The Case of the Missing Pi Day 4s

Yesterday was Pi Day, 3/14 (those who prefer days before months can have Pi Approximation Day, 22/7) and in celebration of this momentous annual event, I invited several of my American colleagues (who have learned to tolerate my numerical eccentricities) over to my house in Canada for an International Pi Day Pie Party, which was a great success.  And, of course, as befitting this event, we had Pie, complete with Pi (to two decimal places) on top:

It's blueberry!So far, so good.  (And for the record, it was very good).  There was only one problem: the local dollar store I went into had a very odd distribution of candle numerals: it had tons and tons of 0, 1, 2, and 9, some 3s, but no 4s, 5s, 6s, 7s, or 8s.  As a professional numbers guy, and also as a guy who needed a 4 for his pi(e), this was deeply disconcerting.

After a moment, I figured out why. Ordinarily, when stores buy products that come in different varieties from wholesalers, the default is to order the same amount of each variety.   In this case, the store had obviously ordered an equal amount of each numeral, but they were being purchased by consumers at different rates.    Now, there is nothing about the properties of the natural numbers that would lead to this observed distribution (if it were Benford’s Law in action, it would be 1 and 2 that would be in short supply). Rather, the explanation is a social one:   Many parents do not buy birthday candles for their child’s first, second, or third birthday, because, while, as my (thankfully childless) brother noted, “Babies love fire!”, parents of toddlers do not.    At the other end, by the time your kid is about 9, and certainly by the double digits, they’ve probably outgrown the ‘giant novelty numeral candle’ phase of their lives.  Ages 4-8 are the sweet spot, and thus these sell out much more quickly.

I also note that, for adults, decadal birthdays like 20 and 30 tend not to attract much numerological attention, whereas 40, 50, and 60 certainly do (not so sure about 70 and 80), and by 90 most of the clientele is deceased.    This doesn’t explain why there were so many 0s available – perhaps purchasers are aware of this phenomenon and order extra zeroes, but don’t take account of differential demand for the tens digits.

Now, if we lived in a perfect world where suppliers and store owners had full information about their stock and made perfectly rational decisions, purchasers would notice such discrepancies and perhaps order more of the missing numerals.  The local dollar store, however, does not occupy such a world.  Fortunately, this being Windsor, Ontario, there was another dollar store across the street, and while it also had a skewed distribution, lo and behold, it did have one lonesome 4 for purchase (seen above).   Thus my Friday Pi Day pi display supply foray was saved.  Yay! (Try saying that in Pig Latin.)

Actually, this is not the first time I had encountered this phenomenon.  Back in 2008, when American gas prices first regularly began to hit $4.00 a gallon, the New York Times reported on, of all things, a shortage of numeral 4s, because their number sets were purchased with an equal distribution across all ten digits (presumably with extra 2s and 3s purchased individually to deal with those dollar amounts).  Once that leading digit got to 4, there was a temporary shortage, leading to some store owners writing their own makeshift 4s until new ones could arrive.

Thus, while we think of linguistic and symbolic resources like numerals as being effectively infinite, in contexts like these, you can indeed have shortages and surpluses.   Thankfully, now that we’re on to the Ides of March and our Pi Day shortage is dealt with for another year, I can store these candles for future use, if I want.  The pie, on the other hand, has gone to a better place.  Because, while you may sometimes need to ration your fours, let’s hope we never live in a world where we have to ration pie.

What’s so improper about fractions?

Yesterday, as part of the Wayne State Humanities Center brownbag series, I gave a talk entitled, “What’s so improper about fractions? Mathematical prescriptivism at Math Corps”, based on my long-term ethnographic research in Detroit.   For those of you who might be interested, you can watch the video below (or on Youtube itself), and the powerpoint is available for download here.

AAA itinerary

For the next several days I will be at the American Anthropological Association annual meetings in Chicago, Illinois.  Unfortunately I am once again ridiculously over-committed with committee work and departmental service and other such fun things, but if any of my readers are going to be there, feel free to track me down.   On Friday afternoon, you could check out my panel, Thinking and Talking about Metalanguage and Metacognition (Conference Room 4C).  Friday evening at the Society for Linguistic Anthropology business meeting, my student Sarah Carson will be receiving the SLA’s undergraduate essay prize (announced here).  Saturday from 10am-2pm, you could come to the exhibit hall where I’ll once again be hosting the Wayne State table at the Graduate School Fair (now with more swag for eager passers-by).

There are so many panels of interest (and so many opinions on what counts as interesting) that I can hardly list all the ones I wish I could go to (see above re: horribly over-committed).  But I do want to draw your attention to one really great panel of interest to the subject matter of this blog, unfortunately tucked away on Sunday morning: More than an Utterance: Indecipherable Scripts and the Materiality of Communication (Conference Room 5G) featuring a thoughtful slate of cross-cultural work on undeciphered and indecipherable writing systems.

I’ve promised myself this year to use my Twitter account to good effect, and so if you’re not already following me @schrisomalis, you could follow me and give me a little extra incentive to actually follow through.

Help ignite the Schwa Fire

A really exciting new digital initiative in linguistics journalism is on the horizon: Schwa Fire.    It’s the brainchild of Michael Erard, a Ph.D. in linguistics and a superlative science writer.  Erard is seeking to fill the gap between the language blogs (of which you’re currently reading one), where content is relatively short and the authors unpaid, and the literary and intellectual magazines like the New Yorker, where there are occasionally linguistic essays of some importance, but not nearly often enough or in enough detail.  Schwa Fire will be a low-cost ($1.99) ad-free digital magazine available on the web and for mobile devices featuring important ideas from people from across the linguistic sciences.

Erard is currently running a major Kickstarter initiative to get his project off the ground, and is over halfway to his $25,000 goal.  I supported it today and I would encourage others with an interest in seeing high-quality, long-form, language-related on-line non-fiction (perhaps with not so many hyphens) to do so as well.

A hithertofore unrecognized neologism

I got a note last week from a correspondent asking me about the word hithertofore, and whether or not it was a ‘proper word’.  I have to admit that at first glance I was very surprised, because of course it was a perfectly good word, and one whose meaning I knew well.   But when the correspondent said that she’d looked around and hadn’t found it, I looked at it again and realized that of course it wasn’t a word.  Or was it?

English has two words with a distinctly archaic flavour that mean ‘up to the present time’, hitherto and heretofore.    These synonyms also start with the same letter, are compounds containing to, and to top it all off, hither and here are also synonyms, so it’s not even semantically odd.   Neither word is especially common, and as you can see from this Ngram, hitherto and heretofore are really quite rare and becoming rarer.    It’s hardly surprising, then, that some speakers and readers might blend these two. Whether we think of it as adding -fore to hitherto, or substituting hither for here in heretofore, doesn’t much matter, as the result is the same, hithertofore.

What should perhaps be more surprising is that hithertofore hasn’t hithertofore been included in any dictionary, not even with a usage note.   It’s not hard to find in use in printed books; Google Books claims 67,500 works containing it (although that number is probably inaccurate) in lots of different genres.  There are plenty of words in big unabridged dictionaries that are far less common than that.   I’ve found it going back at least as far as 1708, and I didn’t have to look very hard. While it seems at a glance that a higher than average proportion of these works are authored by non-native English speakers, I also would argue that one has to be relatively fluent to even make such an error, conflating two already-unusual words.

Note, though, that its Ngram, rather than slowly declining from the 19th century until today like those of its two constituents, shows it to be largely a product of the mid-20th century, peaking around 1970.  This suggests, firstly, that perhaps it was at its most popular when its two constituents had declined enough in frequency that they had fallen out of regular use (and were thus prone to confusion), but were still common enough to be intermixed.   It hit its sweet spot half a century ago, but now the two well-accepted words themselves are falling out of use in favour of previously  or other terms, so hithertofore may actually have lost its chance to become another widely used variant (even at its most popular, it was less than 1% as frequent as heretofore).    I still think it’s a neat example of the way that memory, meaning, and phonology can lead to the appearance of nearly-invisible blends, and given that it is a relatively common error, it could probably use some lexicographical attention.