For newer sources: Academic discussions build upon each other. While the point might have been made in 1968 it was probably expanded in 1997 and quoting the 1968 one would exclude the 1997 authors reflections which might have influenced the writer. It also invites more recent discussion and excludes the skipping over of more recent opinions. This increase the likelihood that your paper will be included in academic discussion. It presents yourself as being up to date with the latest research and able to posit a new line of inquiry previously not attempted.

For older sources: this is quite straightforward, translations are always open to interpretation. If you've got a 1st-century source written in completely legible modern English you could use it as a primary source. But I extremely doubt it. Academics also work in increments and using a non-published, self-made translation of the 1st-century text would need a separate academic discussion about its validity. To discuss a certain topic it is better to use a well justified translation by someone else than to combine two completely different studies into one single paper.

This

a 2001 book, when discussing point X, references a 1997 book making point X, which, in turn, turns out to be a reference to a 1968 book discussing the same point X

is lit review. This is 100% of the first chapter of my dissertation. This is there to show the different approaches historians have taken to a problem, and how positions evolved or have been discounted, and how my approach supplements/corrects/is different from earlier works. Here's a concrete example:

On the causes of the Social War of 91-88 BCE, Mommsen took an approach that understood the Italian rebels as a sort of proto-nationalist movement that reflected the nationalist sentiments of his own time (1860-something). This nationalist frame, Mouritsen (1996) argues, was adopted consciously or unconsciously by every scholar who wrote on the Social War since Mommsen. Mouritsen wanted to move away from the nationalist frame as anachronistic. Mouritsen builds a chain from modern scholarship to what he perceived as the error introduced by Mommsen to show how modern approaches, in his opinion, miss the point.

It can help to build credibility by showing that other people agree with you, and helps share and acknowledge ideas. When researching a topic you tend to read supporting material about it, which you often go on to credit and reference in your own. Also depending on where you are primarily sources might not be readily available.

Sometimes point X is a matter of serious discussion in its own right. If the reader is interested in aspects of point X, it is useful to have multiple mentions of point X. Put it down to generosity and a desire to give credit where credit is due. Also, there are times when the anecdote mentioned is not as readily available as the author might hope. If the editor's deadline is bearing down and interlibrary loan has not come back with the original first century book, the author has little choice other than to cite the citer.

I'll give you an example. There is a frequently repeated story that Ottoman Sultan Bayezid II welcomed the Jews expelled from Spain in 1492 with the comment " those who say that Ferdinand and Isabella are wise are indeed fools; for he gives me, his enemy, his national treasure, the Jews." Or words to that effect. Great quote, I used it myself in my book on the 1565 siege of Malta, but my cite is secondary material because despite years of searching, I have never found a contemporary source. (BTW, if anyone reading this has one, I would love to hear from you.)