The Phenomenology of Error

Joseph M. Williams
This article was originally published in CCC. The original citation is as follows:
Williams, Joseph. "The Phenomenology of Error." College Composition and Communication 32 (1981): 152-68.

I am often puzzled by what we call errors of grammar and usage, errors such as different than, between you and I, a which for a that, and so on. I am puzzled by what motive could underlie the unusual ferocity which an irregardless or a hopefully or a singular media can elicit. In his second edition of On Writing Well (New York, 1980), for example, William Zinsser, an otherwise amiable man I'm sure, uses, and quotes not disapprovingly, words like detestable vulgarity (p. 43), garbage (p. 44), atrocity (p. 46), horrible (p. 48), oaf (p. 42), idiot (p. 43), and simple illiteracy (p. 46), to comment on usages like OK, hopefully, the affix -wise, and myself in He invited Mary and myself to dinner.

The last thing I want to seem is sanctimonious. But as I am sure Zinsser would agree, what happens in Cambodia and Afghanistan could more reasonably be called horrible atrocities. The likes of Idi Amin qualify as legitimate oafs. Idiots we have more than enough of in our state institutions. And while simply illiteracy is the condition of billions, it does not characterize those who use disinterested in its original sense.1

I am puzzled why some errors should excite this seeming fury while others, not obviously different in kind, seem to excite only moderate disapproval. And I am puzzled why some of us can regard any particular item as a more or less serious error, while others, equally perceptive, and acknowledging that the same item may in some sense be an "error," seem to invest in their observation no emotion at all.

At first glance, we ought to be able to explain some of these anomalies by subsuming errors of grammar and usage in a more general account of defective social behavior, the sort of account constructed so brilliantly by Erving Goffman.2 But errors of social behavior differ from errors of "good usage." Social errors that excite feelings commensurate with judgments like "horrible," "atrocious," "oaf(ish)," and "detestable" are usually errors that grossly violate our personal space: We break wind at a dinner party and then vomit on the person next to us. We spill coffee in their lap, then step on a toe when we get up to apologize. It's the Inspector Clouseau routine. Or the error metaphorically violates psychic space: We utter an inappropriate obscenity, mention our painful hemorrhoids, tell a racist joke, and snigger at the fat woman across the table who turns out to be our hostess. Because all of these actions crudely violate one's personal space we are justified in calling them "oafish"; all of them require that we apologize, or at least offer an excuse.

This way of thinking about social error turns our attention from error as a discrete entity, frozen at the moment of its commission, to error as part of a flawed transaction, originating in ignorance or incompetence or accident, manifesting itself as an invasion of another's personal space, eliciting a judgment ranging from silent disapproval to "atrocious" and "horrible," and requiring either an explicit "I'm sorry" and correction, or a simple acknowledgment and a tacit agreement not to do it again.3

To address errors of grammar and usage in this way, it is also necessary to shift our attention from error treated strictly as an isolated item on a page, to error perceived as a flawed verbal transaction between a writer and a reader. When we do this, the matter of error turns less on a handbook definition than on the reader's response, because it is that response -- "detestable," "horrible" -- that defines the seriousness of the error and its expected amendment.

But if we do compare serious nonlinguistic gaffes to errors of usage, how can we not be puzzled over why so much heat is invested in condemning a violation whose consequence impinges not at all on our personal space? The language some use to condemn linguistic error seems far more intense than the language they use to describe more consequential social errors -- a hard bump on the arm, for example -- that require a sincere but not especially effusive apology. But no matter how "atrocious" or "horrible" or "illiterate" we think an error like irregardless or a like for an as might be, it does not jolt my ear in the same way an elbow might; a between you and I does not offend me, at least not in the ordinary sense of offend. Moreover, unlike social errors, linguistic errors do not ordinarily require that we apologize for them.4 When we make media a singular or dangle a participle, and are then made aware of our mistake, we are expected to acknowledge the error, and, if we have the opportunity, to amend it. But I don't think that we are expected to say, "Oh, I'm sorry!" The objective consequences of the error simply do not equal those of an atrocity, or even of clumsiness.

It may be that to fully account for the contempt that some errors of usage arouse, we will have to understand better than we do the relationship between language, order, and those deep psychic forces that perceived linguistic violations seem to arouse in otherwise amiable people.5 But if we cannot yet fully account for the psychological source of those feelings, or why they are so intense, we should be able to account better than we do for the variety of responses that different "errors" elicit. It is a subject that should be susceptible to research. And indeed, one kind of research in this area has a long tradition: In this century, at least five major surveys of English usage have been conducted to determine how respondents feel about various matters of usage. Sterling Leonard, Albert Marckwardt, Raymond Crisp, the Institute of Education English Research Group at the University of Newcastle upon Tyne, and the American Heritage Dictionary have questioned hundreds of teachers and editors and writers and scholars about their attitudes toward matters of usage ranging from which referring to a whole clause to split infinitives to enthuse as a verb.6

The trouble with this kind of research, though, with asking people whether they think finalize is or is not good usage, is that they are likely to answer. As William Labov and others have demonstrated,7 we are not always our own best informants about our habits of speech. Indeed, we are likely to give answers that misrepresent our talking and writing, usually in the direction of more rather than less conservative values. Thus when the editors of the American Heritage Dictionary asks its Usage Panel to decide the acceptability of impact as a verb, we can predict how they will react: Merely by being asked, it becomes manifest to them that they have been invested with an institutional responsibility that will require them to judge usage by the standards they think they are supposed to uphold. So we cannot be surprised that when asked, Zinsser rejects impact as a verb, despite the fact that impact has been used as a verb at least since 1601.

The problem is self-evident: Since we can ask an indefinite number of questions about an indefinite number of items of usage, we can, merely by asking, accumulate an indefinite number of errors, simply because whoever we ask will feel compelled to answer. So while it may seem useful for us to ask one another whether we think X is an error, we have to be skeptical about our answers, because we will invariably end up with more errors than we began with, certainly more than we ever feel on our nerves when we read in the ways we ordinarily do.

In fact, it is this unreflective feeling on the nerves in our ordinary reading that interests me the most, the way we respond -- or not -- to error when we do not make error a part of our conscious field of attention. It is the difference between reading for typographical errors and reading for content. When we read for typos, letters constitute the field of attention, content becomes virtually inaccessible. When we read for content, semantic structures constitute the field of attention; letters -- for the most part -- recede from our consciousness.

I became curious about this kind of perception three years ago when I was consulting with a government agency that had been using English teachers to edit reports but was not sure they were getting their money's worth. When I asked to see some samples of editing by their consultants, I found that one very common notation was "faulty parallelism" at spots that only by the most conservative interpretation could be judged faulty. I asked the person who had hired me whether faulty parallelism was a problem in his staff's ability to write clearly enough to be understood quickly, but with enough authority to be taken seriously, He replied, "If the teacher says so."

Now I was a little taken aback by this response, because it seemed to me that one ought not have to appeal to a teacher to decide whether something like faulty parallelism was a real problem in communication. The places where faulty parallelism occurred should have been at least felt as problems, if not recognized as a felt difficulty whose specific source was faulty parallelism.

About a year later, as I sat listening to a paper describing some matters of error analysis in evaluating compositions, the same thing happened. When I looked at examples of some of the errors, sentences containing alleged dangling participles, faulty parallelism, vague pronoun reference, and a few other items,8 I was struck by the fact that, at least in some of the examples, I saw some infelicity, but no out-and-out grammatical error. When I asked the person who had done the research whether these examples were typical of errors she looked for to measure the results of extensive training in sentence combining, I was told that the definition of error had been taken from a popular handbook, on the assumption, I guess, that that answered the question.

About a year ago, it happened again, when a publisher and I began circulating a manuscript that in a peripheral way deals with some of the errors I've mentioned here, suggesting that some errors are less serious than others. With one exception, the reviewers, all teachers at universities, agreed that an intelligent treatment of error would be useful, and that this manuscript was at least in the ballpark. But almost every reader took exception to one item of usage that they thought I had been too soft on, that I should have unequivocally condemned as a violation of good usage. Unfortunately, each of them mentioned a different item.

Well, it is all very puzzling: Great variation in our definition of error, great variation in our emotional investment in defining and condeming error, great variation in the perceived seriousness of individual errors. The categories of error all seem like they should be yes-no, but the feelings associated with the categories seem much more complex.

If we think about these responses for a moment we can identify one source of the problem: We were all locating error in very different places. For all of us, obviously enough, error is in the essay, on the page, because that is where it physically exists. But of course, to be in the essay, it first has to be in the student. But before that, it has to be listed in a book somewhere. And before that in the mind of the writer of the handbook. And finally, a form of the error has to be in the teacher who resonated -- or not -- to the error on the page on the basis of the error listed in the handbook.

This way of thinking about error locates error in two different physical locations (the student's paper and the grammarian's handbook) and in three different experiences: the experience of the writer who creates the error; in the experience of the teacher who catches it; and in the mind of the grammarian -- the E. B. White or Jacques Barzun or H. W. Fowler -- who proposes it. Because error seems to exist in so many places, we should not be surprised that we do not agree among ourselves about how to identify it, or that we do not respond to the same error uniformly.

But we might be surprised -- and perhaps instructed -- by those cases where the two places occur in texts by the same author -- and where all three experiences reside in the same person. It is, in fact, these cases that I would like to examine for a moment, because they raise such interesting questions about the experience of error.

For example, E. B. White presumably believed what he (and Strunk) said in Elements of Style (New York, 1979) about faulty parallelism and which vs. that:

Express coordinate ideas in similar form. This principle, that of parallel construction, requires that expressions similar in content and function be outwardly similar. (p. 26)
That, which. That is the defining or restrictive pronoun, which the non-defining or non-restrictive . . . The careful writer . . . removes the defining whiches, and by so doing improves his work. (p 59)
Yet in the last paragraph of "Death of a Pig,"9 White has two faulty parallelisms, and according to his rules, an incorrect which:
. . . the premature expiration of a pig is, I soon discovered, a departure which the community marks solemnly on its calendar . . . I have written this account in penitence and in grief, as a man who failed to raise his pig, and to explain my deviation from the classic course of so many raised pigs. The grave in the woods is unmarked, but Fred can direct the mourner to it unerringly and with immense good will, and I know he and I shall often revisit it, singly and together, . . .
Now I want to be clear: I am not at all interested in the trivial fact that E. B. White violated one or two of his own trivial rules. That would be a trivial observation. We could simply say that he miswrote in the same way he might have mistyped and thereby committed a typographical error. Nor at the moment am I interested in the particular problem of parallelism, or of which vs. that, any more than I would be interested in the particular typo. What I am interested in is the fact that no one, E. B. White least of all, seemed to notice that E. B. White had made an error. What I'm interested in here is the noticing or the not noticing by the same person who stipulates what should be noticed, and why anyone would surely have noticed if White had written,
I knows me and him will often revisit it, . . .
Of course, it may be that I am stretching things just a bit far to point out a trivial error of usage in one publication on the basis of a rule asserted in another. But this next example is one in which the two co-exist between the same covers:
Were (sing.) is, then, a recognizable subjunctive, & applicable not to past facts, but to present or future non-facts. (p. 576)
Another suffix that is not a living one, but is sometimes treated as if it was, is -al . . . (p. 242)
H. W. Fowler. A Dictionary of Modern English Usage. Oxford, 1957.
Now again, Fowler may have just made a slip here; when he read these entries, certainly at widely separate intervals, the was in the second just slipped by. And yet how many others have also read that passage, and also never noticed?

The next example may be a bit more instructive. Here, the rule is asserted in the middle of one page:

In conclusion, I recommend using that with defining clauses except when stylistic reasons interpose. Quite often, not a mere pair of that's but a threesome or foursome, including the demonstrative that, will come in the same sentence and justify which to all writers with an ear. (p. 68)
and violated at the top of the next:
Next is a typical situation which a practiced writer corrects for style virtually by reflex action. (p. 69)
Jacques Barzun. Simple and Direct. New York, 1976.
Now again, it is not the error as such that I am concerned with here, but rather the fact that after Barzun stated the rule, and almost immediately violated it, no one noticed -- not Barzun himself who must certainly have read the manuscript several times, not a colleague to whom he probably gave the manuscript before he sent it to the publisher, not the copy editor who worked over the manuscript, not the proof reader who read the galleys, not Barzun who probably read the galleys after them, apparently not even any-one in the reading public, since that which hasn't been corrected in any of the subsequent printings. To characterize this failure to respond as mere carelessness seems to miss something important.

This kind of contradiction between the conscious directive and the unreflexive experience becomes even more intense in the next three examples, examples that, to be sure, involve matters of style rather than grammar and usage:

Negative constructions are often wordy and sometimes pretentious.
1. wordy Housing for married students is not unworthy of consideration.
concise Housing for married students is worthy of consideration.
better The trustees should earmark funds for married students' housing. (Probably what the author meant)
2. wordy After reading the second paragraph you aren't left with an immediate reaction as to how the story will end.
concise The first two paragraphs create suspense.
The following example from a syndicated column is not untypical.
Sylvan Barnet and Marcia Stubbs. Practical Guide to Writing. Boston, 1977, p. 280.
Now Barnet and Stubbs may be indulging in a bit of self-parody here. But I don't think so. In this next example, Orwell, in the very act of criticizing the passive, not only casts his proscription against it in the passive, but almost all the sentences around it, as well:
I list below, with notes and examples, various of the tricks by means of which the work of prose construction is habitually dodged . . . Operators or verbal false limbs. These save the trouble of picking out appropriate verbs and nouns, and at the same time pad each sentence with extra syllables which give it an appearance of symmetry . . . the passive voice is wherever possible used in preference to the active, and noun constructions are used instead of gerunds . . . The range of verbs if further cut down . . . and the banal statements are given an appearance of profundity by means of the not an formation. Simple conjunctions are replaced by . . . the ends of sentences are saved by . . .
"Politics and the English Language"
Again, I am not concerned with the fact that Orwell wrote in the passive or used nominalizations where he could have used verbs.10 Rather, I am bemused by the apparent fact that three generations of teachers have used this essay without there arising among us a general wry amusement that Orwell violated his own rules in the act of stating them.

And if you want to argue (I think mistakenly) that Orwell was indulging in parody, then consider this last example -- one that cannot possibly be parodic, at least intentionally:

Emphasis is often achieved by the use of verbs rather than nouns formed from them, and by the use of verbs in the active rather than in the passive voice.
A Style Manual for Technical Writers and Editors, ed. S. J. Reisman. New York, 1972. pp. 6-11.
In this single sentence, in a single moment, we have all five potential locations of error folded together: As the rule is stated in a handbook, it is simultaneously violated in its text; as the editor expresses in the sentence that is part of the handbook a rule that must first have existed in his mind, in his role as writer he simultaneously violates it. And in the instant he ends the sentence, he becomes a critical reader who should -- but does not -- resonate to the error. Nor, apparently, did anyone else.

The point is this: We can discuss error in two ways: we can discuss it at a level of consciousness that places that error at the very center of our consciousness. Or we can talk about how we experience (or not) what we popularly call errors of usage as they occur in the ordinary course of our reading a text.

In the first, the most common way, we separate the objective material text from its usual role in uniting a subject (us) and that more abstract "content" of the object, the text, in order to make the sentences and words the objects of consciousness. We isolate error as a frozen, instantiated object. In the second way of discussing error, a way we virtually never follow, we must treat error not as something that is simply on the surface of the page, "out there," nor as part of an inventory of negative responses "in here," but rather as a variably experienced union of item and response, controlled by the intention to read a text in the way we ordinarily read texts like newspapers, journals, and books. If error is no longer in the handbook, or on the page, or in the writer -- or even purely in the reader -- if instead we locate it at an intersection of those places, then we can explain why Barzun could write -- or read -- one thing and then immediately experience another, why his colleagues and editors and audience could read about one way of reflexively experiencing language and then immediately experience it in another.

But when I decided to intend to read Barzun and White and Orwell and Fowler in, for all practical purposes, the way they seem to invite me to read -- as an editor looking for the errors they have been urging me to search out -- then I inform my experience, I deliberately begin reading, with an intention to experience the material constitution of the text. It is as if a type-designer invited me to look at the design of his type as he discussed type-design.

In short, if we read any text the way we read freshman essays, we will find many of the same kind of errors we routinely expect to find and therefore do find. But if we could read those student essays unreflexively, if we could make the ordinary kind of contract with those texts that we make with other kinds of texts, then we could find many fewer errors.

When we approach error from this point of view, from the point of view of our pre-reflexive experience of error, we have to define categories of error other than those defined by systems of grammar or a theory of social class. We require a system whose presiding terms would turn on the nature of our response to violations of grammatical rules.

At the most basic level, the categories must organize themselves around two variables: Has a rule been violated? And do we respond? Each of these variables has two conditions: A rule is violated or a rule is not violated. And to either of those variables, we respond, or we do not respond. We thus have four possibilities:

la. A rule is violated, and we respond to the violation.
lb. A rule is violated, and we do not respond to its violation.

2a. A rule is not violated, and we do not respond.
2b. A rule is not violated, and we do respond.

Now, our experiencing or noticing of any given grammatical rule has to be cross-categorized by the variable of our noticing or not noticing whether it is or is not violated. That is, if we violate rule X, a reader may note it or not. But we must also determine whether, if we do not violate rule X, the same reader will or will not notice that we have violated it. Theoretically, then, this gives us four possible sets of consequences for any given rule. They can be represented on a feature matrix like this:

That is, the first kind of rule, indicated by the line marked , is of the following kind: When violated, [+V], we respond to the violation, [+R]. When it is not violated, [-V], we do not respond, [-R]. Thus the same rule results in combinations of features indicated by (a-d). Rule type is characterized by a rule that when violated, [+V], we do not notice, [-R]. But when we do not violate it, [-V], we do not notice it either, [-R]. Thus the single rule combines features indicated by (b-d). The other rules follow the same kind of grid relationships. (As I will point out later, the problem is actually much more complex than this, but this will do as a first approximation.)

I do not assert that the particular items I will list as examples of these rules are universally experienced in the way indicated. These categories are based on personal responses, and it is possible that your responses are quite different than mine. But in fact, on the basis of some preliminary research that I shall report later, I would argue that most readers respond in the ways reflected by these categories, regardless of how they might claim they react.

The most obviousest set of rules be those whose violation we instantly notes, but whose observation we entirely ignore. They are the rules that define bedrock standard English. No reader of this journal can fail to distinguish these two passages:

There hasn't been no trainees who withdrawed from the program since them and the Director met to discuss the instructional methods, if they met earlier, they could of seen that problems was beginning to appear and the need to take care of them immediate. (+V, +R)

There haven't been any trainees who have withdrawn from the program since they and the Director met to discuss the instructional methods. If they had met earlier, they could have seen that problems were beginning to appear and that they needed to take care of them immediately. (-V, -R)

Among the rules whose violation we readily note but whose observance we do not are double negatives, incorrect verb forms, many incorrect pronoun forms, pleonastic subjects, double comparatives and superlatives, most subject-verb disagreements, certain faulty parallelisms,11 certain dangling modifiers,l2 etc.

The next most obvious set of rules are those whose observation we also entirely ignore, but whose violation we ignore too. Because we note neither their observation nor their violation, they constitute a kind of folklore of usage, rules which we can find in some handbook somewhere, but which have, for the most part, lost their force with our readers. For most readers, these two passages differ very little from one another; for many readers, not at all:

Since the members of the committee had discussed with each other all of the questions which had been raised earlier, we decided to conduct the meeting as openly as possible and with a concern for the opinions of everyone that might be there. And to ensure that all opinions would be heard, it was suggested that we not limit the length of the meeting. By opening up the debate in this way, there would be no chance that someone might be inadvertently prevented from speaking, which has happened in the past. (+V, -R)

Because the members of the committee had discussed with one another all the questions that had been raised earlier, we decided to conduct the meeting in a way that was as open as possible and concerned with the opinion of everyone who might be there. To ensure that all opinions would be heard, someone suggested that we not limit the length of the meeting. By opening up the debate in this way, we would not take the chance that someone might be inadvertently prevented from speaking, something which has happened in the past. ( -V, -R)

I appreciate the fact that some readers will view my lack of sensitivity to some of these errors as evidence of an incorrigibly careless mind. Which errors go in which category, however, is entirely beside the point.13 The point is the existence of a category of "rules" to whose violation we respond as indifferently as we respond to their observance.

A third category of rules includes those whose violation we largely ignore but whose observance we do not. These are rules which, when followed, impose themselves on the reader's consciousness either subliminally, or overtly and specifically. You can sense the consequence of observing these rules in this next "minimal pair":

I will not attempt to broadly defend specific matters of evidence that one might rest his case on. If it was advisable to substantially modify the arguments, he would have to re-examine those patients the original group treated and extend the clinical trials whose original plan was eventually altered. (+V, -R)
I shall not attempt broadly to defend specific matters of evidence on which one might rest one's case. Were it advisable substantially to modify the arguments, one should have to re-examine those patients whom the original research group treated and extend the clinical trials the original plan of which was eventually altered. (-V, +R)
I appreciate that many of you believe that you notice split infinitives as quickly as you notice a subject-verb error, and that both should be equally condemned in careful prose. At the end of this paper, I will try to offer an argument to the contrary -- that in fact many -- not all -- of you who make that claim are mistaken.

The exceptions are probably those for whom there is the fourth category of error, that paradoxical but logically entailed category defined by those rules whose violation we note, and whose observance we also note. I think that very few of us are sensitive to this category, and I think for those very few, the number of items that belong in the category must, fortunately, be very small. Were the number of items large, we would be constantly distracted by noticing that which should not be noticed. We would be afflicted with a kind of linguistic hyperesthesia, noticing with exquisite pleasure that every word we read is spelled correctly, that every subject agrees with its verb, that every article precedes its noun, and so on. Many of us may be surprised when we get a paper with no mispelled words, but that pleasure does not derive from our noticing that each word in turn is correctly spelled, but rather in the absence of mispelled words.

In my own case, I think I note equally when an infinitive is split, and when it is not. In recent months, I also seem to be noticing when someone uses that in the way that the "rule" stipulates, and I notice when a writer uses which in the way which the "rule" prohibits. I hope I add no more.

I suspect that some readers put into this category the regardless / irregardless pair, media as a singular and as a plural, perhaps disinterested / uninterested. I offer no pair of contrasting examples because the membership of the category is probably so idiosyncratic that such a pair would not be useful.

Now in fact, all this is a bit more complicated than my four categories suggest, albeit trivially so. The two-state condition of response: {+/-}, is too crude to distinguish different qualities of response. Responses can be un-favorable, as the ordinary speaker of standard English would respond un-favorably to

Can't nobody tell what be happening four year from now.
if it appeared in a text whose conventions called for standard English. A response can be favorable, as in the right context, we might regard as appropriate the formality of
Had I known the basis on which these data were analyzed, I should not have attempted immediately to dissuade those among you whom others have . . .
(We could, of course, define a context in which we would respond to this unfavorably. )

Since only the category of {+response} can imply a type of response, we categorize favorable and unfavorable response, {+/-favorable}, across only {+response}. This gives us four more simple categories:

{+violate, -favorable}
{-violate, +favorable}
{+violate, +favorable}
{-violate, -favorable}
The first two I have already illustrated:
{+v, -f}: He knowed what I meaned.

{-v, +f}: Had I known the basis on which . . . I should not etc.

This leaves two slightly paradoxical categories, which, like Category IV: those rules whose violations we notice and whose observations we notice too, are populated by a very small number of items, and function as part of our responses only idiosyncratically. In the category {-violate, -favorable}, I suspect that many of us would place It is I, along with some occurrences of whom, perhaps.

The other paradoxical category, {+violate, +favorable} is not illustrated by It's me, because for most of us, this is an unremarked violation. If it elicits a response at all, it would almost invariably be {-favorable}, but only among those for whom the me is a bête noir. In fact, I can only think of one violation that I respond to favorably: It is the than after different(ly) when what follows is a clause rather than a noun:

This country feels differently about the energy crisis than it did in 1973.
I respond to this favorably because the alternative,
This country feels differently about the energy crisis from the way it did in 1973.
is wordier, and on principles that transcend idiosyncratic items of usage, I prefer the two less words and the more certain and direct movement of the phrase. My noticing any of this, however, is entirely idiosyncratic.

As I said, though, these last distinctions are increasingly trivial. That is why I refrain from pursuing another yet more finely drawn distinction: Those responses, favorable or unfavorable, that we consciously, overtly, knowingly experience, and those that are more subliminal, undefined, and unspecific. That is, when I read

It don't matter.
I know precisely what I am responding to When most of us read a shall and a shifted preposition, I suspect that we do not consciously identify those items as the source of any heightened feeling of formality. The response, favorable or unfavorable, is usually less specific, more holistic.

Now what follows from all this? One thing that does not follow is a rejection of all rules of grammar. Some who have read this far are undoubtedly ready to call up the underground grammarians to do one more battle against those who would rip out the Mother Tongue and tear down Civilized Western Values. But need I really have to assert that, just because many rules of grammar lack practical force, it is hardly the case that none of them have substance?

Certainly, how we mark and grade papers might change. We need not believe that just because a rule of grammar finds its way into some handbook of usage, we have to honor it. Which we honor and which we do not is a problem of research. We have to determine in some unobtrusive way which rules of grammar the significant majority of careful readers notice and which they do not. One way to do this research is to publish an article in a journal such as this, an article into which have been built certain errors of grammar and usage. The researcher would then ask his readers to report which errors jumped out at them on the first reading. Those that you did not notice should then not be among those we look for first when we read a student's paper.

One curious consequence of this way of thinking about error is that we no longer have to worry about defining, rejecting, quibbling over the existence of a rule. We simply accept as a rule anything that anyone wants to offer, no matter how bizarre or archaic. Should anyone re-assert the 19th-century rule against the progressive passive, fine. Upon inspection it will turn out that the rule belongs in the category of those rules whose violation no one notices, and whose observation no one notices either. As I said, it may be that you and I will find that for any particular rule, we experience its violation in different ways. But that is an empirical question, not a matter of value. Value- becomes a consideration only when we address the matter of which errors we should notice.

Done carefully, this kind of classification might also encourage some dictionary makers to amend their more egregious errors in labeling points of usage. The AHD, for example, uses "non-standard" to label

. . . forms that do not belong in any standard educated speech. Such words are recognized as non-standard not only by those whose speech is standard, but even by those who regularly use non-standard expressions. ( )
The AHD staff has labeled as non-standard, ain't, seen as the past tense of see, and don't with a singular subject. It has also labeled as non-standard irregardless, like for as, disinterested for uninterested, and see where, as in the construction, I see where . . . Thus we are led to believe that a speaker who would utter this:
I see where the President has said that, irregardless of what happens with the gasoline shortage, he'll still be against rationing, just like he has been in the past. He seems disinterested in what's going on in the country.
would be just as likely to continue with this:
I ain't sure that he seen the polls before he said that. He don't seem to know that people are fed up.
Indeed, we would have to infer from this kind of labeling that a speaker who said "I ain't sure he seen . . . " would also be sensitive to mistakes such as disinterested for uninterested or like for as. In matters such as this, we see too clearly the very slight scholarly basis upon which so much of this labeling rests.

Finally, I think that most of this essay is an exercise in futility. In these matters, the self-conscious report of what should be counted as an error is certainly an unreliable index to the unself-conscious experience. But it is by far a more satisfying emotion. When those of us who believe ourselves educated and literate and defenders of good usage think about language, our zealous defense of "good writing" feels more authentic than our experience of the same items in unreflective experience of a text. Indeed, we do not experience many of them at all. And no matter how wrong we might discover we are about our unreflective feelings, I suspect we could be endlessly lectured on how we do not respond to a less in front of a count noun, as in less people, but we would still express our horror and disgust in the belief that less is wrong when used in that way. It simply feels more authentic when we condemn error and enforce a rule. And after all, what good is learning a rule if all we can do is obey it?

If by this point you have not seen the game, I rest my case. If you have, I invite you to participate in the kind of research I suggested before. I have deposited with the Maxine Hairston of the University of Texas at Austin (Austin, Texas 78712), a member of the Editorial Board of this journal, a manuscript with the errors of grammar and usage that I deliberately inserted into this paper specifically marked. How can I ask this next question without seeming to distrust you? If you had to report right now what errors you noticed, what would they be? Don't go back and reread, looking for errors, at least not before you recall what errors you found the first time through. If you would send your list (better yet, a copy of the article with errors noted on first reading circled in red) to Professor Hairston, she will see that a tally of the errors is compiled, and in a later issue will report on who noticed what.

If you want to go through a second time and look for errors, better yet. Just make clear, if you would, that your list is the result of a deliberate search. I will be particularly interested in those errors I didn't mean to include. There are, incidentally, about 100 errors.

[Note from the Web editor: I have not had access to the list of errors, and in scanning and transcribing this some may have been inadvertently corrected, others introduced. If you know of instances of this, please let me know, at]


1. I don't know whether it is fair or unfair to quote Zinsser on this same matter:

OVERSTATEMENT. "The living room looked as if an atomic bomb had gone off there," writes the inexperienced writer, describing what he saw on Sunday morning after a Saturday night party that got out of hand. Well, we all know that he's exaggerating to make a droll point, but we also know that an atomic bomb didn't go off there, or any other bomb except maybe a water bomb . . . . These verbal high links can get just so high -- and I'm already well over the limit -- before the reader feels an overpowering drowsiness . . . . Don't overstate. (p. 108)

2. Erving Goffman, Frame Analysis: An Essay on the Organization of Experience. (New York: Harper and Row, 1974).

3. Some social errors are strictly formal and so ordinarily do not require an apology, even though some might judge them "horrible": a white wedding gown and a veil on a twice-divorced and eight-month pregnant bride, brown shoes with a dinner jacket, a printed calling card.

4. Some special situations do require an apology: When we prepare a document that someone else must take responsibility for, and we make a mistake in usage, we are expected to apologize, in the same way we would apologize for incorrectly adding up a column of figures. And when some newspaper columnists violate some small point of usage and their readers write in to point it out, the columnists will often acknowledge the error and offer some sort of apology. I think William Safire in The New York Times has done this occasionally.

5. Two other kinds of purely linguistic behavior do arouse hostile feelings. One kind includes obscenities and profanities. It may be that both are rooted in some sense of fouling that which should be kept clean: obscenities foul the mouth, the mouth fouls the name of a deity. The other kind of linguistic behavior that arouses hostility in some includes bad puns and baby talk by those who are too old for it. Curiously, Freud discusses puns in his Wit and the Relation to the Unconscious (under "Technique of Wit") but does not in "The Tendencies of Wit" address the faint sense of revulsion we feel at a bad pun.

6. Sterling Leonard, Current English Usage, English Monograph No. 1 (Champaign, Ill.: National Council of Teachers of English. Chicago, 1932); Albert H. Marckwardt and Fred Walcott, Facts About Current English Usage, English Monograph No. 7 (Champaign, Ill.: National Council of Teachers of English. New York, 1938); Raymond Crisp. "Changes in Attitudes Toward English Usage," Ph.D. dissertation, University of Illinois, 1971; W. H. Mittins, Mary Salu, Mary Edminson, Sheila Coyne, Attitudes to English Usage (London: Oxford University Press, 1970); The American Heritage Dictionary of the English Language (New York: Dell, 1979). Thomas J. Cresswell's Usage in Dictionaries and Dictionaries of Usage, Publication of the American Dialect Society, Nos. 63-64 (University, Alabama: University of Alabama Press 1975), should be required reading for anyone interested in these matters. It amply demonstrates the slight scholarly basis on which so much research on usage rests.

7. William Labov, The Social Stratification of English in New York City (Washington, D.C.: Center for Applied Linguistics, 1966), pp. 455-81.

8. Elaine P. Maimon and Barbara F. Nodine, "Words Enough and Time: Syntax and Error One Year After," in Sentence Combining and the Teaching of Writing, eds. Donald Daiker, Andrew Kerek, Max Morenberg (Akron, Ohio: University of Akron Press, 1979) pp. 101-108. This is considered a dangling verbal: For example, considering the way Hamlet treats Ophelia, there is almost corruptness in his mind. Clumsy yes, but considering is an absolute, or more exactly, metadiscourse. See footnote 12. This is considered a vague pronoun reference: The theme of poisoning begins with the death of old King Hamlet, who was murdered by his brother when a leperous distilment was poured into his ear while he slept. Infelicitous, to be sure, but who can possibly doubt who's pouring what in whose ear (p. 103)? Counting items such as these as errors and then using those counts to determine competence, progress, or maturity would seem to raise problems of another, more substantive, kind.

9. Essays of E. B. White (New York: Harper and Row, 1977), p. 24. 9

10. Orwell's last rule: Break any of these rules sooner than say anything outright barbarous, does not apply to this passage. Indeed, it would improve if it had conformed to his rules:

I list below, with notes and examples, various of the tricks by means of which a writer can dodge the work of prose construction . . . such writers prefer wherever possible the passive voice to the active, and noun constructions instead of gerunds . . . they further cut down the range of verbs . . . they make their banal statements seem profound by means of the not un- formation. They replace simple conjunctions by . . . they save the ends of sentences . . .
Should anyone object that this is a monotonous series of sentences beginning with the same subject, I could point to example after example of the same kind of thing in good modern prose. But perhaps an example from the same essay, near the end, will serve best (my emphasis):
When you think of a concrete object, you think wordlessly, and then, if you want to describe the thing you have been visualizing, you probably hunt about till you find the exact words that seem to fit it. When you think of something abstract you are more inclined to use words from the start, and unless you make a conscious effort to prevent it, the existing dialect will come rushing in and do the job for you . . .
Nine out of ten clauses begin with you, and in a space much more confined than the passage I rewrote.

11. Virtually all handbooks overgeneralize about faulty parallelism. Two "violations" occur so often in the best prose that we could not include them in this Category I. One is the kind illustrated by the E. B. White passage: the coordination of adverbials: . . . unerringly and with immense good will. The other is the coordination of noun phrases and WH-clauses: We are studying the origins of this species and why it died out. Even that range of exceptions is too broadly stated, but to explain the matter adequately would require more space than would be appropriate here.

12. Handbooks also overgeneralize on dangling constructions. The generalization can best be stated like this: When the implied subject of an introductory element is different from the overt subject of its immediately following clause, the introductory element dangles. Examples in handbooks are always so ludicrous that the generalization seems sound:

Running down the street, the bus pulled away from the curb before I got there.

To prepare for the wedding, the cake was baked the day before.

Some handbooks list exceptions, often called absolutes:
Considering the trouble we're in, it's not surprising you are worried.

To summarize, the hall is rented, the cake is baked, and we're ready to go.

It exceptions can be subsumed into a more general rule: When either the introductory element or the subject of the sentence consists of meta-discourse, the introductory element will not always appear to dangle. By meta-discourse I mean words and phrases that refer not to the primary content of the discourse, to the reference "out there" in the world, the writer's subject matter, but rather to the process of discoursing, to those directions that steer a reader through a discourse, those filler words that allow a writer to shift emphasis (it, there, what ), and so on, words such as it is important to note, to summarize, considering these issue, as you know, to begin with, there is, etc. That's why an introductory element such as the following occurs so often in the prose of educated writers, and does not seem to dangle (meta-discourse is in bold face):
To succeed in this matter, it is important for you to support as fully as possible . . .

Realizing the seriousness of the situation, it can be seen that we must cut back on . . .

As I will point out later, the categories I am suggesting here are too broadly drawn to account for a number of finer nuances of error. Some violations, for example, clearly identify social and educational background:
He didn't have no way to know what I seen.
But some violations that might be invariably noted by some observers do not invariably, or even regularly, reflect either social or educational background. Usages such as irregardless, like for as, different than,etc. occur so often in the speech and writing of entirely educated speakers and writers that we cannot group them with double negatives and non-standard verb forms, even if we do unfailingly respond to both kinds of errors. The usage note in the American Heritage Dictionary (Dell Paperback Edition, 1976; third printing, November, 1980) that irregardless is non-standard and "is only acceptable when the intent is clearly humorous" is more testimony to the problems of accurately representing the speech and writing of educated speakers. On February 20, 1981, the moderator on Washington Week in Review, a Public Broadcasting System news program, reported that a viewer had written to the program, objecting to the use of irregardless by one of the panelists. To claim that the person who used irregardless would also use knowed for knew or an obvious double negative would be simply wrong. (I pass by silently the position of only in that usage note. See footnote 13, item 9.) The counter-argument that the mere occurrence of these items in the speech and writing of some is sufficient testimony that they are not in fact educated is captious.

13. Here are some of the rules which I believe belong in this Category II: (1) Beginning sentences with and or but; (2) beginning sentences with because (a rule that appears in no hand-book that I know of, but that seems to have a popular currency); (3) which / that in regard to restrictive relative clauses; (4) each other for two, one another for more than two; (5) which to refer to a whole clause (when not obviously ambiguous); (6) between for two, among for more than two. These next ones most readers of this journal may disagree with personally; I can only assert on the basis of considerable reading that they occur too frequently to be put in any other category for most readers: (7) less for fewer; (8) due to for because, (9) the strict placement of only; (10) the strict placement of not only, neither, etc. before only that phrase or clause that perfectly balances the nor. The usage of several disputed words must also suggest this category for most readers: disinterested / uninterested, continuous / continual, alternative for more than two. Since I have no intention of arguing which rules should go into any category, I offer these only as examples of my observations. Whether they are accurate for you is in principle irrelevant to the argument. Nor is it an exhaustive list.

14. The rules that go into Category III would, I believe, include these. Again, they serve only to illustrate. I have no brief in regard to where they should go. (1) shall / will, (2) who / whom, (3) unsplit infinitives, (4) fronted prepositions, (5) subjunctive form of be, (6) whose / of which as possessives for inanimate nouns, (7) repeated one instead of a referring pronoun he / his / him, (8) plural data and media, singular verb after none.

Author Note:

Joseph M. Williams is Professor of English and Linguistics at the University of Chicago. He is the author of Style: Ten Lessons in Clarity & Grace. He is now pursuing the idea of form in non-fictional prose. This paper is based on a paper he presented at CCCC in 1980. sert on the basis of considerable reading that they occur too frequently to be put in any other category for most readers: (7) less for fewer; (8) due to for because, (9) the strict placement of only; (10) the strict placement of not only, neither, etc. before only that phrase or clause that perfectly balances the nor. The usage of several disputed words must also suggest this category for most readers: disinterested / uninterested, continuous / continual, alternative for more than two. Since I have no intention of arguing which rules should go into any category, I offer these only as examples of my observations. Whether they are accurate for you is in principle irrelevant to the argument. Nor is it an exhaustive list. 

14. The rules that go into Category III would, I believe, include these. Again, they serve only to illustrate. I have no brief in regard to where they should go. (1) shall / will, (2) who / whom, (3) unsplit infinitives, (4) fronted prepositions, (5) subjunctive form of be, (6) whose / of which as possessives for inanimate nouns, (7) repeated one instead of a referring pronoun he / his / him, (8) plural data and media, singular verb after none.

Author Note:

Joseph M. Williams is Professor of English and Linguistics at the University of Chicago. He is the author of Style: Ten Lessons in Clarity & Grace. He is now pursuing the idea of form in non-fictional prose. This paper is based on a paper he presented at CCCC in 1980.