Passive Verbosity Again

I have a correspondent I call Faxman who is a professor of accounting. He has the laudable desire to improve his M.B.A. students’ ability to write clear prose. This is a worthy endeavor, and I was rather shocked to learn that his efforts have led to (can you believe this?) complaints from students and a warning from his dean.

Faxman advises his students to avoid the passive. He wrote to me accusing me of straw-man argumentation in my recent paper on usage authorities’ hatred of passives, but what he meant (I consulted him via email) was that I discussed arguments from the usage literature that seemed feeble to him (about passives being weak, static, dull, vague, cowardly, bureaucratic, evasive, avoided by good writers, and so on), but did not discuss his favorite.

Faxman’s favorite is the argument from verbosity: Passives are more wordy, he thinks, so changing passives to actives will shorten a text in a way that leaves meaning unchanged but improves readability. I didn’t address that point because it seemed to me self-evidently mistaken, but I’m happy to do so.

Faxman’s own presentation of the length point seemed to me “a shocking piece of numerical dishonesty coming from an accounting professor”; but I discovered to my surprise that Faxman did not follow my reasoning (briefly adumbrated in last week’s post). I must assume my explanation was not clear. Perhaps I should have exhibited a table of active clauses, word-counts, and percentage length increases for the corresponding passives:

John loves Mary. 3 67%
All cows eat grass. 4 50%
Pigs usually enjoy eating acorns. 5 40%
Several monkeys were using old typewriters. 6 33%
. . . . . . . . .
In Chomsky’s view the rise of supranational corporations committed to the globalization of profit-seeking and the depredation of the earth’s resources poses entirely new dangers both for human freedom and for the natural environment. 34 < 6%
. . . . . . . . .

You can see that the percentage increase in length of the passive shrinks as sentence length expands. In the limit it disappears, much like the $3 charge that a venal ATM at the Newark Liberty International Airport charged me last month (on a $20 withdrawal that’s 15 percent, but on $400 it’s only 0.75 percent, and in the limit it’s nothing).

Some back-of-envelope calculations show in more detail why changing passives to actives in a typical text saves only a trivial number of words. Assume an average sentence length of 20 words (this is about right for nonfiction writing). That’s 50 sentences per 1,000 words.

But sentences can have more than one clause, and any clause might have a transitive head verb usable in a passive construction. Let’s assume that each sentence has on average two passivizable transitive verbs (I think that is quite generous). In each 1,000 words, then, I’ll assume there might be 100 potentially passivizable verbs.

In typical prose only about 12 percent of those verbs will in fact head passive clauses; but Faxman is concerned with writing that overuses the passive, so let’s assume 15 verbs heading passive clauses in each 1,000 words.

Of those 15, only some will be long passives, with a by-phrase. My investigations suggest it may be 5 percent or even below that. But let’s be generous again, and not just double that but quadruple it: I’ll assume, implausibly, that 20 percent of the 15 verbs are heads of long passives. That makes three long passives per 1,000 words. (For the short passives you typically get no saving or an actual increase: changing In 1963 Kennedy was assassinated to In 1963 Oswald assassinated Kennedy doesn’t alter the word count, and naming the culprit as Lee Harvey Oswald would actually increase the word count from 5 to 7.)

For each of the long passives, changing them to their active counterparts will save either one word (by) or two (if there is a be). In a long passive clause without be, such as The dictator had the coup leader taken away by soldiers, the active version (The dictator had soldiers take the coup leader away) has only one word less, but I’ll ignore that, and assume every long passive has a be as well as a by. That amounts to pretending that we’ll save six words per 1,000 words of text (I think it would in practice be much less).

Under one set of quite reasonable assumptions that are generous to my opponent, therefore, the shrinkage of typical text will be at most 0.6 percent.

The belief that you can make text noticeably more concise simply by changing all the passives to active is fanciful. It may well be good to Omit Needless Words and write as tersely as you can, but eliminating passives will have almost no role to play in that virtuous enterprise.

Return to Top