R is a really interesting sound. Or rather, it's a letter that has an interesting range of possible pronunciations. It can be rolled or glided in the back of the throat like in some dialects of German, trilled in the front like in various Spanish dialects, or pronounced as a fairly tight, fluid liquid up against either the palate or alveolar ridge, like in English. In some Asian languages, R and L aren't well defined, and so speakers of those languages may have difficulty pronouncing them distinctly. What results is hybrid sound that sounds a little too much like R when it should be L, and a little too much like L when it should be R. So people often think Asians are "swapping" L and R, like they're just confused or something, when this really isn't the case. English speakers have a difficult time rolling Spanish R's in the front or German R's in the back. Some French and German speakers have a difficult time pronouncing the tight alveolar/palatal liquid R's that English speakers use pervasively. And yet, for as difficult a sound as R can be, it pops up all over the damn place.
From 2005 to 2010, I worked for the Indiana University Center for Postsecondary Research, an office based out of the School of Education that runs that National Survey of Student Engagement. The office regularly employs graduate students who work together on teams that handle an alphabetical subset of the hundreds of colleges that participate in the survey process each year. One of these graduate students for a couple of years was a guy with the last name Kuykendall. Now this name might normally be pronounced something like "kigh-ken-dall" or "koy-ken-dall", but this particular guy's family pronounced it "ker-ken-dall", and that made me incredibly curious, so of course I had to ask where his family was from: Gary, Indiana. Interestingly, that dialect wouldn't necessarily convert an /er/ sound to /oi/ (some accents in New England might pronounce something like "burger" as "boygah"), but what it might do, is take correct "oy" sounds and hypercorrect them to "er" interpretations, since hyper-corrective r-insertion is something that can happen when rhotic speakers (those who pronounce their R's in these contexts) hear and interpret non-rhotic pronunciations (where the R has, in contrast, been dropped) in contexts where there could either be an omitted R or none at all. The result? "Kuykendall" ends hyper-rhoticized and pronounced as "Kerkendall".
R-insertion like this doesn't really seem to happen with other letters. I mean, there are cases in which other letters or sounds might be falsely interpreted to exist somewhere because of a mistake or misunderstanding of some sort. For example, an assistant manager at an old movie theater job that I had in the 90s wrote the word "outfit" as "alfit" one time in a note to staff regarding Halloween costumes. But that wasn't an insertion of L to satisfy some pronunciation constraint as much as it was just a mishearing of the word "outfit" being spoken with southern English: often pronounced "ow-fit" or something like that. After all, L and T are surprisingly similar sounds. If you close your L the rest of the way instead of letting air glide over your tongue, you end up with a T. Not a huge leap to assume a slack pronunciation of "outfit" is actually something that sounds like "alfit" if you haven't seen the word written for some reason or another.1 But again, this was purely a misunderstanding, and not a linguistic choice to modify a pronunciation to fit a phonetic context.
So why R in particular? Maybe it's because for as difficult as imitating specific types of R can be for various speakers of dialects, the muscular mechanics behind actually producing the kind of R-insertions we're talking about are relatively easy. ("Easy" in the sense that it's not physiologically straining, but not "easy" in the sense that it's easy to duplicate if you don't use a particular variant of it in your language.) You just move your tongue up lightly toward the roof of your mouth and an R starts coming out. What this means is that it becomes a minimal consonantal insertion between two vowel sounds, as well as in a variety of other contexts in which speakers of certain dialects may feel the need to default to some sort of break-up of two vowels, or perhaps an easier transition into a following syllable.
R-insertion can happen both when an R is re-inserted into a place from which it is omitted in other contexts (Linking R), or it can happen in places in which there was initially no R (Intrusive R), whether that is for ease of pronunciation or simple hyper-correction. Consider these examples:
- "Tuner amp", when pronounced with a non-rhotic English accent might sound like "tuner amp" anyway, because the R appears intervocalically. "Tuner" on its own may sound like "Tuna". This is an example of Linking R.
- "Champagne supernova in the sky", on the other hand, has no R after the end of "nova", but receives one in the famous song to facilitate the transition from "a" to "i" in the following word: "Champagne supernova[r] in the sky...
- The R interpreted in "Kuykendall" is kind of a hybrid. It wasn't originally there, so it's not Linking R. But on the other hand, it's only Intrusive R because it didn't belong there in the first place, not because the speakers that inserted it actually needed it there to satisfy a pronunciation tendency of their own. This particular R-intrusion seems to be purely a result of hyper-correction based on a false assumption, and it stuck as the family's pronunciation of their own name evolved in the area in which they settled.
When we look at accents that do or don't pronounce certain Rs in the United States, it's worth noting whether the languages that influence them do or don't, as well. For example, some southern dialects will leave out R-sounds when pronouncing something like "Car", as would the kinds of English-accented speakers whose accents influenced them historically:
"Get in the car and go." vs "Where's my ca[r]?"
On the other hand, while you might imagine Irish accents to have had an influence on Bostonian English, Boston accents don't really have as much in common as you'd think, and this shows in the fact that the typical Irish accent is rhotic, whereas strong Boston accents tend to omit R in the relevant places. Speaking of Irish accents, there's an interesting CNN bit on how the distinction between rhotic and non-rhotic dialects can affect the ease with which speakers of one accent can imitate another. It's short and worth a watch.
On a final note, if you find exploring different accents to be interesting, then you might enjoy playing with this Map of English Speech Accents based on data from George Mason University's Speech Accent Archive. Please don't read into the similarity scores feature much at all for now. It's a work in progress (and the notes indicate as much), but it's certainly interesting to click around and get a feel for what accents sound like in different areas of the world.