Posted by Neal on March 17, 2010
“Doug?” I asked as I knocked on his door.
“It’s going to be 65 degrees today, but it’s only 35 now, so you’ll want a long-sleeve shirt. Also, it’s St. Patrick’s Day, so you need to decide whether you’re going to wear green, or consciously boycott green.”
That’s how I have to put it to Doug. Otherwise, he’s likely to take the same attitude that I had toward St. Patrick’s Day when I was a kid. “Wear green? Why should I? I don’t have to wear green!” My attitude was a result of my unhappy introduction to the holiday: I walked into my first-grade classroom on what I thought was an ordinary morning, only to be greeted by three pinches within a minute. The teacher rescued me by giving me a green construction-paper shamrock to pin on, saying something about “Patrick’s day.” What, my friend Patrick Montgomery had a day in his honor? What did that have to do with getting pinched for not wearing green?
When Doug and Adam started going to school, I made sure they didn’t have to learn about St. Patrick’s Day the hard way. Forewarned, they can wear green or not, but at least not be surprised if people comment if they don’t.
The name Patrick reminds me of a phonology question I’ve had for a while. In the English I speak, the /t/ in a /tr/ cluster is actually pronounced [ʧ] (i.e. the “CH” sound). So Patrick is more like Patchrick; trick like chrick; troop like chroop, etc. In phonetic terms, the [t] is becoming an affricate — that is, a sound that begins with a stop consonant (in this case [t]) and ends with a fricative (in this case [ʃ]). (I’ve blogged about affricates before, here and here.)
The same thing happens when I pronounce /d/ in a /dr/ cluster. Just as the voiceless /t/ turns into the voiceless affricate [ʧ], the voiced /d/ turns into the voiced affricate [ʤ]. That’s the stop [d] plus the “zh” fricative [ʒ], or in other words, what we English speakers know as the “J” sound. So hadron sounds like hajron; drip like jrip; droop like jroop, etc.
I’m not alone in these pronunciations; most speakers I’ve heard have also done it; and I’ve even heard of children who are learning to write spelling tr and dr as chr and jr. Why would this happen? To pronounce /tr/ or /dr/, your tongue’s starting position is with the tip just behind for top front teeth and the back of the tongue lowered. What comes next depends on how you pronounce your /r/s. If you make a “bunched /r/” like I do, then your tongue tip will lower and “bunch up” toward the back of your tongue, which will rise up toward the palate. If you have the “retroflex /r/”, then the back of your tongue will stay more or less where it is, and the tip of your tongue will move toward the palate, possibly enough that your tongue is curling backwards a little bit. (For more on bunched /r/ and retroflex /r/, read John Wells’s blog post or this explanation from Kevin Russell (scroll down to “retroflex”).) In either case, the tip of your tongue needs to move from its /t/ position. If, instead of picking it up from the /t/ position and placing it at the /r/ position, you just slide it from one to the other, air will escape through the gap between the tongue tip and the alveolar ridge (the part of your palate just behind your top front teeth). When that happens, you have a fricative.
As I think about this, a question I have is why the fricatives we get are [ʃ] and [ʒ], instead of [s] and [z]. I think it’s because a bunched /r/ is made with the tongue rising toward the palate, and [ʃ] and [ʒ] are, too. So as you get to the /r/, the air is escaping over a tongue that’s closer to the position for [ʃ, ʒ] than it is to the position for [s, z]. Supporting evidence comes from the fact that this same affrication occurs when /t, d/ come before another palatal sound: /j/. ([j] represents the consonant Y sound.) That’s how we get the pronunciation witcha for wit’ your, and didja for did ya. (And the “ch” sound in spatula and tarantula.) This hypothesis would also predict that /tr/ and /dr/ affrication occurs only with speakers who have bunched /r/, not those with retroflex /r/, so there’s a research project for someone. (If it’s been done, please leave a comment giving the source!)
However, one fact is still unexplained. It turns out that (at least for me) there is one exception to this affrication rule: When the /tr/ cluster is part of an /str/ cluster, I don’t affricate. I don’t pronounce string as [sʧrIŋ] (s-chring); I pronounce it as [strIŋ]. I don’t pronounce stray as [sʧre] (s-chray); I pronounce it as [strej]. There is another phonological rule that affects /t/ at the beginning of English words, and also has an exception when the /t/ is preceded by /s/, so maybe these two rules are related. Moreover, there are other speakers who don’t have an exception for /str/, and (as I learned in a January thread on the American Dialect Society listserv), there are even speakers who let the palatalizing of the /t/ pass on through to the /s/, resulting in [ʃʧr]. Larry Horn noted it in Michelle Obama’s plea for aid for Haiti, when she says, “Schools de[ʃʧroyed" and "parent[ʃʧr]ying to feed their children”. But those topics will have to wait for another post.