You asked about maths and CS in particular. xLeitix wrote about CS; I am going to write about math from my perspective publishing in the field.
In math, there is much less of a need to cite "recent" work - although there is still some need, as I will explain below. In many areas of math (perhaps not all), the referee really can verify the arguments by logical reasoning, so references are less important to verify that the argument is correct. (Of course, if you use previous results, they will need to be cited.)
For this reason, citation practices in math are well known to be different than other fields. We publish less and have fewer citations on average than some other fields (so our journals have lower impact factors) and our citations are, on average, to "older" papers compared to other fields.
As a perhaps extreme example, I published a paper in 2010 with 14 references, of which 11 were published before the year 2000. The paper is in a respected, selective journal (an "A" journal in the Australian Math Society ranking). This paper is an outlier, though, compared to my other papers. My most recently accepted paper has 8 references: 1 is still a preprint, 2 were published in the last 4 years, 2 are from the 1990s, and 3 are from the 1970s. I don't think that is very far from normal in my area of mathematics.
When I referee papers in math, I look for references that:
Provide appropriate links to background material (these are particularly helpful for non-experts who read the paper).
Give appropriate credit for previous work.
Motivate the new work by showing how it relates to previous work. The 3 "newer" references in the paper from 2010 that I mentioned were exactly for this purpose. They showed how the problem we were studying had been posed by others, and how our work was related to published open questions in another area of mathematics.
When someone submits a paper to a selective math journal, the referees and editors will look for all these things. A paper that is full of brilliant technical results, but for which the editors can't see any motivation or interest, may end up being rejected because there are other papers that also have brilliant technical results, but which have clear motivation and are likely to be of interest to many others. Because journals have space limitations, correctness on its own is often not sufficient for a paper to be accepted.
At less selective math journals, correctness on its own may be the main criterion, but I would still expect a referee to comment on a under-referenced paper.
As for the situation in the question, it helps to remember that philosophy is generally focused on the types of problems that cannot be solved by mere logical reasoning. The same holds for many areas of the humanities, as well. In these fields, one cannot simply prove one's argument from commonly held axioms - the problems being studied are not amenable to simple logical analysis like mathematics problems. Each paper is viewed as a contribution to a discussion about the topic.
This leads to another key difference between citations in math and in some other fields. In mathematics, we usually try to cite the original source of an idea, to give credit to the first person to define or prove something. In other fields, the practice is instead to cite the most recent references on the idea, because they give a better representation on the current state of the discussion about the topic.