Some notes on inventorying postings about topics in the world of grammar, usage, and style — mostly about how these inventories are assembled and the imperfections that result.
I maintain a whole lot of inventories, mostly of examples of specific phenomena. Almost all of these inventories are “fortuitously” collected: I include things I’ve read or heard (including things I produced myself) when they catch my attention. This is not a fair sampling (in a statistical sense) of anything, but is does provide a rough picture of what’s out there, and you can use the inventories to formulate hypotheses about the frequencies of variants, their geographic and social distributions, the syntactic and discourse conditions on them, and so on.
Of course, I also do systematic corpus searches. And very occasionally I do what amounts to a systematic collection of fortuitous examples. For example, for a year I did my best to record every example of “stranded to” (Verb Phrase Ellipsis leaving the infinitive marker to as a remainder, as in “I don’t want to”) that came past my eyes (not my ears as well; that would have been unthinkable). Afterwards, I had a massive coding task, which is still not complete.
This discussion of inventories of examples was intended to ease into some of the issues about inventories of postings that I’m about to discuss.
Ideally, such an inventory would catch all the postings about some topic. To this end, I’ve kept some files listing postings as they appear. But it’s time-consuming to update such files as I go along, so I almost surely miss some.
Later I’ll go back — or have someone else go back — and search through Language Log (both classic and new) and this blog for postings with key expressions in them (“Omit Needless”, for example). But these searches will miss some, or even many, genuinely relevant postings, if these postings happen not to use the key expression. So a posting that cites someone’s claim that a word is “unnecessary” or that an expression is “redundant” or “pleonastic” or “wordy”, but without using the Strunkian “Omit Needless” slogan, won’t get caught in the net.
The problem is especially acute when the principle in question doesn’t have a long-standing and reasonably well-known label (like Omit Needless Words (ONW)).
Take Include All Necessary Words (IANW), or more generally Include All Necessary. The label isn’t original with me, though I’m not sure where I picked it up, but I’m the person who’s mostly been using it, and then only within the last few years. So when Tim Moon, this summer’s intern on the OI! project, searched Language Log and this blog for {“Include All Necessary”}, he didn’t get a lot, and it was all from me (results here).
It’s not that IANW advice isn’t all over the place. It’s just that it is very rarely couched in those terms. Instead, you get formulations involving just necessary (a very common lexical item), exhortations that people shouldn’t omit (or leave out or delete) certain words, and so on. There are lots of ways of couching the idea, and they’re not easy to search for.
Now, in the OI! project, we (mostly Rachel Cristy and Tim Moon) began assembling references to phenomena in the handbooks, categorizing them by higher-level principles (ONW, IANW, plus a collection of troublesome “redundancy/pleonasm” cases). All of these involve involve choices between roughly equivalent alternatives that share most of their elements (too big a dog vs. too big of a dog; a couple ideas vs. a couple of ideas).
In phase 1, things were organized by handbook, and within each handbook file by higher-level category. In phase 2, the task was to turn the “handbook categorization” into a “phenomenon categorization” (brief discussion of the phases here). So new folders and files were created — the folders labeled by the higher-level categories, the files labeled by phenomenon names.
This is a tricky business. Many usage manuals are arranged as dictionaries (but many are not, and as I explained here, that made it hard for the project to use them), so they have headers for the phenomena, but few of these headers are common to different manuals; very few are labels that linguists would use (and sometimes there’s no consensus among linguists); and many are just exemplars of the phenomena in question.
So as manual headers we get between you and I (for nominative conjoined objects),” preposition at end” (for stranded prepositions), too big of a dog (for of-marked exceptional degree modification), and so on.
Of course, I don’t expect manuals intended for general users to employ the technical terminology of linguistics, although I would feel better if the writers of the manuals gave some indication that they knew about linguists’ discussions of the phenomena. (This is a complex topic, and really has to do with conceptualizations of phenomena, not choice of terminology, so I’ll leave the topic for another day.)
A further reason why we can’t just adopt headers from the handbooks is that handbook entries are very often organized by what the writers take to be the most relevant word in expressions, with diverse (and often unrelated) phenomena jumbled together. In a reference work like MWDEU, this makes some sense, but in material that proposes to tell the unenlightened what to di it’s not very helpful.
So an entry on of might tell you not to use it in off of (“I jumped off of the roof”) and other P + of combinations, and not to use it in too big of a dog), while telling you that you must use it in a couple of ideas. ONW in the first two cases (which have nothing to do with one another, as far as I can tell), IANW in the second. (Look here for some more ONW cases involving of.)
The allusions to “unbrella principles” like ONW and IANW are unhelpful because they assume the unenlightened know all of the grinding details of formal written standard English (there are, at least, tens of thousands). It’s no good telling people to Omit Needless Words and Include All Necessary Words (as principles of grammar, not rhetoric; brevity and clarity are of course important precepts, and need to be taught, but they have nothing to do with the “correctness” of too big of a dog or a couple ideas), unless the people you’re addressing already know which words are “omissible” and which “necessary”.
So what we end up with is advice that amounts to “use of correctly” and the like.
[Digression. For some years I’ve been wondering about how this strange situation could have come about, until it came to me that we have a point-of-view problem here: American handbooks for college students, back roughly a hundred years, are organized from the point of view of the writing teacher, not the writer. Generations of these handbooks have similar charts on their inside front covers, giving codes for various issues in writing. The purpose of these labels (and a complex system of sub-labels) was to mark a bit of student writing as an offense against one of the (higher-level) principles at issue. That is, their purpose is to say You Have Offended Against #N.
This looks to me like what Geoff Pullum calls “nerdview” — the insiders’ articulation of affairs, with little care for the users’ viewpoint. (The term was introduced here, and there are a number of later postings.)]
To return to the original topic: phase 2 of the OI! project was fraught with complexities. Now to consider why certain phenomena get the attention they do.
Leave a Reply