It's been quite fashionable to equate language change and evolution lately. In many ways language can be described as a living 'being', depending on your definition of life of course.
Personally though; I feel that this idea is fundamentally flawed. Sure up to some extent language change is very similar to evolution but not in all ways.
What is similar, is that both languages and organisms constantly change; and make offspring (for languages that is new people learning the language, whether a newborn, or a grown person).
Some languages seem to be 'fitter' than others, and therefore manage to spread further, and survive longer. But here there's already a very apparent difference between languages and organisms. Organisms are fitter because of their features, while the 'fitness' of languages is dependant on the fitness of the host, that is to say the speakers. Maybe of course the 'fitness' of a language is dependant on how easy it is to learn. But this doesn't seem to be right. By that reasoning we would all be Esperanto! Or any other 'easy' language. But the ease of a language is completely dependant one exposure, relatedness to the speaker's first language. And for infants there's absolutely no distinction. Every child learns their native language with about the same speed.
But to me the most fundamental difference is, the way change takes place. Evolution in organisms is something fairly random. Mutation happen randomly, and the results of such a mutation can either be good, bad, or neither. Whether such a change is good or bad, is completely random, we have no influence on it whatsoever.
Now for language change. These changes aren't truly random, the thrive towards an easier pronunciation, less energy. There's not many languages that will change *dw > erk. Simply because the latter is a lot harder to pronounce than the first. There's then a certain 'non-randomness' to such language changes. We pronounce an intervocalic [t] d]. Or an intervocalic [d] as a [ð]. Why? Because it takes less energy.
Besides such lenitions, sometimes there's also fortitions, but they are rather rare. But we could say those are done to make things clearer.
What I'm trying to say is, the start of a certain shift, is unpredictable and random; but there's reason behind what kind of shifts happen when. Especially with complete phoneme shifts, like the Armenian Consonant shift, and the Germanic Consonant shift, there's a clear interrelation between the shifts, the apparently influenced each other. This would be impossible if shifts were truly random.
Also, if we'd assume the position that changes are completely random. From a theoretical point of view, a language could start changing further and further, so far until at some point it has become so incredibly complex, that even the native speakers will no longer understand each other; and by this such a language would die out.
There's many reasons why languages die out, but a scenario as above is absolutely unattested. Thus I think we can not think of language change as something completely evolutionary.
Another point which is apparently an evolutionary thing, is the phenomenon of 'sprachbunds'. A sprachbund is a geographical region where languages, even when historically completely unrelated start taking over each other's features. A typical European sprachbund feature are voiced fricatives; which are quite rare outside of Europe. Typical of the Indian sprachbund is the existence of retroflex stops, which are not native to a language like Sanskrit.
This is similar to how some organisms can start to look very similar due to the same kind of environment, while not directly related. Recent studies, for example, have shown that the Large Flying Fox (Pteropus Vampyrus) is genetically more similar to apes than it is to bats. We explain this then that they lived in similar environments which made similar changes just as profitable.
The difference here, is though, that Sprachbunds could be seen as 'similar environments', but this is because it is the same environment. Equating these things, would be the same as assuming that all animals in the region where Bats evolved eventually all started looking like Bats, regardless whether it's a bacteria, a reptile or a mammal. That does sound absurd doesn't it?
We could also turn it around, let's say that sprachbunds, because they are often associated with regions, are in fact influenced by similar environments. By that reasoning we'd expect the High German sound shift to take place in similar mountainous regions too. Or even more extreme, that people's languages start sounding like High German in that region. That is of course equally ridiculous.
Therefore, I don't think there's enough of a base to assume that languages 'evolve', similar to how organisms evolve. Therefore thinking of language as a life form, symbiotic, or parasitic is very difficult to keep up.
I know that there's much space for discussion here, so please do.