I have completed a reanalysis of the verbs in President Obama's speech after the BP oil disaster, and can add a further note to Mark's analysis of Kathleen Parker's unbelievably irresponsible prattle about how the frequency of passive constructions chosen by his speechwriters shows that President Obama talks like a girl (is "suffering a rhetorical-testosterone deficit").
I can report that I found a way of counting under which one can vindicate Paul JJ Payack's 13 percent figure, which Mark found inexplicable. But a morass of inexplicable stupidity remains nonetheless.
I find it hard to say which of three things appals me the most in this story: the idiocy of assuming without evidence that females use the passive voice more; the irresponsibility of suggesting that the President of the United States is talking wimpy on the basis of that unsupported idea; or the dishonesty of a national newspaper opinion column pretending to have linguistic evidence for a point of view without taking responsibility for checking that the evidence exists.
Kathleen Parker cited Payack as the source for the claim that Obama's post-oil-spill speech "featured 13 percent passive-voice constructions". Trying to assess the truth of this, Mark counted up the sentences in the speech (there are 135) and scored a sentence as passive if it had one or more tensed constructions containing passive clauses in it anywhere. He came up with 11%, and noted that when the same methodology was applied to a comparable post-disaster speech by George W. Bush he came up with closer to 18%.
For my first attempt at replicating Payack I counted on a more fine-grained basis. I counted the lexical verbs in the text of the speech that Obama read — that is, the verbs that are not auxiliaries. Auxiliaries can begin a closed interrogative clause, as in Have you eaten?, and can take the n't suffix as in I haven't eaten, whereas lexical verbs don't have these properties. The reason for counting only lexical verbs is that they are the only ones that are potential candidates for appearing as the verbs of passive clauses. Auxiliary verbs can never be heads of passive clauses.
I counted conservatively: no occurrence of the copular verb be was included, even if it was the only verb in its sentence (because the copula can always occur before the subject in an interrogative). And no verb used as a modifier in a noun phrase was counted (so in the approaching oil I did not count approaching as a verb, although it is in fact a verb in participial form functioning as an attributive modifier). I found 519 verbs all told. The lexical verbs number 331.
I then counted passive clauses as defined in The Cambridge Grammar of the English Language. There are 29 passive clauses. (I don't include the copula that often precedes a passive clause as part of that clause, so sometimes the only constituent in a passive clause is the verb that is its head. I exhibit all of the relevant data below, in case you want to see what falls under the definition of a passive clause.)
Now, the ratio of passive clauses to lexical verbs is 29/331 = 8.76%, so that's too small to replicate Payack.
And the ratio of passive clauses to sentences, on the other hand, is 29/135 = 21.5%, so that's too big.
Returning to the data, therefore, I tried removing the intransitive lexical verbs from the count, and considering only transitive ones — those that have direct objects. (Some details: I included ditransitives, which have two objects, but I did not include potentially transitive verbs when used intransitively. Thus I would count help as transitive in help our neighbors but not in help stop the oil; I would count make as transitive in make no mistake or make BP pay but not in make sure it never happens again. Thus again I counted conservatively, pushing the passive percentage up.) There are 206 transitive verbs in all, and 177 of them are used in the active voice.
So the ratio of passive clauses to transitive verbs is 29/206 = 14%. That gets us close to Payack's number. And since we already know from Ben Zimmer that Payack isn't strong on telling passives from actives, all we have to assume is that he either missed a few or misanalyzed some intransitive verbs as transitive. Adding a dozen misdiagnosed verbs to the total considered would bring the ratio to about 13%. So that might be the explanation of how Payack gets his figure.
We could now proceed (though it does take an hour or two of hard work per speech) to compare Obama's speech with one by George W. Bush, and get, no doubt, a result paralleling Mark's (that Bush is way higher in passive frequency), since my methodology is not likely to change the relative numbers as opposed to the absolute percentage.
However, it is important to see that to engage in such numerology is in a sense just playing along with foolishness and irresponsibility. The real point emerges if you look at the passive examples themselves. I will list all of them, with the passive clause (the passive participle together with its complements including the agent by-phrase if there is one) underlined in each phrase quoted. Judge for yourself the extent to which these phrases look as if they were "used to either deflect responsibility, or to have no particular 'doer' of an action" — Payack's version of a familiar but thoroughly ignorant claim about the function of passives:
Passive clauses (underlined) in Obama's post-oil-leak speech
- Seventeen others were injured
- a team led by Dr. Steven Chu
- a relief well … that's expected to stop the leak
- an effort led by Admiral Thad Allen
- millions of gallons of oil have already been removed from the water
- five and a half million feet of boom has been laid
- the second thing we're focused on
- areas where the beaches are not yet affected
- their way of life may be lost
- whatever resources are required
- the workers and business owners who have been harmed
- this fund will not be controlled by BP
- to ensure that all legitimate claims are paid
- the account must and will be administered by an independent third party
- The plan will be designed by states, local communities, tribes, fishermen, businesses, conservationists and other Gulf residents
- the necessary precautions would be taken
- known as the Minerals Management Service
- a philosophy that says corporations should be allowed to play by their own rules and police themselves
- industry insiders were put in charge of industry oversight
- Oil companies … were essentially allowed to conduct their own safety inspections
- the days of cheap and easily accessible oil were numbered
- the path forward has been blocked — not only by oil industry lobbyists, but also by a lack of political courage and candor
- an entire way of life being threatened by a menacing cloud of black crude
- there are costs associated with this transition
- the same thing was said about our ability to produce enough planes and tanks in World War II
- The same thing was said about our ability to harness the science and technology to land a man safely on the surface of the moon
- a tradition that was brought to America long ago by fishing immigrants from Europe
- It's called "The Blessing of the Fleet"
- a blessing that's granted
Tell me the truth: can you truly say that you think phrases like seventeen were injured, or a team led by Dr. Chu, or expected to stop the leak, or a way of life may be lost, or the resources that are required, or his days are numbered, or threatened by a menacing cloud, or costs associated with it, or brought from Europe, sound girly?
Are Payack and Parker completely brainless? Didn't they glance at the text of the speech and think about how men talk? Did the Washington Post truly imagine something serious was being said about men's as opposed to women's styles of speech?
And girliness aside, is the alleged tendency to "deflect responsibility" in evidence here? Take a sentence like At this agency [the Minerals Management Service—GKP], industry insiders were put in charge of industry oversight. Obama's passive clause is clearly locating responsibility: whoever was in charge on any particular occasion at the Minerals Management Service, Obama is accusing them of toadying to oil companies and making sure no stringent supervision would occur. Or take the sentence And this fund will not be controlled by BP. It's a passive construction, sure; but it directly and explicitly takes responsibility, making a promise that whoever controls the fund it will not be BP.
Interpreting a raw frequency count of passive clauses as a measure of shiftiness or evasion is outright and obvious stupidity. But that's what Payack does. Parker merely stretches things to draw an even sillier conclusion (one that Payack cannot be blamed for) by confusing use of passive clauses with speaking like a woman.
We have said it before on Language Log, and I'll say it again now: when you find journalists and columnists telling you things that have anything to do with language, put your hand on your wallet, because honesty and integrity are about to go out the window. Where language is concerned, people simply make stuff up.