I don’t know Navajo so I’m not sure. It was explained to me by the professor who first told me about it as a matter of analogical extension, e.g., the battery might be conceived of as the “heart” because it powers the car like the heart powers the body, etc.
Well, okay, but my point is that no language is actually pure, in the sense of not having taken on any foreign words or influences. There are, of course, dead languages to which new (foreign or native) terms aren’t added because there’s no one around to need them, but as far as natural spoken living languages are concerned, no language is pure. It’s a matter of degree as to how each preserves assumed (proto)-_____ features (morphology, phonology, etc.), but that certainly doesn’t lead us in the direction of saying “this language is pure because it has preserved X”. By way of illustration from phonology, we can look to the Indo-European set of unvoiced stops, which are provided in the Guide to Appendix I of the 2006 edition of the American Heritage Dictionary as follows:
I.E. → p t k kw (kw = labialized k; the w should be a superscript)
From that set of sounds, the following correspondences have been established (+ = dead language):
Hittite+ → p t k ku
Tocharian+ → p t/c/ts k/ś k/ś
Lithuanian → p t š k
Greek → p t k p/t/k
Latin+ → p t c qu
Old Irish+ → – t c c “–” = p was lost in Old Irish]
Old English+ → f th h hw
There were many more languages in the table, but above is a representative sample of some of the sound changes that occurred in the divergence from (P)I-E in a number of its descendants. It is important to note that Hittite is said to have likely gone extinct by 1100 BC (the last records stop ~ 1200 BC), and Tocharian (a branch with two or three languages, rather than a single language) by 9th century AD, so you can measure a sort of relative rate of change between languages this way, i.e. Hittite had preserved the majority of the I-E sound correspondences, and died out early on, preventing any further changes. Tocharian, which went extinct some several hundred years later, shows considerably more variation (where I-E had t, Tocharian had sometimes t, sometimes c, sometimes ts). And so on and so forth, throughout the ages. The point is: Even at the level of individual sounds (and so certainly at the level of word formation), there is no stability or purity, only differing rates of change. I am not a historical linguist, but if I had to guess I would say that the situation in Icelandic is at least helped by the geography of the place. In the same way that islands produce a baffling array of flora and fauna not found in other environments (e.g., Madagascar), languages that are literally cut off from foreign influence also develop or retain archaic features in surprising ways. It might have also helped that the Icelanders were under Norse and Danish rule for so long. The Scandinavian languages form a dialect continuum similar to any other group of related languages (e.g., the South Slavic dialect continuum in the Balkans), which gives their language an insular quality: Their nearest neighbors are Scandinavians, just like they are, so most of the new words that enter their language would likely enter via Norse or Danish, which share a common origin with Icelandic, meaning that there is less phonological “re-tooling” that needs to happen to create an “Icelandic” form of whatever the word is. So you have loan words that look and behave like native words and might not be obvious to anyone who doesn’t know their history or the mechanisms of borrowing. Wikipedia has an interesting example of one such word:
Sapir and Zuckermann (2008) demonstrate how Icelandic camouflages many English words by means of phonosemantic matching. For example, the Icelandic-looking word
eyðni, meaning “AIDS”, is a phonosemantic match of the English acronym AIDS, using the pre-existent Icelandic verb
eyða (“to destroy”) and the Icelandic nominal suffix -
ni. (
source)