SpecGram Vol CLXVI, No 2 Contents Letters to the Editor

Why Linguistics is Not a Science

The SpecGram Editorial Board

In a couple of recent editorials we have answered several of the questions most frequently submitted by SpecGram readers. Since the publication of those editorials, by far the most common question received in our offices has been, “Could please furnish us with your bank account number so we can transfer payment to you?” We cannot in good conscience accede to this request, as it violates a number of constraints and therefore suffers from what we like to call “fatal infelicity.”

Another frequent question, though, is more worthy of our attention, (though only due to its being fifth on the frequency list) and it is to that more worthy question that we now turn both our attention and that of our attentive readers, to wit, youPL:

“Is linguistics a science?”

Faithful readers will recall that this question was recently posed by precocious eight-year-old Virginia O’Hanlon (whose parents ought to be watching their computer more closely). As an elaboration of that message, we wish to provide here a fuller answer for the adults among our readers (though Virginia is certainly welcome to follow along, too).

We are honestly a bit hesitant to address this query more fully, as we feel that the truth may be (as the Australians say) “hard to swallow.” However, the “service” mindset which has been thrust upon us by our publisher apparently dictates that we nonetheless undertake such an address. So here goes.

No, linguistics is not a science.

There are, of course, innumerable published claims to the contrary, but let us assure you, Drear Reader, that they are false. If you obstinately refuse to accept this clear testimony, let us further undertake to show that these claims are false.

Claim 1: Linguistics textbooks always claim that linguistics is a science.

This is not strictly true. In fact, most linguistics textbooks carefully do not assert an equational construction here. Rather, they typically retreat to an attributive, with a formulation such as “linguistics is the scientific study of language.” This is a non-trivial difference. That which is “scienti-fic” is only “like unto” the root “science,” and need not be mistaken for the real thing, in the same way that something which is “terrific” is not “terrible” and something which is “appl-ic-able” is not an “apple.” Textbooks which baldly claim that linguistics “is” a science are few, far between, and invariably intended for the mass market, not the classroom.

Claim 2: Linguists receive grants from the National Science Foundation.

This is true, but beside the point. Let’s face it, the National Science Foundation is an American institution, and the average American completes a less rigorous science curriculum at the secondary level than the average European does in kindergartenand we need not even mention Asia here. Furthermore, the NSF, as a government agency, operates under a classified definition of “science” which is not open to public examination, and the agency therefore cannot be used as evidence of anything other than militaristic and indeed hegemonic intentions.

Claim 3: Linguists publish articles in scientific journals.

Aha! There’s that attributive again. We’ve already dealt with that. And in any case, linguists primarily publish articles in linguistic journals. Let us not go down that circular path.

Sibelan Forrester, 2012, “A Russian in Persia, and Persia in Russian: Translating Yuri Tynianov’s ‘The Death of the Vazir-Mukhtar’ ”.

Lecture delivered as part of the “New Directions in Russia, Eastern Europe, and Eurasia” series at the University of Illinois at Urbana-Champaign, November 1, 2012.

Chiasmus of the Month
January 2013

There has recently been a massive increase, we acknowledge, in the publication of linguistic articles in journals devoted to biology, medicine, and other fields within the hard sciences. However, this is not evidence of linguistics being a science, but rather of scientists attempting to delve into linguistics, in which attempt they generally meet with disastrous results (as the SpecGram LingNerds have demonstrated in many episodes of the Language Made Difficult talk show, freely downloadable as part of the SpecGram podcast).

Claim 4: Linguistic Theories, being scientific, are explanatory.

In reality, linguistic Theories (at least, the ones that make a big show of calling themselves that) don’t spend much time explaining anything. Explanation1 involves asking “why something happens”, and answering that question with an account of “how things actually lead to that something happening”. Whether this turns out to be accurate or not is not the issue; it’s not science unless you’re trying to get there. Without that, you’re just doing a particularly uninteresting form of literary criticism that has the distinction of being even less useful than regular literary criticism and without the potential of making the reader aware of power differentials (you can try to add that last part, but you’ll get banished to uttermost sociolinguistics, where Theories die of reality-overdoses).

What linguistic Theories spend most of their time doing instead is modeling. Models can represent explanations, but they don’t have to and they usually refuse to; they instead invoke categories and processes wildly unlike absolutely anything that could possibly actually be going on, and get off the hook for it by claiming they don’t care what actually happens anyway. Once you get past “Hey, maybe humans are better at something related to complex language, and that’s why we have it,” you’ve pretty much gotten past the explanatory parts of Theory, although you can get some mileage out of claiming that language does some stuff that it can’t be proven to do, and then explaining why it does those things.2

Claim 5: Linguistic Theories, being scientific, are falsifiable and evolve to adapt based on new evidence.

Linguistic Theories adapt to become less falsifiable; this is an adaptation to new evidence only to the extent that it allows a fit theory to become completely immune to it. In other words, a linguistic Theory doesn’t evolve to become another linguistic Theory; it evolves to become a religion. Universal Grammar, for example, is on the very cusp of apotheosis, and should be wrestling with Freudian Psychology any day now for who gets the nicest corner of Olympos’s sub-subbasement.

Well, Drear Reader, we think that pretty much wraps up the question and tucks it in for the night. Open and shut case, really. Linguistics is not a science. Now stop calling it one.



1 We at SpecGram steadfastly use an approved, in-house definition of “explanation,” which has the advantage of meaning pretty much what most people who speak English mean by it, as far as we can tell. One could, of course, simply tinker with one’s definition to render one’s theory explanatory. We won’t print one’s article if one does, though. Unless we’re short on submissions, or we think it’s so funny we’ll print it anyway.

2 “Why is there an invisible Lutheran wildebeest living by each functional head? The functional head actually just specifies the wildebeest. Other theories are mere surface accounts that don’t explain the presence of the wildebeest.”

Letters to the Editor
SpecGram Vol CLXVI, No 2 Contents