Unfree variation, summer of 2007

From a wildly defensive comment of mine on my 9/16 posting “The pengring”, a comment about my loss of an audience in linguistics:

Story. My proposal for a course at the 2007 Linguistic Institute (at Stanford) was rejected on its merits, but was added at the last minute (after I appealed in wounded anger) because I was a local faculty member. (This is the position of the applicant for college admissions who would not be admissible on the merits but is accepted as a legacy.) The course was, by all accounts, a tremendous success and I’m proud of what I did. (I’ll post the course description and my course-end summary of its content; all of the material for the course is in a Page on this blog.)

So here it is.

Background: from the Linguistic Society of America site, in the report on the 2007 Linguistic Institute:

The 2007 Linguistic Institute, which took place from July 1-27, was hosted by the Department of Linguistics, Stanford University. Its theme was “Empirical Foundations for Theories of Language”, reflecting a focus on emergent directions for linguistic research. Courses and activities were chosen to inspire the broadening and clarification of the empirical basis of linguistic study, alongside the development or refinement of theoretical models.

…a large part of the slate of regular session courses were chosen following a call for proposals, which then went through a review and revision process, with the remainder taught by invited or on-site instructors.

My rejected proposal, which ended up serving as the course description:

The advice literature on English grammar, style, and usage is full of cases where it is claimed that two expressions – restrictive relativizers which and that, sentence-initial but and however, complementizers if and whether – are free variants, identical in semantics and discourse function; the advice is then to cease using one of them in certain syntactic contexts (always use that for restrictive relatives, always use but sentence-initially) or to distinguish them merely stylistically (if is informal, whether formal, so don’t use if in formal contexts).

As Bolinger observed long ago, usually these “free variants” turn out to be subtly different in their semantics or discourse function, and often in their syntax as well; and sometimes these differences are related to the prosody of the variants. They are in fact unfree variants.

This seminar investigates an assortment of cases where two or more expressions serving as alternatives seem to be in free variation, differing at most in stylistic value; explores the semantic, discourse function, syntactic, prosodic, and processing factors that favor the choice of one variant over the other; and evaluates claims about the stylistic values of the variants.

When the course was added to the catalogue for the Institute, these notes were added to the description:

This course focuses primarily on lexical variants, while LSA.341, Paraphrase and Usage, focuses primarily on alternative syntactic constructions.

Course Areas: Language Variation, Morphology/Syntax, Semantics/Pragmatics. Prerequisites: Basic level linguistics.

For assignments, the students were provided with links to publicly available corpora that they could search; and with tools for simple statistical analysis. The assignments:

1. For credit students (and auditors who want to join in the fun): pick one topic from the first two sections of the file “A list of some variants in English” [link provided]. I’d suggest you avoid the five topics that I’ll be talking about [whichthat, muchalot, howeverbut, countmass, P+of], unless you think you have a fresh slant on one of them. You can also suggest an appropriate topic that’s not on the list.

Your task is to discover something about how the variants you chose differ – in syntactic distribution, meaning, discourse function, style, whatever. It’s fine to pick a topic that’s been studied a lot by other people; the point is for you to experience the process of discovery, not (necessarily) to make a contribution to scholarship in a few weeks of summer school. (You might reinvent the wheel. Even so, there’s a fair chance that your wheel will be an interesting one.) So I will actively discourage you from rummaging through reference grammars, advice manuals, the technical literatures on grammar and variation, etc.

Start by making some hypotheses about what’s going on and then checking these hypotheses against easy-to-use corpora. See what you can find doing Google searches. These are quick, though dirty, but you can often make real headway on a topic that way. We’ll talk about some other possibilities in class.

We’ll have an initial discussion of the topics in class on 7/12. And then students will present their findings at the last class meeting, on 7/26.

2. [the intention of which is to promote active listening to language use around you, something I want to encourage for all linguists] Your second assignment is to keep a data log and e-mail it in two installments to me (by Friday evening 7/13 and 7/20). This is a record of things you notice in what you hear and read: variants that catch your attention, including those you think might be inadvertent errors, those that are not put the way you would put them, those that sound unfamiliar to you, etc. Try to provide as much information as you can about where you collected your examples and who produced them. Discuss your examples briefly: what do you think is going on? (I can cope easily with .doc, .txt, and .rtf files.)

And the summary (with references to case studies discussed in class):

LSA.376: Some thoughts for the last class (7/26/07)

1.  Overall strategy in looking at variants X and Y: Look for differences in them, beyond differences in their syntax and their sociolinguistic status. That is, look at the places where they can in principle alternate in the speech/writing of (some) individuals.

Expect that they will be interchangeable in many of these contexts, but that there will be some contexts in which one is preferable, or even the only acceptable choice.

2. These differences can be, at least:

— sensitive to lexical choices in the surrounding context (degree muchalot: preferences according to modified A)
–sensitive to syntactic factors: (VP adverbial muchalot: preferences according to position)

— semantic/pragmatic (differences in sense, entailments, presuppositions, conventional implicatures)
— discourse-functional

— stylistic (differences in formality, colloquialness, etc.) (whichthat, muchalot, howeverbut, P+of)
— prosodic/phonological (whichthat, muchalot, howeverbut, P+of)

— tied to processing factors (in production or comprehension)

3. Typically, it will turn out that many different factors are relevant, and these factors will have different weights.

4. One very common difference is that, in semantics/pragmatics or discourse function, Y = X + something; from the muchalot.notes file:

difficult = hard + something
nearly = almost + something
subordinator once = after + something
restrictive relativizer which = that + something
sentence-initial connective however = but + something
determiner a lot = much/many + something

and now possibly:

out of = plain out + something

(Bear in mind that variants can differ in more than one way.)

5. But X and Y can simply be different, though maybe overlapping on occasion: double categorization of some nouns (e-mail, spam) as Count and Mass.

(Lecture notes for each of the five case studies are also available on this blog.)

5 Responses to “Unfree variation, summer of 2007”

  1. Ellen Kaisse Says:

    That sounds like a great course, and it’s not in the area I had trouble with and stopped following closely, which was morphology. Unfree variation indeed sounds like an outgrowth and logical development of themes from your earlier work.

    • arnold zwicky Says:

      Presumably, then, it was stuff like my *quicklier paper that turned you off. A meticulous attempt, in agonizing detail, to show how theoretical assumptions guide morphological analysis. The referee who hated it so much took the position that we all have to work work within the currently accepted framework of theoretical linguistics (meaning generative linguistics in whatever framework is dominant at the moment), so that it’s just foolish to entertain alternative assumptions; to advance the analysis I preferred (on the basis of my preference for certain carefully exposed theoretical positions), I would have to defend in detail, with evidence, every one of my theoretical assumptions that was dissonant with current received opinion — an obviously unachievable task.

      The editor let my piece pass anyway. I would be the first to admit that the sort of scrupulous analysis I went through makes for very hard reading, and massively unrewarding reading as well, if my theoretical preferences are not yours. But I tried to diligently expose the connections between theoretical commitments and morphological descriptions, something almost no one ever does (because it is wearyingly difficult, especially when you think your unexamined theoretical commitments are just obviously, even necessarily, correct).

      If I’m lucky, maybe four or five people read my piece through (Mark Aronoff, Paul Kiparsky, and …?). But then I keep reading that most journal articles in most fields get only a handful of readers, so I’m no more worthless than most academics.

      As for the course, I remind you that my Stanford colleagues did in fact reject it, and never reversed that opinion. Even though the committee in question found it worthless, they let me teach it anyway, because I’m Arnold Zwicky and sort of famous and I made a big aggrieved fuss, not because they thought it really was worthwhile. In addition, several of those papers rejected for conferences were in fact for sections of that course — and now exist only as abstracts (available on this blog) or as postings to Language Log or this blog (also available to the public), the venues that became my substitute for conferences and journal publication.

      I have never looked back, though every so often someone tells me that that’s just because I can’t make it in the real world, in competition with real linguists. I’m trying to become fully comfortable with the idea that what I am now is an entertainer, though one with an enormous amount of knowledge and expertise in several fields.

      • arnold zwicky Says:

        “became my substitute for conferences and journal publication”: well, not *quite*. For some years I gave a series of papers, on a variety of topics, at the annual Stanford Semantics Festival (the abstracts for all of them, I think, are available on this blog) — the result of the tolerance of my Stanford colleagues who ran these meetings, especially Beth Levin. I have no illusion that it was the brilliance of my abstracts that got them accepted year after year in this one place, and I thank my colleagues for their kindness to me.

      • Ellen Kaisse Says:

        I was never a morphologist, so what grabbed me or didn’t grab me isn’t very significant. (What is significant and vexing is the failure of your colleagues to recognize what a good course proposal that was, as witness the success of the course and, from what little I know, the continued relevance of the topic 14 years later. I don’t know the ‘quicklier’ paper. The referee’s argument doesn’t make a lot of sense to me.) The conference paper I best remember, and perhaps there was another in a similar vein, had what seemed to me at the time to be a taxonomical direction — you were identifying what kinds of phenomena could occur in morphology, giving examples, and giving names to those that didn’t have one. Now that I put that down in words, I can’t say what about it wouldn’t have seemed interesting to me, aside from the fact that, as I said, pure morphology was not my bag. But from what you say, the source of rejections was not predominantly that kind of paper.

      • arnold zwicky Says:

        To EMK about my turn-off morphology paper(s): giggle. I have no idea what papers you’re talking about, but then I’m not very good at remembering what I’ve written. On several occasions someone has said that Arnold Zwicky said so-and-so, paraphrasing what sounded like a really good idea, something I’d be delighted to be credited with, except that I had no recollection of it — and then they went away and dug up the exact quote and its source and read it accusatorily to/at me.

        With the blogs, it’s much worse. Between LLog and AZBlog, I’ve written about 11,000 postings, so I’ve forgotten enough details to paper, oh, Australia over with them.

Leave a Reply to Ellen Kaisse Cancel reply


%d bloggers like this: