June 17, 2005

Descent into the advice literature

grano In this week's mail: two observations from Stanford student Tommy Grano on the perils of the advice literature on grammar, style, and usage: one illustrating Do As I Say, Not As I Do, and one illustrating the Ad Hoc Instruction.  Both illustrate the unfortunate consequences of approaching matters of usage through (largely unarticulated) "theory", rather than by observing the practice of the relevant speakers and writers.

First, on the question of using while as a logical connective ('(al)though') rather than as a temporal connective ('during'), Grano writes:

I... had an unpleasant run-in with the advice literature a few days ago while looking for a GRE prep book. While one of them advised against using "while" as a logical connective, I promptly found two such usages in the introduction to the same book. Now how can I trust the book in other areas that I'm not as savvy about??

An excellent question.  Do As I Say, Not As I Do is regrettably prevalent, in the advice literature on language as in parental advice to the young.  People who tell you to replace every occurrence of restrictive which by that use restrictive which in their own advice manuals.  People who tell you that possessives can't be antecedents for pronouns use possessives this way in THEIR manuals.  People who tell you not to strand prepositions strand them all over the place; I mean, nobody says or writes things like Of what could you have been thinking when you wrote that?  And on and on.

We're dealing with self-delusion here.  The advisers have an explicitly formulated "rule", which they subscribe to so thoroughly that they believe they follow it themselves; in my experience, they also believe that they notice all violations of the rule (though a fair number pass by unremarked).  But in fact, when they're not consciously monitoring language, they mostly produce and process it without reference to the rule.

Why should they think they do adhere to the rule?  Usually because they believe that there is a "theoretical" basis for the rule.  In the case of logical while (and its sibling, logical since), there are two supporting assumptions: that any potential for ambiguity should be avoided, and that words should be used in their historically "original" meanings.  Now, neither of these hypotheses bears close examination, in general or in the specific case of logical while, but they are remarkably difficult to dislodge, because a great many people believe that this is the way language OUGHT to be.

On to Grano's second tale from the world of the advice manuals:

Then there was one usage book saying that in order to determine the grammaticality of a sentence, sometimes you have to add words (e.g., he's as tall as me --> *he's as tall as me am), and sometimes you have to subtract words (e.g., me and Sandy went to the store --> *me went to the store). I found that amusing...either add or subtract words, depending on which action will result in whatever judgment the book was going after.

Taken at face value, these instructions do seem entirely ad hoc.  Expand here, trim there.  Certainly, students must find this advice baffling.

But there are justifications, rarely articulated with care, that lie behind the instructions -- that there's a class of words in English (one of which is as) that occur in combination with a NP only by ellipsis from a full finite clause (with the elliptical version maintaining the formal features of the full version), and that when NPs are coordinated, each must be separately licensed in the context for the whole coordination.  Indeed, these justifications are widely assumed to be, in some sense, universal, because they are taken to be logical necessities; this is the way language HAS to be.  Once again, the "theoretical" justifications don't bear close examination, but they are remarkably difficult to dislodge, because they seem almost self-evident to many people.

As a piece of practical advice, what I say to students preparing for these exams is that they should get as many questions as they can from actual exams (with answers) and study them to see what points of grammar, style, and usage are in fact being tested.  Their task is to psych out the exam; the details of real (formal standard written) English aren't exactly irrelevant, but when the crunch comes these details might have to be set aside in favor of learning arbitrary stipulations.  It doesn't please me to be giving such advice; I'm appalled at the way "grammar" is taught and tested.  But to urge students to revolt at exam time would only be to disadvantage them.  They are not the part of the system that needs changing.

zwicky at-sign csli period stanford period edu

Posted by Arnold Zwicky at June 17, 2005 01:45 PM