5 Jul 2007

Documenting the College Board's Mistakes

Submitted by Karl Hagen
This is part 2 of a series.
(Part 1, Part 3)

The instructions for the writing sections of the the SAT ask the test taker to use the conventions of "standard written English" in finding correct answers. But as I noted in part 1 of this article, the public specifications as to what that means, are vague.

To fill in the lacuna, an obvious place to turn is the College Board's online SAT course, which provides detailed explanations to the multiple-choice writing problems of fifteen different practice tests. Because they use a certain level of technical vocabulary, these explanations provide valuable insight into how the test makers view the subject matter. This insight, however, is not one that inspires great confidence in the test-makers understanding of grammatical principles.

The explanations adopt a grammatical framework that is largely traditional. Many are consistent with the terminology used in Warriner (1965). Even taking the traditional account as given, the explanations contain numerous inconsistencies and a troubling number of outright errors. Beyond internal inconsistencies, there are a number of well-known flaws in the traditional scheme that the College Board uses. Most of these flaws are not of recent discovery. They are well documented in standard reference works such as Quirk et al. (1985) and Huddleston and Pullum (2002), works that anyone expert in English grammar should be intimately familiar with. Indeed, some of the explanations evidence familiarity with at least some aspects of these more up to date works, which makes the choice of an obsolete scheme all the more unaccountable. These flaws are not just matters of arcane technical differences in opinion. Following the flawed account can lead, and in some cases clearly has lead, to incorrect reasoning about whether an item is acceptable or not, result in unfairly arbitrary questions.

I will treat these different types of problems in turn, beginning with outright errors, then considering internal inconsistencies, and finally turning to fundamental inadequacies with the grammatical framework chosen.


The following are errors of analysis or assertion no matter what specific grammatical framework one follows. References to test questions and explanations are as described in part 1.

Pronouns are said to have subjects

"The singular pronoun 'its' agrees with its singular subject, 'each'" (SG-4.6.16).

In this particular case, each is in the subject of the clause, but pronouns don't have subjects, they have antecedents. See also OC-1.7.14.

Participles are said to have tense

"The present tense of the verbal form 'being' is inconsistent with the past tense of the verb 'avoided'" (SG-8.3.3).

The writer has undoubtedly been mislead by the term present participle, but participles are tenseless forms of the verb (Huddleston & Pullum, 2002, p. 80). If the explanation were true, we would have to say that sentences such as the following involved an unacceptable tense shift:

"Following in her mother's footsteps, Jane became an evolutionary biologist."

The same confusion is repeated at SG-6.4.2, OC-3.10.3, OC-3.10.13, OC-4.5.9, OC-5.6.5, OC-5.6.19

Verb tense/aspect is misclassified

In the traditional scheme used by the College Board aspect is not distinguished from tense, but the following are mistakes in any analytical scheme:

In OC-p.5.17, "had been" is called past progressive. It is, in fact, past perfect.
In OC-2.6.31, "have achieved" is called a "completed past tense," by which the author means "past perfect." It is, in fact, present perfect.
In OC-6.5.2, "had been" is called a present perfect. It is, in fact, a past perfect.

Modal auxiliaries and past-tense verbs are said to agree with subjects

"The verb phrase 'must compete' agrees with its subject, 'they.'" (OC-1.7.17)

Modal auxiliaries do not vary their form to match the number of the subject, and therefore cannot be said to agree (Huddleston & Pullum, 2002, p. 499). See also OC-4.5.26 for the same confusion.

The same objection applies to past-tense verbs (other than forms of be), and yet in OC-2.6.25, we are told,
"The singular pronoun 'he,' referring to 'M.R. Harrington,' agrees with the singular verb 'named.'"

The subject of a sentence in the passive voice is said to perform an action

The sentence in question: "After two terms in the Texas State Senate, Barbara Jordan elected to the United States House of Representatives, where she served from 1973 to 1979." (OC-4.5.19)

The explanation: "A main verb ('was elected') is needed to indicate that Barbara Jordan, the subject of the sentence, is performing the action described in the sentence."

'Was' is required because the predicate needs a tensed verb. Barbara Jordan is not performing an action in this sentence (i.e., not the actor). She experiences the election. Indeed, traditional definitions of the passive voice stress this semantic relationship as the passive voice's signal feature.

Adjectives are classified as verbs

"The singular verb, 'was able,' agrees with the singular subject, 'doctor.'" (SG-6.4.15.B)

Only was is a verb here. Able is an adjective.

Content clauses are mistaken for relative clauses

In the following sentence, the that-clause is incorrectly called a relative clause:

"The educator’s remarks stressed that well-funded literacy programs are needed if everyone is to gain the skills required for survival in society." (OC-6.5.10)

Distinguishing relative and content clauses is something that always gave my students trouble, but it's actually not particularly esoteric a distinction, and anyone who is qualified enough to write this sort of explanatory key ought to know the difference. There is no pronoun or gap in this subordinate clause. Notice that "well-funded literacy programs are needed...society" is structurally complete. Nothing has been replaced with a relative pronoun. That in this clause is just a subordinator.

The same mis-analysis occurs in SG-6.4.26, OC-1.7.16, OC-1.10.6, OC-6.5.22

Adverbs are confused with pronouns

"Where," a relative adverb, is mis-labeled in SG-3.10.11, OC-p.5.17, OC-p.5.23. "There" is mis-labeled in SG-4.10.6.

Relative pronouns are labeled conjunctions

See SG-7.3.16.

Demonstratives are labeled relative pronouns

See SG-1.7.28, SG-5.4.1, SG-5.4.14, SG-7.3.11.

Whether you want to call that in a construction like "that of his ancestors" a demonstrative determiner (Huddleston & Pullum, 2002, p. 352) or a demonstrative pronoun, as in the traditional account, it is certainly not a relative pronoun.

Corrections are suggested that result in ungrammatical sentences

Question SG-5.10.12 reads as follows with choice E substituted in (text from choice E is underlined):

Giraffes born with very long necks were able to stay alive when food was scarce and therefore have this desirable trait inherited in their offspring.

The explanation claims that this choice "involves inconsistent verb tenses. The verb 'have' is not consistent in tense with the rest of the verbs in the sentence. It should simply be 'were.'"

The verb tense is inconsistent, but replacing "have" with "were" is, if anything, worse ("and therefore were this desirable trait inherited...") Calling the tense inconsistent is an inadequate explanation anyway, since changing the tense of "have" to "had" makes the tense consistent, but does not fix the basic problem with the resultative use of 'had' (it sounds illogically as if the inheritance was the outcome of deliberate intent, e.g., "He had Bill fired.")

Another example of an inadequate correction can be found in OC-6.5.4. The sentence (again with choice E substituted) reads:

Many psychologists do not use hypnosis in their practices, their knowledge of it being very little results in wariness of it.

The explanation reads, "Choice (E) involves improper pronoun case. "It" should be "its."
Although there are two its in this sentence, the first one must be intended. The author of the explanation is undoubtedly thinking of the traditional rule for the subject of a gerund. Yet making this substitution produces no improvement in the sentence:

"...their knowledge of its being very little results in wariness of it."

To the extent that this version makes any sense at all, it changes the sentence's meaning in unintended and illogical ways, appearing to suggest that the psychologists know that hypnosis is little. The explanation writer has mis-analyzed the sentence, treating "it being very little" as a single constituent, when the correct grouping is as follows:

[their knowledge of it] [being [very little]]

In other words, this is what is traditionally known as an absolute construction.

Moreover, the change doesn't do anything about the fundamental problem with this version of the sentence: apart from its painful awkwardness, it contains a comma splice.

Assertions about language are made that clearly derive solely from intuition, with no effort made at verification

In SG-3.6.8, to explain why the answer choice "Telling stories giftedly" does not work, the assertion is made that "'[g]iftedly' is not a word." According to my dictionary, that's simply false. It is an uncommon word, but it does appear from time to time in published prose. For example from a 1947 article in Time:

The aging Welsh wonder-boy of British journalism was as giftedly gabby as ever but no longer so leftish.

Or the following comment about William McGonagall from Wikipedia (quoting Stephen Pile, The Book of Heroic Failures):

He also campaigned vigorously against excessive drinking, appearing in pubs and bars to give edifying poems and speeches. These were very popular, the people of Dundee possibly recognising that McGonagall was "so giftedly bad he backed unwittingly into genius".

The actual problem with "giftedly" in this context is not that it isn't a word, but that it appears to work much better as a modifier of adjectives than of verbs.

Another instance of jumping to false conclusions occurs in SG-6.4.8, objecting to the form acuter: "the comparative form of 'acute' is 'more acute.'" Two-syllable adjectives follow no wholly regular rule as to whether they can inflect for grade (Huddleston & Pullum, 2002, pp. 1583-4), and in such cases, it helps to see what established authors do. A quick search of Project Gutenberg turns up many hits by authors such as Saki, Thomas Hardy, and Bulwer-Lytton. The instance in Saki, from When William Came, will suffice to show that acuter is entirely standard:

When one is abroad, among foreigners, one's country's misfortunes cause one an acuter, more personal distress, than they would at home even.

A fused sentence is labeled "not a sentence."

See SG-5.10.7. A run-on sentence is still a variety of sentence, just not one that is correctly punctuated.

Some explanations are non sequiturs

The question (choice E substituted for the original text):

Samuel Adams was by no means the first American to espouse the democratic cause, but he having been the first to conceive the party machinery that made it practical. (OC-6.5.2)

The explanation: "Choice (E) creates a fragment. There is no subject for the verb 'having been.'"

The initial observation is true. This version is a fragment. But the reason given to explain why it is a fragment does not follow. There is a subject: he. The sentence is a fragment because having been has no tense.

Typographical errors abound

It should be noted that there are a number of typographical errors that suggest these explanations have not undergone editorial scrutiny. There are a noteworthy number of spelling errors, which are particularly surprising, since these are the sort of problems that an ordinary word processor should flag. A few examples: "campagning" (SG-1.7.32), "addesses" (SG-2.6.30); "Morever" (SG-2.6.35); "Athough" (SG-5.10.2, twice), "prounoun" (SG-5.10.8) "prepositon" (SG-7.3.8) "setence" (SG-7.3.9); "pharse" (SG-7.3.14); " awkard" (SG-7.10.8)

In some cases explanations have obviously been pasted into the wrong spot. For example, the explanation for choice C in SG-1.7.19 is a word-for-word copy of the explanation for choice A in the same question, referring to the pronoun "which," although the words that require explaining are completely different ("raised by"). Some of the errors enumerated above may therefore be the result of carelessness rather than actual mis-analysis. Nonetheless, they create confusion for anyone reading the explanations.

Next time, inconsistent explanations.