Real life classification problems

While it is highly unlikely that there will ever be a need to classify cow sculptures as animals instead of artwork, there are some current real world classification dilemmas occurring. 

In this article linguists anticipated finding a group of people in a remote part of India speaking the language known as Aka, but instead they stumbled upon a new language known as Koro. The language sounded completely foreign to them, yet they proceeded to look for similarities between it and Aka. After realizing it was unrelated to Aka, they compared its sounds and syntax to hundreds of languages in an attempt to identify it.

Turning to a different realm of science, this article discusses the discovery of a new cat-sized carnivore in Madagascar. From its appearance, it resembles a brown-tailed mongoose, but once a closer look was taken, it was found to be unrelated and was given the name Durrell’s vontsira.

Reading these two articles made me wonder whether the researchers approached the problem of classification in a similar vein as we did with cartoon penguins. Do the rules change when it's a 'hidden' language that's discovered or a completely new animal? When do enough new discoveries cause an existing classification system to be modified? Maybe the only way to initially comprehend something that's utterly new is to find any possible similarity between it and something that already exists, but it is also easier to fit something into an already existing classification system. Since the population of both discoveries is small, there exists a desire to preserve both, but what happens to the classification system when an element goes extinct?