Correlations in Linguistics
by “Big” Stas T. Stickle & “Lil” Rel E. Vance
Published 2024. 318 ± 2.89 pages
As *ptooey* statistics *bleurgh* has become an accepted part of linguistics, there has been much exploration of the parameter space. Yet very little space has been dedicated to mapping all the hitherto unexplained correlations in Linguistics. As a service to the field, we present in this tome a list of the most interesting ones we have unearthed, as well as some interesting negative results. A few sample excerpts are provided below.
The correlation between the number of errors and the readiness to write your own Excel formula to calculate correlations, rather than using the builtin one, is NaN.00000%.
The clarity of the syntactic analysis of your data is inversely correlated with how strongly that analysis supports your syntactic theory. (r = –0.60).
Having a Ph.D. in Linguistics is not strongly correlated with future earnings (r = –0.32) or seniority in the workplace (r = +0.22), but is correlated with selfperceived coolness (r = +0.77) and the number of IPA segments you can pronounce (r = +0.88).
The obscurity of the transcription system in historical records is positively correlated with how tonedeaf and tineared the transcriber was (r = +0.45) and how valuable the data could be to your phonological theory (r = +0.58).
The complexity and subtlety of your grammatical analysis correlates strongly (r = +0.91) with how well computational models can address the grammatical feature in question, rendering your analysis practically moot.
There is an almost perfect negative correlation (r = –0.98) between the readyness with which someone points our grammer and speling errors and the amount of friends they have.
The enthusiasm of your committee members correlates moderately negatively with their knowledge and expertise in the tiny subsubsubfield you’ve chose to write your doctorate about (r = –0.67).
The sequence of classes taken as an undergraduate (potential) major plays a weak role in determining whether students go on to study graduate linguistics, with a syntaxfirst approach having a negative (r = –0.18) correlation with eventual graduate study. However, we hypothesize that there is an underlying variable that splits the population into two subsets for which the correlation is much stronger. The intersection of the above set with the admittedly unscientific variable of “people we’d like to have a few beers with” results in a strong negative correlation (r = –0.61), while the inverse variable, “people we’d avoid at all costs, even if there’s free beer at the department party” has an even stronger positive correlation (r = +0.70). We are presently collaborating with the psychology department to find a scientifically acceptable variable that results in such strong correlations, with a peerreviewed publication hopefully to follow.
The correlation between frequency of plosive enunciation and number of close friends is r = –0.75.
We examine the context of citations of linguistics papers (co)authored by researchers in other departments to infer how linguists value the contributions of outsiders to the field. Almost all departments are viewed as making small net positive contributions, with the largest value added (in order) by researchers in archaeology, cognitive science, anthropology, foreign languages and literature, music, astronomy, biological sciences, theoretical mathematics, and hotel management. Departments perceived to contribute negatively to the literature include physics, computer science, and—in a distant last place—economics.
