Word lengths
« previous post | next post »
A comic-strip question from Rooster Teeth, relayed by reader AB:
Short answer: phonology is an error-correcting code.
Long answer: another time.
July 12, 2011 @ 7:52 am · Filed by Mark Liberman under Linguistics in the comics
« previous post | next post »
A comic-strip question from Rooster Teeth, relayed by reader AB:
Short answer: phonology is an error-correcting code.
Long answer: another time.
July 12, 2011 @ 7:52 am · Filed by Mark Liberman under Linguistics in the comics
Powered By WordPress
RP said,
July 12, 2011 @ 8:17 am
It strikes me that "town" at least arguably consists of the same number of phonemes as "jep" – and that in an alternative spelling system, they might have the same number of letters, or "jep" might even have more, making the question more one about spelling (the question about phonology still being a good question, but a separate one).
Leonardo Boiko said,
July 12, 2011 @ 8:38 am
I assume spelling will also have similar error-correcting properties, given whole-word reading. Since we identify common words as a unit by their sillhouete, it pays to have words with very distinct shapes, even if this make them longer or irregular.
J.W. Brewer said,
July 12, 2011 @ 9:17 am
Isn't this a language-specific issue? My impression (which could be wrong) is that e.g. all of the phonotactically-possible monosyllables in Mandarin have been "used up" in all four tones, whereas English (with a much larger range of possible monosyllables due to its different phonotactic constraints) has some available inventory not yet given any semantic assignment. Sometimes new assignments are given: e.g. "spork" as portmanteau for spoon+fork worked well because it produced a perfectly cromulent monosyllable that was previously unassigned and thus available for use.
[(myl) And as a result, despite nominal morphemic monosyllabism, most actual Chinese words are several syllables long — the polysyllabic space is sparsely-enough populated for forward error correction to work.]
Sili said,
July 12, 2011 @ 9:28 am
"Jep" is an informal affirmative in Danish. Adopted from American as it happens.
Gadi said,
July 12, 2011 @ 9:37 am
@Sili, which highlights, as RP mentioned, the need to differentiate between orthography and phonology. Reminds me of Robin Soderling's famous "yoking" at the French Open a few years back.
Personally "Jep" looks to me like the name the English upper class would have used for their far-east enemy during the second World War.
Robert Coren said,
July 12, 2011 @ 10:27 am
@Gadi: Really? To me it looks like a diminutive for "Jephtha" or maybe "Jehosaphat".
Theodore said,
July 12, 2011 @ 10:36 am
I look forward to the “long answer” when it comes. (See my comment to this previous LL post: Why journalists need to know morphology.)
Dan T. said,
July 12, 2011 @ 10:40 am
"Blog" is another example of a formerly-unused monosyllable that got assigned an English meaning in relatively recent times. And "spam" got assigned a proprietary trademarked meaning back in the 1930s, but gained a generic-word meaning for "junk e-mail" in the 1990s.
PubliusFL said,
July 12, 2011 @ 11:07 am
Gadi: "Personally 'Jep' looks to me like the name the English upper class would have used for their far-east enemy during the second World War."
Or an Australian, what with the salary-celery merger.
dl said,
July 12, 2011 @ 11:37 am
The namespace problem is significant in computer programming and for the web.
By 2000 all 17,576 of the possible 3 letter (xyz.com) domains were registered. http://www.3character.com/3-letter-domains-the-history-and-the-future.html
Martienne said,
July 12, 2011 @ 12:39 pm
As a fan of both Language Log and Rooster Teeth, this post just made me all happy.
I also look forward to the long answer.
Q said,
July 12, 2011 @ 12:46 pm
Actually, J.W. Brewer, not all possible combinations of sounds and tones are used in Mandarin. This site http://www.studypond.com/pinyin.aspx has a good browser-based (computer) script that shows you what combinations of initials, vowels and finals are used in Mandarin. Not all tones are used for all syllables, and there are some combinations of tones and syllables which are so rare that you wouldn't encounter them except maybe in a place name or other proper noun.
If you think about Biángbiáng_noodles it wouldn't be a stretch to say that sounds like piang, diang, tiang are pronouncable in Mandarin. And as Victor Mair will no doubt say, the range of syllables is far wider than you would expect especially when you take into account dialects and colloquialisms. Don't forget erhua.
MattF said,
July 12, 2011 @ 1:05 pm
On the other hand, filling up the namespace can have unforeseen consequences:
iiiiiiii.com
(That's eight i's).
CClinton said,
July 12, 2011 @ 1:05 pm
I find it interesting also that, as the comic shows, there are cases where "to spell something" can't be applied to sequences of letters that don't correspond to a "word."
And if I was more semantics savvy then maybe I'd have more to say on that.
Spell Me Jeff said,
July 12, 2011 @ 2:30 pm
I suspect that the longer explanation gets us into the reasons cops and soldiers use words like 'affirmative,' 'roger,' and '10-4' in place of "yes."
iching said,
July 12, 2011 @ 3:10 pm
@PubliusFL: I am an Aussie and have the salary/celery merger myself (hence pronounce Alan/Ellen, Hal/hell the same, with a short "a", as in "bad") but I always distinguish jap/jep, bad/bed etc. Just sayin'.
KevinM said,
July 12, 2011 @ 3:18 pm
@iching. Perhaps the 1940's bad/bed jap/jep merger was already less upper class and more common among middle class types with genteel pretensions? (Something for which you Aussies, to your credit, are not known.) Think Trevor Howard and Celia Johnson in "Brief Encounter."
Robert Coren said,
July 12, 2011 @ 3:25 pm
@iching (and others): My recollection of the New Zealand accent (which I know is significantly different from the Aussie, but it's closer than it is to my accent) is that while I hear [ɛ] where I would say [æ], the vowel I normally pronounce as [ɛ] is likely to become [ɪ], or occasionally even [i]. I like to say that, to an American ear, a New Zealander who had accidentally forgotten their knapsack might say "I lift my beckpeck", and my inquiries at the hotel desk might well result in my being directed to the beast local ristaurant.
So there wouldn't be a Alan/Ellen merger, although there might be confusion (to an Up-Topper, anyway).
Ran Ari-Gur said,
July 12, 2011 @ 4:50 pm
@Robert Coren: As you may be aware, the phenomena that you describe are known as "TRAP raising" and "DRESS raising" (or, together, "front vowel raising"); armed with that bit of terminology, a bit of Googling finds this chart, according to which the phenomena are less likely to occur before approximants (such as /l/) than any other manner of articulation, and less likely to occur before coronals (such as /l/) than any other place of articulation. All told, I don't think it's surprising that some speakers would have a salary–celery merger, raising TRAP but not DRESS before /l/.
Keith said,
July 12, 2011 @ 6:05 pm
I love the way children can ask such questions; they have a logic that often astounds adults.
One day, my son of about two or three proudly announced to my mother that he could hop very well without falling. He proceeded to demonstrate, hopping on his right foot. My mother, knowing from playing football with him that he is right-footed, wanted to know if he could do equally well on the left foot, and asked, "can you hop with both feet?" His reply was, "if I do that with both feet, it's called jumping."
K.
Ran Ari-Gur said,
July 12, 2011 @ 6:10 pm
Correction: the above chart is just for TRAP-raising. For DRESS-raising, the chart is here. DRESS-raising is not so discouraged by following /l/, compared to other following sounds, as TRAP-raising is; but overall, TRAP-raising affects a greater proportion of TRAP tokens than DRESS-raising does DRESS tokens, so it still doesn't seem surprising that some speaker would have a partial, phonologically-conditioned merger of TRAP and DRESS.
army1987 said,
July 12, 2011 @ 6:57 pm
Essentially, NZers say "sacks, sex, six, sucks" more or less the way I say "sex, six, sucks, sacks". Usually they speak slowly enough for me to “undo” this chain shift in my mind, but once one of them told a tongue-twister including lots of /s_ks/ words and I experienced a weird variation on the Stroop effect as a result. :-)
Ray Girvan said,
July 12, 2011 @ 9:22 pm
@KevinM: the 1940's bad/bed … merger
There's a scene in the 1952 film Angels One Five where the hero, Pilot Officer T.B. Baird, meets his new RAF colleagues and instantly gets the nickname "Septic", partly due to the initials and partly because of "Baird" being a homophone for "bad" in the RP of the time.
Adrian Morgan said,
July 12, 2011 @ 9:53 pm
Regarding comments of PubliusFL/itching and for the education of others, the shall-shell merger is common among speakers living in or near Melbourne, Victoria, but is not a dialectal feature in Australia's other states. (In other words, it's a local sound change.) It's basically a widening of the [e] vowel when it occurs before [l].
People from Melbourne therefore often pronounce the name of their own city as "Mal-bourne". They also pronounce the name of the suburb Malvern as "Mole-vern", presumably so that Melbourne and Malvern are more distinct. By contrast, there's a Malvern in Adelaide too, and it's always Mal-vern.
Andrew Filer said,
July 12, 2011 @ 10:13 pm
Speaking of NZ accents and monosyllables, in the TV program Flight of the Concords, one of the two main characters (NZers living in New York) is named Brett, but is often misheard as "Brit" by New Yorkers.
I would think that a much greater portion of the monosyllabic namespace is taken up by proper names and nicknames — Brett, Jed, Jim, and the like.
Mark S. said,
July 12, 2011 @ 11:45 pm
@J.W. Brewer: MYL has already noted that most Mandarin words are polysyllabic. It's also interesting, though, that all possible monosyllables are not found in all four tones.
Mandarin has about 1,300 syllables (interjections and loan words complicate the count a little), which is several hundred fewer than if the 410 or so basic syllables that exist in one tone or another were represented in all four tones.
Coincidentally, I've just posted some info on one factor behind this: The where and why of missing second tones.
RP said,
July 13, 2011 @ 3:22 am
@Adrian Morgan, I am not sure it is safe to conclude that natives of Melbourne pronounce "Malvern" as "Mole-vern" for the reason you suggest. After all we have a Malvern here in the UK and I would pronounce it either "Mawlven" or "Molven", never "Malven". Cf "all", "tall", which have "aw", and "Walter", which for me has (short) "o", or "salt", which I pronounce "solt" but some people (perhaps my dad, 60 and originally from London) might say "sawlt". I would pronounce "Melbourne" with an "e" and would definitely confuse "salary" with "celery". But my birthplace of "Derby" would be "Dahby", in line with the odd conventions that surround certain place-names.
RP said,
July 13, 2011 @ 3:22 am
Sorry – I meant to say, would definitely *not* confuse "salary" with "celery".
Pflaumbaum said,
July 13, 2011 @ 6:17 am
I have a book on the phonetics of English from the 1920s, aimed at helping elocution teachers guide speakers towards RP, and the author has the sound in bat or Jap as /æ/, just as it is in RP today. She regards raising it towards /ɛ/ as an affectation. Whether this was true, or still held in the 1940s, I don't know, but it shows some prescriptive pressure against it even from arch RP-supremacists.
As far as Celia Johnson as Laura Jesson goes, her happily married (14 seconds into the trailer) sounds like [ˈhæpɨli ˈmæɾɨd] to me.
Rodger C said,
July 13, 2011 @ 12:39 pm
Then there's the Catalan comic strip, Jep i Fidel. Jep is (was? I haven't known any Catalan-speaking children since the 70s) a cheerful jock type, Fidel is his thin studious friend, and of course they get along perfectly.
Mark Mandel said,
July 13, 2011 @ 1:13 pm
@MattF: ¡Ay ay ay ay ay ay ay ay!
Aviatrix said,
July 13, 2011 @ 6:23 pm
"Jep" in English is the common short for for the manufacturer of aviation approach charts. As in "Do you have jeps or NOAA plates?"
Nyq Only said,
July 13, 2011 @ 8:48 pm
As a British person in Australia I assumed that it was just too expensive to transport a full set of vowels to the colonies :)
Martin B said,
July 13, 2011 @ 10:43 pm
@Nyq Only:
Half right. The shipment crashed in New Zealand, where they became scrambled. Only a few survived to make the onward journey to Australia.
Matthew said,
July 15, 2011 @ 12:08 pm
Thanks, Adrian, for your clarification about the celery/salary merger – as a native New South Welshman, I experienced a bit of a WTF moment when I read iching's comments on the Celery/Salary merger (a native Australian? really?)
I've never noticed it before in our Melburnian cousins. I guess phoneme mergers are not as easy to spot as you'd think.
i like talking about language with a six-year-old « i like trams said,
October 23, 2011 @ 3:33 am
[…] isn't yos a word? This is an interesting question. Have a look at this post on Language Log: Word lengths. I'm not convinced by Mark Liberman's short answer (in my experience, the short answer […]