If you're an untenured faculty member, you already know that one part of your process for achieving tenure involves "outside letters", that is, letters about your tenure case written by faculty outside of your institution. What else you may know about these letters seems to vary pretty widely.

It is relevant to what I say here that I work in the University of California system, and UC system policy is particularly lenient regarding who gets to see tenure letters. I read my own tenure letters (which were, technically, redacted—more on this below) and I have read the tenure letters submitted for all of my department colleagues who were up for tenure since I started at UC in 2012.¹ So, my experience reflects two departments (linguistics at UC Berkeley, linguistics at UC Santa Cruz) and cases across a range of subfields. I have also written a bunch of tenure letters over the last ten years, and in that process learned about aspects of how the process works at other places.

So, based on all that, and in honor of "letter writing season" (aka summer), here's how the sausage gets made.

A bunch of tenured people get asked to write a letter for your case; some of them say yes; most of those who say yes turn in the letters on time; those that are asked, say yes, and turn in their letters on time are the ones who write the letters that matter for your case.

There is variation in how the list of letter writers gets made. You really have to find out how your institution does this. Typically, the candidate makes a list, and someone else (the chair, or some subset of the faculty in the dept) makes a list, and the two lists get combined into a final list. Universities have rules about how many names have to be on the final list and what proportion of those names have to come from each side's starting list. In some places, names that are on both lists count as belonging to the candidate's list, and names on the candidate's list are limited to (say) 50%. This gives the candidate an incentive not to name [too many of] the obvious people, in the hope that the chair/dept list will name those people, and they will not count "against" the candidate. UC Berkeley has the enlightened policy of counting names that are on both lists as belonging to the chair/dept list, which means there is no reason for the candidate not to name the obvious names. Again: find out how your institution does this!

Once a final list is made, the people on the list are contacted by email and asked to write a letter for your case. It matters when in the season of such things (usually: spring/summer) this happens, because people are usually more responsive to the first such request than the fifth. (So if you, the candidate, have any sway over the timing of all this, try to get the lists in order as early as possible so that the emails can go out as early as possible. That makes it more likely that you will get the letter writers you want.)

(Should you tell people that you are putting down their names as a possible tenure letter writer? Or ask them if it's ok to do that? My advice is that you not do this. I, personally, would not find it weird to be asked, but multiple senior people have told me they would find it very uncomfortable to be asked. So I think the consensus is that asking is not an appropriate thing to do. [updated 6/17/26])

I mentioned that UC is notable in its willingness to let candidates read their tenure letters. UC policy is to "redact" the letters by cutting off the letterhead, the signature block, and anything below the signature block. Writers can say how they know the candidate below the signature block if they want that material redacted. I put "redact" in scare quotes for this whole process because this still leaves it totally obvious who wrote the letter in many if not most cases. This is linguistics, we are a tiny field. People have to try really hard to be anonymous in tenure letters, and insofar as they are writing a positive letter (as they mostly are—see below) they may not see the value in putting their effort into that.²

Private schools are much less forthcoming about tenure letters. I am told at some institutions the candidate is not even able to see any quotes from the letters for fear that this will de-anonymize the letter writers. So, again, find out what your institution does.

Now you might say: AR, you've just told us that your primary experience with tenure letters is in this extreme non-redaction situation where writers know the candidate is going to read their letter in (almost) its entirety. Doesn't this bias your overall sense of what is in the tenure letters? There is probably some truth to this, but I also think there is a major factor determining what is in tenure letters that holds true regardless of policies about redaction. Namely, requests to letter writers come with particular instructions, and much of what is in the letters depends on the instructions.

When someone is asked to write a letter for your case, they don't just get an email saying Do you think X should get tenure? Instead, they get a formal letter, on letterhead, that clarifies exactly what the university is asking them to do. The university has policies about what this letter is like. It always includes a request to assess the research, and if possible the other stuff, done by the candidate. (Outside colleagues typically don't know if someone is a good teacher or whether they are doing any service. So letter requests typically say: tell us about this if you know, but if not, no worries.) It nearly always includes a request to compare the candidate to other people. And it sometimes but not always includes a request to say whether or not the candidate would get tenure at the letter writer's institution, and/or to say whether, in the letter writer's opinion, the candidate should get tenure at the institution they are in fact at.

"Assess the research" is a vague ask. Some places are ok with that vagueness, and some attempt to make it more precise by asking about other nebulous properties such as excellence and quality. There is usually a question about "impact". Sometimes there is a question about "leadership". There's a lot of variation in how letter writers respond to all of this. Some will "assess" your research mostly just by summarizing it. (Everyone will summarize your research in some way, regardless of what else they may or may not add. Summaries of your research take up the vast majority of all text in the letters.) Some will respond to questions about impact by reciting your citation numbers from google scholar or some other source; some will talk about how important it is that X or Y paper responds to what you published; others will talk about the impact of your work on their own. Leadership, some writers will just make assertions about ("X is a leader") and some will just choose not to address.

Then we come to the dreaded comparison-to-other-people part. I am told that some institutions provide a specific list of names of other scholars that the writer is supposed to compare the candidate to, though I have personally never seen this.³ More commonly, it is up to the letter writer to decide the comparison class. Generally there is not a single obvious way to do this, so this is where things get strategic. If the writer thinks the case is weak, but still wants to write a positive letter, they will compare the candidate to people in the field who got tenure even with weak cases. Someone who writes a lot of letters has a mental file of such people. What they put in their letter basically is: sure this person has only published one article, but look at Assoc. Prof. Y, they only published one squib! On the flip side, someone who thinks the case is adequate/strong will compare the candidate to people in the field whose files are also generally adequate/strong, arguing that the candidate either holds their own among their distinguished peers or surpasses them.

And then the last parts: would this candidate get tenure at your institution? Should they get tenure at ours? This is where things get real in terms of what is actually going to happen with this letter, and the letter writer knows it. If the case is strong, no worries, the writer will come out and say that the person would get tenure at their (the writer’s) home department and should get tenure where they are. If the case is middling or weak, the writer will sometimes come out and say that the person should not get tenure, but more often will hedge or dodge the question. Would this person get tenure at your institution? is a question to which one can answer I don't know.

There is a reason Grice 1975 talks about academic letters as a place to find conversational implicature. People don't want to write a letter saying that Person X should not get tenure. They will avoid the question or be indirect in order to avoid saying this. They will strategically choose a weak comparison class to avoid saying this. Have I come to this conclusion just because my experience is coming from basically unredacted letters of the UC sort? Maybe, but I don't think that that’s all there is to it. When you write a tenure letter for someone, even if you are 100% sure they will never ever read it, you know what is at stake for them. If they don't get tenure, they will probably have to change professions and otherwise majorly reorient their life. This will be painful for them. If you think that they did not actually meet the standard for tenure,⁴ you feel a professional obligation not to lie about your assessment, but you also know that you are delivering a truth that a specific human (whose work you've just read) will be very unhappy to receive.

An untenured linguist at a private school once told me they were sad to learn that they would not be allowed to read their tenure letters, because they hoped the letters would contain a lot of useful feedback about their work. They absolutely will not. A tenure letter is not a journal review that says: I liked arguments 1, 2, and 4, and here are four problems with argument 3. Whether letter writers say anything at all about the problems with argument 3 depends on whether they think your case is strong or not. If they think it is strong, they might acknowledge some issue with argument 3 (though they also might not), setting this against the beauty, the elegance, the field-rocking transformative character of arguments 1, 2, and 4. The problem with argument 3 is a sentence or two at best—the risk of a negative implicature is far too high for the writer to really dig into why argument 3 is a problem. If they think your case is weak, but they still want to write a positive letter, they definitely will not get into argument 3; it's too much of a risk of spilling over into overall-negative territory. And lastly, in the case where the writer either recommends or implicature-recommends that you not get tenure, they are not going to make their case by going through why this one argument in this one specific paper is off the mark in some specific way. Maybe they would say: there are interesting things about this paper, but it has a serious flaw, argument 3. Here's one sentence about that flaw. The writer knows that the letters are being read not just by linguists but by deans and dean-ish people, and that those people do not give a damn about argument 3, but rather about "excellence", "quality", "leadership", "impact".

Beyond this, what the letters don't say is a question of what's in the instructions to the writers, and what sorts of implicatures those writers are aiming for or inadvertently generating. (You can be sure that inadvertent implicatures are something writers are worried about!)

...and I were untenured, I would: