indo-European words, or as English amassed the staggering 800,000 odd words of its vocabulary: mementoes of the great road trip of civilisation.


Silvio Zinsstag,

teacher for ancient languages