I think that the negative effects of language death are overstated, and due at least four factors.
The first is presentism: we take a "snapshot" of the world as it appears now and assume that it is the natural, permanent order and the status quo must be preserved. When one reads documents like the European Charter for Regional and Minority Languages, it is clear that some non-majority or non-official languages are perceived to belong, whereas others are not. The reasons are normative and seem to relate to the concept I brought up in class of "squatters rights". English "belongs" in the United Kingdom because it has been there for hundreds of years, whereas Hindi has only been spoken in the UK for perhaps half a century, so it is not afforded any sort of protections. While this case may seem clear-cut, in other areas, it is highly prone to political manipulation. Based on history, should Turkish be a protected minority language in the Balkans or should Arabic be a protected minority language in Southern Spain? Obviously political contingents in both these settings would disagree. What about Romani? Do the Ronani people "belong" in Europe, given their long history but ultimately non-European origins? Should Tahitian be protected in Europe because it is spoken in French Polynesia?
The second factor, as I see it, is due to attributional biases. Due to technology, global scholarship, and a variety of other influences that make global communication possible, we are able to see linguistic shift as never before. Because we can observe it, we assume that we are in some way responsible. An environmental analogy may help here, borrowing from the discourse of ecolinguistics. Yes, in some cases, a species may go extinct due to explicit human interaction. The tree frog dies because its environment was destroyed, an artificial constraint on the species not due to natural selection. Yet the tree frog could just as easily go extinct because of a change in weather or other animals predation patterns, for instance. Or, it could evolve over generations (while we generally think of evolution as occurring over millions of years, if you go back and think about Darwin's finches it can occur over a much shorter time frame as well) to be the point where its evolutionary line may split or it may not even be considered the same species. It is perhaps worthwhile to note that scientific classifications of animals are also somewhat normative just like linguistic distinctions (though perhaps more grounded in hard science, i.e. DNA than linguistic distinctions). The difference between the moose and the elk has less to do with genetic differences than whether it resides in America or Europe. Similarly, whether one speaks Moldovan or Romanian has to do with what side of the border one lives on.
Ultimately, though, the ecological comparison fails. Language is a purely human phenomenon, so any human influence on language is not artificial in the same way as, say, a bulldozer in the Amazon rainforest is. Furthermore, this distinction between "human" and "natural" is a bit fuzzy. If we were to nuke the Yucatan peninsula and cause the extinction of an entire class of animals, we would say it was an artificial human occurrence. Yet that is effectively what happened 65 million years ago, killing the dinosaurs.
The third factor is a bias among linguistics to assume that language death is bad. Given what linguistics do, obviously language death is something to be feared. A linguist is no more likely to accept the fact that some languages should be allowed to die than a doctor is likely to accept the proposition that a patient with cancer should be refused treatment. Yet ultimately, the patient will die. Similarly, although linguistics can protect languages, language shift is inevitable.
Finally, the language that we use to describe language shifts has clearly negative connotations. "Language death," "linguistic genocide," and "linguistic social Darwinism" all have more-or-less unpleasant connotations. "Language extinction" is more clinical but still negative. I use the term "language shift" to reflect the fact that languages evolve over time, but this term fails to capture the fact that sometimes languages do abruptly seek to exist when the last speaker dies.
Another thing worth considering is the relationship between "language" as a living phenomenon - continuing an ecological metaphor - and language as a static concept - i.e. English, with a dictionary and a uniform set of rules. Codifying a language in essence removes it from its competitive environment of natural selection, akin to raising and animal in captivity. How does this relate to language shift? Does a language that is no longer spoken but still written down - for example Latin or Sanskrit - "die"? Is it still "alive"? Are we looking at the unliving skeletons of the Latin language, codified and preserved in museums, or is its essence still around?