Intellectually, you want to go to the program with the best advisor who will help guide your research. Working with a superstar has the obvious benefits of letters of recommendation for fellowships and jobs in the future. The cache of being affiliated with a superstar faculty may also help at conferences and other gatherings.
In the real world, however, your one faculty mentor can turn out to be a huge pain. They can leave the university for greener pastures at another department that swoops in and offers to double their salary. They can show no interest in you once you arrive on campus. Heck, they can get hit by a bus or start having serious health problems. Clearly, these aren't fun things to ponder, but they are the realities of working with and relying on others.
Though your most important relationship in graduate school will be with your graduate advisor (and this is true in the sciences, social sciences and especially the humanities) you will have the opportunity to work with numerous other faculty in your time in graduate school. It is to your benefit to share some research interests with them as well.
Let me give you a couple of examples. The most obvious example is the simple fact that you will be asked to put together committees of faculty for your qualifying exams. These faculty can behave professionally, explaining to you clearly what will be expected to you - or they can choose to be a thorn in your side for a year or half year. They can tailor your reading lists to your interests or make you read arcane texts from the 1950s. Before that, you will probably take a series of courses in your department. Though you can typically take these courses from people outside of your research interests - you might as well take these classes with interesting people, right?
Finally, you will be asked to put together a dissertation committee later in your career. For our institution this consists of at least a couple of faculty in addition to your advisor.
The point here is that there are a number of dangers on relying on just one faculty member in graduate school. What if they go on sabbatical to Antarctica? What if they get sick? What if they retire? The better option is to look for a program with a number of faculty who will work with you over the course of your time in graduate school. If you can find a program with 3-4 people you'd like to work with instead of 1-2 - you're probably in a little bit better shape.
