October 6, 2011

Continua, Planes, and False Dichotomies

On Twitter, Erin Brenner asked, “How about a post on prescriptivism/descriptivism as a continuum rather than two sides? Why does it have to be either/or?” It’s a great question, and I firmly believe that it’s not an either-or choice. However, I don’t actually agree that prescriptivism and descriptivism occupy different points on a continuum, so I hope Erin doesn’t mind if I take this in a somewhat different direction from what she probably expected.

The problem with calling the two part of a continuum is that I don’t believe they’re on the same line. Putting them on a continuum, in my mind, implies that they share a common trait that is expressed to greater or lesser degrees, but the only real trait they share is that they are both approaches to language. But even this is a little deceptive, because one is an approach to studying language, while the other is an approach to using it.

I think the reason why we so often treat it as a continuum is that the more moderate prescriptivists tend to rely more on evidence and less on flat assertions. This makes us think of prescriptivists who don’t employ as much facts and evidence as occupying a point further along the spectrum. But I think this point of view does a disservice to prescriptivism by treating it as the opposite of fact-based descriptivism. This leads us to think that at one end, we have the unbiased facts of the language, and somewhere in the middle we have opinions based on facts, and at the other end, where undiluted prescriptivism lies, we have opinions that contradict facts. I don’t think this model makes sense or is really an accurate representation of prescriptivism, but unfortunately it’s fairly pervasive.

In its most extreme form, we find quotes like this one from Robert Hall, who, in defending the controversial and mostly prescription-free Webster’s Third, wrote: “The functions of grammars and dictionaries is to tell the truth about language. Not what somebody thinks ought to be the truth, nor what somebody wants to ram down somebody else’s throat, not what somebody wants to sell somebody else as being the ‘best’ language, but what people actually do when they talk and write. Anything else is not the truth, but an untruth.”1In Harold B. Allen et al., “Webster’s Third New International Dictionary: A Symposium,” Quarterly Journal of Speech 48 (December 1962): 434.

But I think this is a duplicitous argument, especially for a linguist. If prescriptivism is “what somebody thinks ought to be the truth”, then it doesn’t have a truth value, because it doesn’t express a proposition. And although what is is truth, what somebody thinks should be is not its opposite, untruth.

So if descriptivism and prescriptivism aren’t at different points on a continuum, where are they in relation to each other? Well, first of all, I don’t think pure prescriptivism should be identified with evidence-free assertionism, as Eugene Volokh calls it. Obviously there’s a continuum of practice within prescriptivism, which means it must exist on a separate continuum or axis from descriptivism.

I envision the two occupying a space something like this:

graph of descriptivism and prescriptivism

Descriptivism is concerned with discovering what language is without assigning value judgements. Linguists feel that whether it’s standard or nonstandard, correct or incorrect by traditional standards, language is interesting and should be studied. That is, they try to stay on the right side of the graph, mapping out human language in all its complexity. Some linguists like Hall get caught up in trying to tear down prescriptivism, viewing it as a rival camp that must be destroyed. I think this is unfortunate, because like it or not, prescriptivism is a metalinguistic phenomenon that at the very least is worthy of more serious study.

Prescriptivism, on the other hand, is concerned with good, effective, or proper language. Prescriptivists try to judge what best practice is and formulate rules to map out what’s good or acceptable. In the chapter “Grammar and Usage” in The Chicago Manual of Style, Bryan Garner says his aim is to guide “writers and editors toward the unimpeachable uses of language” (16th ed., 5.219, 15th ed., 5.201).

Reasonable or moderate prescriptivists try to incorporate facts and evidence from actual usage in their prescriptions, meaning that they try to stay in the upper right of the graph. Some prescriptivists stray into untruth territory on the left and become unreasonable prescriptivists, or assertionists. No amount of evidence will sway them; in their minds, certain usages are just wrong. They make arguments from etymology or from overly literal or logical interpretations of meaning. And quite often, they say something’s wrong just because it’s a rule.

So it’s clearly not an either-or choice between descriptivism and prescriptivism. The only thing that’s not really clear, in my mind, is how much of prescriptivism is reliable. That is, do the prescriptions actually map out something we could call “good English”? Quite a lot of the rules serve little purpose beyond serving “as a sign that the writer is unaware of the canons of usage”, to quote the usage entry on hopefully in the American Heritage Dictionary (5th ed.). Linguists have been so preoccupied with trying to debunk or discredit prescriptivism that they’ve never really stopped to investigate whether there’s any value to prescriptivists’ claims. True, there have been a few studies along those lines, but I think they’re just scratching the surface of what could be an interesting avenue of study. But that’s a topic for another time.


1 In Harold B. Allen et al., “Webster’s Third New International Dictionary: A Symposium,” Quarterly Journal of Speech 48 (December 1962): 434.
Descriptivism, Prescriptivism 25 Replies to “Continua, Planes, and False Dichotomies”
Jonathon Owen
Jonathon Owen


25 thoughts on “Continua, Planes, and False Dichotomies

    Author’s gravatar

    Loved this post! One of your best so far, methinks. My only question is, who “lives” in the bottom-left quadrant?

    Author’s gravatar

    Answer to Brinestone’s question: The language of the future! Only somewhat kidding. I’d venture to suggest that what determines the axis one falls on has to do with whether one believes language to be a living thing or merely a relic of a civilization. And in that context, I’d defend those linguists who feel there is something deeply wrong with extreme prescriptivists whose attitudes see no difference between English and a dead language like Latin. In fact, they want English to be more like Latin, they want it to be dead. Well, that’s just my two turtleddoves.

    Author’s gravatar

    Jonathan, the quadrant makes much more sense than the continuum. I see copyeditors occupying various places in the upper right quadrant, though sometimes they slip over into the upper left quadrant and become assertionists.

    Now we just need someone to come up with the questions that, when answered, will place a person somewhere on the map. 🙂

    Thanks for a great post,

    Author’s gravatar

    Excellent, thoughtful, and thought-provoking post.

    In that lower-left quadrant, I would place those people who have for years insisted that English is and should only be a Germanic language, and that English speakers should strike from their vocabularies any words with a Latin or Greek base. (Unfortunately, I’m at a complete loss of the name of this movement at the moment.) The composer Percy Grainger was one of these types; he would use, for example, “louden” and “soften” instead of the Italian “crescendo” and “decrescendo” in his music.

    People like this neither tell you how people use language nor tell you how to write well.

    I’d probably also put those people who push for simplified, phonetic spellings of English words. But you don’t want to get me started on that idiocy.

    Author’s gravatar

    After a little more research, it seems, perhaps, that “movement” might be a little too broad for Percy Grainger’s linguistic beliefs. His choices were rooted in racism and bigotry, and he seems not to have had too many followers . . . unless you count the Third Reich.

    Author’s gravatar

    We’re using the wrong term. It’s not prescriptionist, but proscriptionist.If you think about it, every so-called prescriptionist or assertionist, acts like Moses coming down from the mountain. Thou shalt not end a sentence with a preposition! Give me a break!

    Author’s gravatar

    Percy Grainger advocated what he called ‘blue-eyed English’.

    There are still folk about who’d like to say ‘Crosslikening speechlore’, instead of ‘comparative linguistics’.

    Author’s gravatar

    I’m not sure anyone exists in the lower-left corner. Perhaps people who make proscriptions against usages that don’t exist to begin with.

    Erin: The thought of making some sort of profiler or linguistic temperament sorter crossed my mind, but I don’t have time for it right now. Maybe I’ll try to tackle that when I’m finally finished with my thesis and have a little more free time.

    Andy: Thanks! I think I’d still put those types in the upper-left corner; they’re still saying that English should be something that it’s not, much like the assertionists. But the motivations are not exactly the same.

    Marc: I thought about diving into the difference between prescriptions and proscriptions but decided against it. Maybe I should have, because it may be a better way to characterize the axis that I’ve labeled “prescriptivism”. Maybe the top should be “what language should be” (prescriptions) while the bottom should be “what language shouldn’t be” (proscriptions). So a lot of what we think of as prescriptions are really proscriptions like “don’t end a sentence with a preposition” or “don’t use disinterested to mean uninterested“, even though those rules often run counter actual usage, placing them in the lower-right quadrant.

    The problem is that it’s not entirely clear what the goal of prescriptivism is. As I discussed in this earlier post, prescriptivism isn’t really a coherent principle but a set of often-conflicting goals, so it’s hard to boil it down to one characteristic.

    Author’s gravatar

    I’m afraid I don’t understand why you think Hall’s argument is duplicitous. What is a truth value, and what does it have to do with prescriptivism?

    Author’s gravatar

    I mean, it seems to me that Hall is talking about untrue statements that we see in prescriptivist writing: disinterested does not mean uninterested, it is wrong to end a sentence with a preposition, good writers don’t use “which” restrictively, etc.

    Author’s gravatar

    Hall is identifying descriptive linguistics with truth (what is) and prescriptivism with what “somebody thinks ought to be the truth” (what should be), which he then calls “untruth”. He may have been thinking about untrue statements like those, but note that he says “anything else is . . . an untruth.”

    A truth value is “a value indicating the relation of a proposition to truth”, according to Wikipedia. Linguistic descriptions are essentially propositions: the language is x. For example, “Many people use disinterested to mean the same thing as uninterested.” You can verify whether or not it’s true.

    A prescription, on the other hand, says the language should be y. “Don’t use disinterested to mean uninterested” or “Disinterested should not be used to mean uninterested.” This kind of statement can’t be assigned a truth value, because it’s not a statement about what’s true or untrue. I think it’s duplicitous to call an opinion an untruth, because opinion is not the opposite of truth. He’s essentially doing what some prescriptivists do when they say “x is not a word”—he doesn’t like prescriptivism, so he simply says it’s false.

      Author’s gravatar

      i came to the same conclusion on my own too.
      descriptivism is for studying, like physics.
      prescriptivism is for building, like engineering.

      programmers use actual programming languages, and there is ALWAYS best practices, and bad practices. there is no avoiding it. even if you disregard everyone else’s opinion on what is or isn’t good or bad practices, you WILL spend more time reading a piece of code if the code uses X feature or practice (e.g. go-to statements, declarative vs imperative style, functional vs object oriented paradigms, etc). And you, being a human, will end up preferring some things over others. voila. prescriptivism.

      What’s nice in programming languages as opposed to spoken/signed languages, is that there are objective metrics and criteria to measure various performance metrics of a piece of code or language – a thing called “cyclometric complexity”, execution times based on big O notation, how verbose something gets, how likely it results in humans writing code causing bugs, etc.

      If we could get some external objective metrics like those for human languages, we would be on our way to start saying one language/feature/construction is more _____ than another.

      for instance, korean featural writing is superior to japanese syllabaries, for the simple reason of math and being transparent about its construction (not anything to do with any ethnic, ethical, or historic reasons at all.

    Author’s gravatar

    Surely the prescriptivist position is not “Disinterested should not be used to mean uninterested”, but “disinterested cannot be used to mean uninterested”, which is indeed a falsifiable statement.

      Author’s gravatar

      i’m all for standardizing un vs dis, a vs anti/im
      there are two levels of prescriptivism here.
      1) consistency
      2) picking one assignment vs the other
      1) we should say either un or dis means not, and the other one should mean the opposite of.
      2) people can have a debate about whether un should mean not or dis should mean not…
      i am a firm prescriptivist on #1 but agnostic about #2.
      Fore example, we say moral as the positive, immoral as the negative, and amoral as simply not having to do with morality. extend this pattern of prefiex for all other words then, but CONSISTENTLY.
      do not have the prefix a- suddenly mean “the opposite of” when it meant “having nothing to do with” for the word moral.
      Another example is dis.
      if people say dislike, it is understood to mean the opposite of liking (e.g. hate), not simply not liking.
      so dis means the opposite thereof.
      But we do not use dis consistently to mean “the opposite of ”
      we say dispassionate about having a position or stake in a debate… but what that word actually means is “not passionate” – the dispassionate person can still be really interested in the debate. It simply means not biased or having too much personal ties to be objective .
      If we were to use prescriptivism proposition #1 of consistency, then we should have dispassionate mean like really being bored to death about the debate.
      There is a 3rd prescriptivist proposition about these negation prefixes:
      that we can do away with redundancy.
      dis = anti
      a = non = un
      get rid of these duplications and redundancies. it just spawns their own irregular usages and confusions as whether they mean the opposite of, or simply not having to do with, or being neutral about something.

    Author’s gravatar

    Why would you say that? Some prescriptions may be stated in such terms, but I’d say most aren’t. For example, here’s Garner: “The distinction is still best recognized and followed because disinterested captures a nuance that no other word quite does.” Later he notes that “disinterested is frequently used (or, in traditionalists’ eyes, misused) for uninterested“.

    And even when prescriptivists make statements like “disinterested cannot be used to mean uninterested“, it’s clear that they mean this in a deontic sense. When challenged and presented with evidence that it can be used that way, prescriptivists generally admit that it can, but it shouldn’t. Thus prescriptivism is really concerned with what should be rather than what is.

    […] Descriptivism and prescriptivism aren’t directly opposed. (from Arrant Pedantry) At times, people suggest that educated linguists are hypocritical for holding a descriptivist stance on language while simultaneously knowing that some ways of saying things are better (e.g., clearer, more attractive) than others. Jonathon Owen shines some light on this by representing the two forces as orthogonal continua — much more light than I’ve shone on it with this summary. […]

    Author’s gravatar

    I love this. Thanks for writing it — such an important topic.

    But it seems to me that your graph falls a bit short in that it provides a space for a value judgment on “bad” prescriptivists, but no corresponding space for “bad” descriptivists. Can’t descriptivism go too far? I think so, but I also think it bugs when descriptivists scream and shout about prescriptivists and their plodding, plodding pedantry. The reality is that these two supposedly contradictory philosophies can live quite comfortably inside one linguist. And should, I think.

    Author’s gravatar


    The graph really isn’t about good or bad descriptivists or prescriptivists, but rather about two different areas of inquiry—what’s grammatical and ungrammatical (in the linguistic sense), and what’s correct and incorrect.

    It’s certainly possible for descriptivists to get it wrong—to say something’s ungrammatical when it’s not, for example—but I’m not sure what you mean by “go too far.” Could you elaborate?

    Author’s gravatar

    This is a very insightful and thought-provoking post, indeed! Now, I’m seeing a coordinate system in language, something that I’ve usually associated with math.

    In my comment on Ask Copy Curmudgeon, I said that I wonder about a few things:

    The comments have discussed might fall into the quadrants, but I’m more interested in who might be standing at that point of intersection. Taking the value of that position as zero, does that mean that this individual does not care about truth and order?

    I see that this is a plausible model to reconcile that P vs. D dichotomy, so this next point is all about application. How else can we use the quadrants and the lines? I’m thinking along the lines of the Johari Window.

    Like I said, it’s thought-provoking.

    Author’s gravatar

    Thanks, Chad, and I’m sorry I forgot about your comment until now.

    I’m not sure it makes sense to think of it as a coordinate system. That is, I don’t really know what it would mean to stand exactly at 0,0. I probably should have described it as a set of quadrants—with two sets of nongradeable binaries—rather than a plane.

    Author’s gravatar

    I don’t think your is/isn’t and should/shouldn’t oppositions really do anything, because you can always formulate one as the other. If subjects should agree with verbs in number, then they shouldn’t be discrepant. If I believe that infinate is a standard spelling, then I don’t know what the standard spelling actually is.

    Perhaps a better approach would be to have two positive-only axes, one for knowledge, one for opinion. Then (0,0) is at the lower left, and means a person who is completely ignorant and indifferent about usage. (1, 1) would be someone who knows all about (a) language and has completely rigid opinions about what it should be. I am perhaps (0.7, 0.4) about English, but a perfect (0, 0) about Badyara.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.