"Hard vowel sounds"

« previous post | next post »

"Red-blue divisions start with newborns’ names; parents show partisan tendencies", Washington Times 6/5/2013:

Names with the soft consonant “l” or that end in a long “a” — for example, President Obama’s daughter Malia — are more likely to be found in Democratic neighborhoods, while names with hard vowel sounds such as K, G or B — think former Alaska Gov. Sarah Palin’s sons Track and Trig — are more popular in Republican communities.

I've pretty much given up on the idea that literate people can be expected to know the difference between voice and tense, or passive and active, or even nouns and verbs. But I thought that consonants and vowels were pretty safe, at least as a taxonomy of orthographic categories. I mean, "AEIOU and sometimes Y", right?

Anyhow, what's even more striking about this article is the fact that it manages never to cite or link to the paper it discusses, which is J. Eric Oliver, Thomas Wood, and Alexandra Bass, "Liberellas versus Konservatives: Social Status, Ideology, and Birth Names in the United States", 2013.

In fact, information about the source is dribbled out bit by grudging bit. In the second paragraph, we're told about "A research paper by a team of University of Chicago political scientists".  In the fourth paragraph, the paper's title is given. In the eleventh paragraph, we get the authors' names. We never get a link.

Can you imagine an article about a politician's new proposal or a company's new product that begins with a general description in the second paragraph, gets around to telling us what the proposal or product is called in the fourth paragraph, and withholds the name of the politician or the company until the eleventh paragraph?

In fact, I believe that Journalists often deal with scientific and technical papers in a similar way: The tip of the informational pyramid is just the fact that a "study" or "paper" exists; then maybe the institution and perhaps the field; the people, the journal, the title, etc., are backgrounded if indeed they're mentioned at all.  It might be an interesting term project for a journalism course to determine whether my impression is valid, and how this treatment compares with stories about new products, new novels, new movies, and so on.

Obligatory screenshot:

[via fev at headsup: the blog]


  1. Robert Coren said,

    June 6, 2013 @ 12:37 pm

    Even more striking is the implication that, if such a correlation even exists, it means anything whatever.

  2. Philip Lawton said,

    June 6, 2013 @ 12:40 pm

    Speaking of journalists dealing with all science in a similar way, the Guardian (UK) published this a few years ago: http://www.guardian.co.uk/science/the-lay-scientist/2010/sep/24/1

  3. Slugger said,

    June 6, 2013 @ 1:22 pm

    This explains why "Barack" is such a popular name in Republican neighborhoods; whereas "Willard" is popular among Democrats.

  4. Yerushalmi said,

    June 6, 2013 @ 1:24 pm

    Wow. I just… wow.

  5. Dick Margulis said,

    June 6, 2013 @ 1:52 pm

    I don't see the problem. KGB are incredibly hard to pronounce as vowels. They are, in fact, not only hard vowels but passive ones. The KGB has been passive since its demise in 1991. You people need to get out more.

    [(myl) OK, you win the internet today.]

  6. DEP said,

    June 6, 2013 @ 2:20 pm

    Don't forget "W" as a vowel. My first grade teacher drilled that into my head, "A-E-I-O-U, and sometimes Y and W." I was 50 years old before I ran across the "Wester Cwm" on a National Geographic map of Mt. Everest. It was like spotting an ivory-billed woodpecker.

  7. DapperAnarchist said,

    June 6, 2013 @ 2:31 pm

    DEP – W is a Welsh vowel. Cwch is a good example – it's a welsh word that's been imported into English, meaning "a loving hug" or "a hug that makes one feel safe and cared for" (at least, thats how it's been explained to me.

  8. Chris said,

    June 6, 2013 @ 2:39 pm

    And where does this leave the recent fad for names that rhyme with "Aidan", such as Braden, Caden, Jaden, et cetera?

  9. dw said,

    June 6, 2013 @ 3:15 pm

    Does "Malia" in fact end in a "long A"? (I presume this means the vowel of words such as "spa").

    I'd always assumed it ended in a schwa.

  10. KevinM said,

    June 6, 2013 @ 3:26 pm

    @dw: On Long Island, we pronounce it like "King Lear."

  11. Chris C. said,

    June 6, 2013 @ 3:43 pm

    In English, a "long" vowel is usually a diphthong.

  12. mollymooly said,

    June 6, 2013 @ 3:46 pm

    @DEP W is, less esotericly, a component of the vowel digraphs AW, EW, OW. In a word like "FEW", one might argue that the phoneme-grapheme correspondences are
    F = /f/
    E = /j/
    W = /u/

  13. dw said,

    June 6, 2013 @ 3:49 pm

    @Chris C — yes, I'm aware that "long A" usually refers to the vowel of "FACE". I couldn't believe that this was how "Malia" is intended to be pronounced, though. FWIW, Wikipedia has "Malia" as /məˈliːə/ (ending like "idea").

  14. Cameron said,

    June 6, 2013 @ 4:12 pm

    I don't think I've never heard "Malia" pronounced with /eɪ /.

    I think it's generally /mʌliːə/ . Though some might use something more like ɑː in the first syllable.

  15. Peter Taylor said,

    June 6, 2013 @ 5:00 pm

    At least this news article specifies the speciality of the researchers as "political scientists". I see many reports of research which don't specify beyond "scientists", leaving the reader guessing as to whether they're publishing speculative nonsense outside their field of expertise.

  16. Ted said,

    June 6, 2013 @ 5:33 pm

    The paper myl links to says "Please Do Not Cite Without Permission."

    I guess the Washington Times didn't have permission.

  17. Jeff Carney said,

    June 6, 2013 @ 5:54 pm

    OK, I'll peeve. Popular articles that are prompted by a scholarly study, but which mix up the study itself and background information to the extent that the reader thinks the background is the focus of the study, and the scholars are specialists in an area in which they do not specialize.

    Case in point. A scholar specializes in conducting surveys. The goal of his team's most recent survey is to determine the public's awareness of "third-hand smoke." It turns out few people even recognize it as a thing, and this is what the study reports. But the journalist summarizing the study realizes that most readers will need a hefty explanation of third-hand smoke before they can place the survey results in a context. The unintended result is that many readers (my students, anyway) come away thinking the study was about the chemical composition of cigarette smoke and its residue, and that the study authors are experts in carcinogens and other toxins.

    Part of the reason I have my students read this article is to show them one of the many traps that await them when they reference popular articles in their research papers.

  18. J.W. Brewer said,

    June 6, 2013 @ 6:07 pm

    The dataset they're primarily using is California-only (pretty sure it's impossible to get a comparable national dataset, but the fact that you don't have non-California data doesn't avoid the question of whether the behavior of Californians is a plausible proxy for that of the national population); someone with greater statistical chops than I have can assess whether the phonological / ideological connection they're seeing is large enough to be significant (in either a technical-stats or common-sense sense),* although even within California they seem to be claiming that ideology is only really a meaningful factor in naming for children of college-educated non-Hispanic white mothers. But just as a reality check, the Washington Times story has a sidebar indicating that "Liam" is currently the most popular boys' name in Wyoming (NB: Not a Blue State), but "Liam" is one of the specific examples the paper gives for a "liberal" boys' name, on account of having supposedly "feminine" phonemes.

    *They assign names a phonemically-based "Male Gendered Score," which allegedly comes in at .48 for conservative-neighborhood boys' names, .44 for liberal-neighborhood boys, .38 for conservative-neighborhood girls, and .35 for liberal-neighborhood girls.

  19. dw said,

    June 6, 2013 @ 6:53 pm

    @Peter Taylor:

    Well, these "political scientists" don't know how to spell "cachet". (p. 9)

  20. Rodger C said,

    June 6, 2013 @ 7:45 pm

    Does this mean elves are liberals and dwarves and orcs are conservatives?

  21. Phil said,

    June 6, 2013 @ 7:53 pm

    Legolas and Elrond would be conservative. And Gimli would be liberal. Or did I get that right? Well, who cares? They didn't get it right either.

  22. Steve Morrison said,

    June 6, 2013 @ 8:12 pm

    So in the word "yaw" the y is a consonant and the w is a vowel, whereas in the word "way" the w is a consonant and the y is a vowel?

  23. Ø said,

    June 6, 2013 @ 8:51 pm

    And in "Yaw-way" you aren't allowed to write any vowels at all.

  24. J.W. Brewer said,

    June 6, 2013 @ 9:11 pm

    The article's primary cited inspiration for the notion of dividing names into more "masculine" and "feminine" based on their phonemes seems to be (Whissell 2001) which is an article titled CUES TO REFERENT GENDER IN RANDOMLY CONSTRUCTED NAMES published in a journal named Perceptual and Motor Skills. Prof. Whissell's homepage describing her research interests ("I have developed a scoring system for the phonaesthetics of sounds in language") is here: http://laurentian.ca/faculties/detail/cwhissell. Is anyone familiar with Prof. Whissell's work on quantitative phonaesthetics and its, shall we say, overall plausibility? I mean, I'm fine with the notion that statistics might show that at present more boys than girls are given K-initial first names whereas more girls than boys are given L-initial first names, and if that were the case it might be plausible that research subjects asked in a vacuum to guess at the gender-identity of made-up names would guess accordingly. But it seems like a non sequitur to get from that to where I think the U of C authors at least would need to get, i.e. that Karens and Kathys are somehow perceived as less feminine / more masculine than Lauras and Lisas. I don't see how that would necessarily follow from the overall K v. L stats. (Of course, why conservatives who are stereotypically committed to traditional sex roles would give their daughters more "masculine" names is yet another mystery . . .)

  25. Mike Maxwell said,

    June 6, 2013 @ 9:49 pm

    Perhaps this is why Ronald Reagan started out a Democrat and ended up a Republican? His name had both kinds of sounds?

  26. Teresa G said,

    June 6, 2013 @ 11:42 pm

    I'm always frustrated by non-linguists' attempts to intuit what makes a name sound more masculine or feminine. They always go about discussing "hard" and "soft" sounds (with no definitions given other than instinct, apparently). Rarely does anyone mention the most distinctive marker of feminine-sounding names (for English speakers) which is ending with an open syllable. A name like "Sasha"–which is a common boy's nickname in Russia, for example–would nearly always be thought of as a girl's name here. Historic boy's names which ended in vowels such as "Leslie" or "Dana" have almost uniformly become thought of as only for girls. Next to ending with an open syllable would be having a sonorant as the final phoneme. Thus names like "Kieran" and "Jordan" also tend to be co-opted for girls. Yet articles like the above ignore this obvious marker in favor of ill-defined generalities like the aforementioned "hard vowels" (very funny, this). It makes me call into question all of the rest of their claims.

  27. Mark Mandel said,

    June 7, 2013 @ 12:18 am

    All right, you've forced my hand. I wasn't going to comment on this post, but here's a comic a friend sent me today … (looks at clock) yesterday.

  28. Mark Mandel said,

    June 7, 2013 @ 12:22 am

    Grmph. The comic is at http://freefall.purrsia.com/ff2400/fc02354.htm

  29. Chris Croy said,

    June 7, 2013 @ 3:41 am

    Laura Wattenberg, who runs the site Baby Name Wizard, has been writing about differences in how Republicans and Democrats name their babies for some time. Her most polished work (PDF) is from 2006. She found that conservatives were more likely to give their kids very uncommon names while liberals tended to pick old-fashioned, Biblical names. Her explanation was conservatives get knocked up at age 20 and give their kids cool names like Ringo while liberals delay childbirth until they're 27 and give their kids uncool names like John, Paul, or George. You can go play with her data yourself (Java).

  30. Rod Johnson said,

    June 7, 2013 @ 8:09 am

    Every time an article like this appears, it saps my will to live a little bit more.

  31. richardelguru said,

    June 7, 2013 @ 8:13 am

    Chris, only one data point ( :-) ), but I'm liberal, had my first when my wife and I were in our thirties. and named him Rædwald.

  32. J.W. Brewer said,

    June 7, 2013 @ 9:47 am

    Wattenberg is working with a national dataset, but one with much less fine-grained detail than the California dataset the U of C people have. Her results are more interesting because they can be presented as, at first glance, counterintuitive. Her results are less interesting because she doesn't even try to correct for demographic variation. (We can all assume that the percentage of boys named Diego is strongly correlated with percentage of Hispanic population, but her national dataset doesn't e.g. permit names from a particular state to be broken down by ethnicity or maternal age, much less by redder or bluer county or precinct within the state.)

    The U of C people seem to have sorted by maternal educational level but not by maternal age. Those of course are correlated with each other, but it would be nice to know if you could tease them apart a bit. Wattenberg's thesis seems to be that age is more important (with educational level just being a statistical proxy) but I don't know that that's ex ante more plausible than the contrary thesis.

    To Teresa's point, at least based on a quick skim, I think the U of C paper's phonemic analysis may not have systematically looked for patterns in where within a name a phoneme occurs (e.g. Valerie, Vanessa, Stephen and David all just got coded as "contains /v/"), which seems like a shortcoming.

  33. Daniel Barkalow said,

    June 7, 2013 @ 10:17 am

    My first thought was that they were talking about Trff Bmzklfrpz, but now I'm not sure if the vowels in his name are "hard" or "soft".

  34. David W said,

    June 7, 2013 @ 12:29 pm

    I think Daniel wins the Internets for that comment. All of the Internets.

  35. Bob Ladd said,

    June 7, 2013 @ 3:14 pm

    @Phil and others: Citing individual counterexamples is beside the point – this is a statistical generalisation. And it is possible to do rigorous work on this kind of thing: see here for an example. Whether the paper under discussion involves rigorous work is a separate question.

  36. Ralph Hickok said,

    June 7, 2013 @ 3:32 pm


    You're thinking of 'cwtch"; "cwch" means "boat."

  37. J.W. Brewer said,

    June 7, 2013 @ 4:01 pm

    Thanks to Bob Ladd for linking the interesting Cutler paper, which is probably still useful even if potentially a bit out of date with respect to ways in which US naming patterns have diverged more from UK patterns in recent decades.* What Cutler et al. don't do, of course, is assert that there is no phonological overlap between male and female names. The differences in distribution are interesting, but, to repeat a point I made above, I simply don't think it follows that the minority of common male names that exhibit the phonological characteristics more common among female names are for that reason perceived as less masculine or more feminine. They *might* be, but that's simply a different question (one which, on quick skim, Cutler et al. do not appear to opine on) where it might be possible to design a suitable experiment to get some relevant data.

    *For example, I think weak-first-syllable boys' names, however historically unusual, may be somewhat more common among those innovated in segments of the African-American community, especially those that seem prosodically reminiscent of French surnames, e.g. DeShawn, DuWayne, and LeBron. I suspect that parents giving their sons such names do not generally worry that they will be perceived as feminine.

  38. Mark F. said,

    June 7, 2013 @ 4:17 pm

    Teresa G. – Without disputing any of your main points, the Takete-Ooloomoo experiment does show that people's intuitions about "hard" vs "soft" sounds are fairly repeatable. I know the experiment didn't say anything about "hard" and "soft", but I bet people would be pretty consistent in which of those words they described as having harder sounds.

  39. J.W. Brewer said,

    June 7, 2013 @ 5:34 pm

    OK, just to follow up on my own prior quibble with Cutler et al., here's a summary of data from a famous/infamous Levitt et al. paper http://www.slate.com/sidebars/2005/04/sidebar_8.html where they used some iteration of the same California dataset to try to pull out the 20 "whitest" and "blackest" boys' names (i.e. the names that were strongest predicters of the race of their bearers). Note that just about half of the "blackest" names have weak first syllables while just about zero of the "whitest" ones do, and note that the "blackest" names fitting this pattern not only include modern creations like DeShawn, but Arabic borrowings like Jamal and Malik and ones taken from whatever one might call the old common stock of "normal" Anglo-American names, such as Tyrone and Maurice.

    Trying to quantify phonological/prosodic distinctions between boy names and girl names without capturing possible systematic differences in patterning along other salient demographic dimensions may lead to unhelpful analysis.

  40. JQ said,

    June 8, 2013 @ 3:06 am

    @Rodger C , Phil

    I'm pretty sure elves and dwarves are all conservative. The only liberal was Sauron.
    (Please ignore the present-day American definitions of these words.)


    For some context, Freefall is very scientifically accurate and the robots in the comic are capable of independent thought.

  41. Gav said,

    June 8, 2013 @ 4:33 am


    there's a town just up the road from here called Ynysybwl. 4 vowels – count them!

  42. maidhc said,

    June 8, 2013 @ 5:18 am

    I'm doubtful. Other studies I've read show that the most likely groups to make up new given names not previously common are African-Americans and Mormons. Of course the political views of these two groups would probably be quite different on average.

    Add to that the faddish nature of baby names in general, and then sample only in California, a state with a unique cultural mix not typical of the rest of the country, and I think you end up with very dubious data.

  43. Lazar said,

    June 8, 2013 @ 6:03 am

    @maidhc: I think another notable (though demographically negligible) group with a penchant for innovative names is very rich people, particularly celebrities. For them it's likely a countersignaling behavior – they're so rich that they can afford to give their children "crazy" names that a dutiful middle-class person would balk at.

  44. Suburbanbanshee said,

    June 8, 2013 @ 3:21 pm

    @Lazar: What you mean is that they don't worry about whether their kids will suffer hell during school or lose out on being hired. Having a distinctive name is advantageous in the entertainment industry.

  45. Joshua said,

    June 8, 2013 @ 7:35 pm

    So Barack Obama, a Democrat, gave his daughters "feminine-sounding" names, and Sarah Palin, a Republican, gave her sons "masculine-sounding" names. Clearly, this must be attributable to their political differences, as opposed to, say, what sex those children were.

  46. marie-lucie said,

    June 9, 2013 @ 12:28 am

    lazar: celebrities … they're so rich that they can afford to give their children "crazy" names that a dutiful middle-class person would balk at.

    I think that 'celebrities' are different from 'very rich people'. Most celebrities are from ordinary backgrounds and have become rich over a short period of time: they are "nouveaux riches", not 'old money' or 'trust funds' kinds of people. Celebrities – some of them famous for being famous – have usually thrust themselves into the public eye through behaviour designed to attract attention. Old money does not usually do so. So celebrities with children tend to give them outlandish names, while old money types give theirs relatively plain names, or names recalling members of their families. This also applies to would-be celebrities such as Sarah Palin, who was not well-known until a few years ago but now qualifies as a celebrity. Her sons' names, although fitting the stereotypical "consonant-rich" masculine pattern, are rather odd.

  47. Ted2 said,

    June 9, 2013 @ 3:44 pm

    The paper in question is dated 04/01/13.

  48. GW said,

    June 9, 2013 @ 6:24 pm

    "I've pretty much given up on the idea that literate people can be expected to know the difference between voice and tense, or passive and active, or even nouns and verbs."

    This American Scholar article illustrates some serious confusion about perfect/imperfect and indicative/progressive: http://theamericanscholar.org/tell-the-riddle-right/?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+theamericanscholar%2FyaSQ+%28The+American+Scholar%29&utm_content=Google+Reader#.UbUN15ywTuR

RSS feed for comments on this post