I'd like to continue the Cocoon's Job Market Boot Camp by discussing recommendation letters. Although the role that recommendation letters should play in hiring is debatable--as letters may be more indicative of letter-writers biases, etc., than good reflections of candidates' abilities--I have many reasons to believe that having good letters is really important. First, after my PhD, it was one of the things that just about everybody impressed upon me. I was told, basically by everyone, "you have to get letters from well-respected people outside of your department." Second, although it is very difficult to tease out causal influences, there was a distinct trend in my own case. The more outside letters I had, the more interviews I seemed to get.
But I am already getting ahead of myself. There are two very different types of situations: (1) the ABD student who has not yet graduated, and (2) the person with the PhD who has been out on the market a while. Let's discuss each situation in turn.
1. Letters of recommendation and the ABD student
As I understand it, it is okay to only have letters from your PhD program if you have not yet graduated, as it is admittedly hard to network while in grad school. That being said, the students I knew in grad school who had letters from people outside the department seemed to fare fantastically well--far better than those who only had letters from within. So, I would definitely advise grad students to try to network early and often, and try to obtain at least one outside letter before hitting the market. I'll say more about how to do this in section 3 below.
One other good thing to do--or so I've heard (more on this shortly)--is to have your department placement-director read and evaluate your letters. I have to be honest. I never did this. However, I've heard of cases where a candidate had a lukewarm letter from someone, and this (apparently) is not good. In the US, at least, it is supposed expected that all of your letters will be stellar, and one lukewarm letter may sink your candidacy.
Third, unless you have no teaching experience at all, you should always have at least one teaching letter written by someone who has actually observed you teach. How many teaching letters do you need? I never procured more than one, and it never seemed to hurt me with teaching oriented schools. As long as the teaching letter is good and by someone who has actually seen you teach, that seems (from my experience) to be good enough.
Anything I missed?
2. Letters of recommendation after the PhD
Once you've completed your PhD, there seem to be two important things to stay on top of: (1) ensuring that your letters are updated every year or two, and (2) obtaining more letters from respected people in the profession. Let me begin with the first of these.
I've been told that it is important to keep your letters up to date, and this pretty much stands to reason. If you've been out on the market for a few years and have accomplished stuff--published some good pieces of work--it would presumably be to your advantage for your letter writers to talk up these accomplishments and your overall development as a philosopher. How do you go about it? Simple: just send an email to your letter writers seeing how things are going with them, and asking if they would mind updating your letter. In my experience, they will usually do so happily, asking for an updated CV and a few pieces of recent work. I might also add that it probably helps if you stay in touch with them independently, checking in with them a few times a year to say hello, etc.
It's also important to keep your teaching letter(s) up to date. If you've been out on the market for a few years, in a VAP position, post-doc, whatever, have someone in a senior position at your current institution watch you teach and write a letter.
I've only had one person respond negatively to a letter-update request, and because my experience here contradicts some of the received advice I gave above (on lukewarm letters), I figured it might be helpful to share my experience. After a couple of years out on the market, when I asked one of my recommenders if they would be willing to update their letter, the response I received was basically, "I don't know how much my letter would help you at this point"--the obvious implication being that the letter would be lukewarm at best (note: I didn't blame this person at all. My dissertation wasn't spectacular by any means, I hadn't published much at that point, and I don't think I ever much impressed them). Anyway, I thanked them for their honesty and proceeded that year without their letter...and got like no interviews. The next year, I asked the same person if I could use their own (lukewarm?) letter, they said yes, and I got a lot more interviews. So, who knows, maybe a lukewarm letter isn't a job-market death sentence, after all.
Finally, then, we come to the big question: how do you go about getting outside letters?
3. Networking and outside letters
I'll be honest about something. For most of my life, I looked at "networking" as sort of distasteful. It always seemed to me like sucking up or brown-nosing, or whatever. I now believe that this is the wrong attitude to have. Although I am still no fan of being a sycophant, crude "social-climber", or whatever (and I recognize that not everyone finds those things so distasteful), I think it is entirely possible to network effectively and get letters without such motives. Early on after my PhD, I didn't network very effectively. Then I just started talking to people at conferences, without any real motive except to get over my shyness and try to be more open! I also met people in other ways. For instance, I applied and was accepted t0 a NEH summer institute in political philosophy, and simply had a great time. The organizers were really great guys, and so at the end of the seminar I simply asked them if they'd be willing to write a letter. They did!
A couple of my other letters came about in similarly unsought, serendipitous ways. When I published my paper on free will, the journal gave me the option of giving ten "colleagues" free access to the publication. Since I hardly knew anyone in the discipline at that point (seriously!), I put in the email of someone who had been a previous professor of mine long ago, whose work I always admired, and whose work actually inspired my paper a bit. I thought, "Hmm...I wonder what they would think of this piece?" Lo and beyond, a couple months later, I received an email from them saying they liked the article. It just so happens then that I had written a response paper to one of their other articles, so I asked if they would read it. They gave me some great feedback, was really encouraging, we exchanged some other stuff, and I've considered them a friend ever since. At some point later on, I asked if they would write a letter for the job-market, and they said yes! Another one of my letters came about in a similar war.
So, I guess that's what I would say about getting outside letters. Some people are cool with being "networkers", and such. I used to keep to myself because I didn't feel cool with that (it always felt sort of icky to me!). Fortunately, though, as I've said above, I think there are non-icky ways to go about it. You don't have to go "looking for letters" at all. Just try to make friends, engage with other people's work and have conversations with them, etc., and letters of recommendation might just be an unexpected (and unsought!) positive result. After you know someone a bit, and have a good relationship with them, if it seems appropriate to ask for a letter, ask away. The worst they can do is say no, and in my experience, this is pretty uncommon. More common, in my experience, are qualified yes answers--with the person I asked saying, "Well, I don't really work in your AOS, but I could write you a letter in terms of what I think of your work." For what it's worth, anytime anyone offered me a qualified yes of this sort, I jumped at the opportunity. It seemed to work well, at least in term of job-market results.
Anyway, these are my thoughts. Have I missed anything? Gotten anything wrong? Fire away!