Last night, the NYT's headline on this AP wire story was "Prosecutors to Reportedly Drop Charges Against Bryant". This morning, it reads "Bryant Charge Dropped; Civil Suit Looms." It changed because because the passage of time changed the perspective, but perhaps it should have changed anyway, for grammatical reasons.
It's not the split infinitive that's at issue here. There's no logical or grammatical reason to forbid splitting infinitives, and sometimes it's even obligatory, as Arnold Zwicky and Geoff Nunberg pointed out here last spring.
Rather, I'm interested in the scope of the adverb. Jeff Erickson at Ernie's 3D Pancakes discussed this question in connection with his tenure letter, which said that "we invite you formally to indicate your acceptance". Jeff asked "Did they formally invite me to indicate my acceptance, or did they invite me to formally indicate my acceptance?", pointing out that these are different things.
It could happen, I guess, that some prosecutors might really form the intention of taking an action that could be described as reportedly dropping some charges -- and perhaps, by pragmatic implication, not really dropping them. But the headline "Prosecutors to Reportedly Drop Charges" is not describing such a Byzantine situation: it clearly and simply means that the prosecutors are planning to drop the charges, but have not yet formally announced it. In due course, the announcement was made, and the headline duly changed to "Bryant charge dropped".
Searching the web for headlines of the form [ NP to reportedly V ] ... or [NP reportedly to V ... ], we can find plenty of both types:
Goldman to reportedly lead Lazard IPO
Thailand's Prime Minister to Reportedly Buy Stake in Liverpool
Mo. 'Pool' Businesses to Reportedly Save $3 Million from 3.7 Percent Workers' Comp Rate CutPennington reportedly to become highest paid Jet
NIH reportedly to release guidelines allowing research on human embryo cells
Paris Hilton sex video reportedly to be offered via new Internet porn site
It's clear that both forms are meant to be interpreted in the same way, expanding the telegraphic headline by supplying a form of to be in front of the infinitive, and taking reportedly to modify the whole thing. Thus the fuller forms are things like "reportedly, Thailand's Prime Minister [is] to buy [a] stake in Liverpool", or "reportedly, Pennington [is] to become [the] highest paid Jet".
Although both forms occur, they're not equally common:
| N reportedly to | N to reportedly | |
| Bush | 21 |
2 |
| Kerry | 0 |
0 |
| U.S. | 121 |
5 |
| China | 55 |
4 |
| France | 5 |
0 |
| U.K. | 24 |
1 |
| Canada | 0 |
1 |
| Germany | 3 |
0 |
| India | 32 |
1 |
| Japan | 49 |
1 |
| Clinton | 18 |
1 |
| Kobe | 11 |
0 |
| TOTAL | 339 (95.5%) |
16 (4.5%) |
Note that many of these examples are not in headlines, and in some of them the noun is not the subject of the infinitive, though an analogous question about scope generally arises:
Dr. Jianli had traveled to China to reportedly investigate large-scale labor unrest and to meet with labor activists.
There may be some effect of copy-editor prejudice against split infinitives here, but I think it's mostly the effect of the (genuine) scope problem, as we can see by comparing what happens to the same list of nouns with the adverb formally, which normally modifies the verb phrase rather than the whole sentence:
| N formally to | N to formally | |
| Bush | 11 |
142 |
| Kerry | 4 |
51 |
| U.S. | 149 |
2,440 |
| China | 7 |
87 |
| France | 7 |
31 |
| U.K. | 17 |
55 |
| Canada | 1 |
161 |
| Germany | 2 |
8 |
| India | 10 |
41 |
| Japan | 20 |
86 |
| Clinton | 3 |
21 |
| Kobe | 0 |
0 |
| TOTAL | 231 (6.9%) |
3123 (93.9%) |
This pretty much reverses the proportions found in the table for reportedly. Note that the table is polluted on the "formally to" side by examples like
The European Commission has decided to ask France formally to review its rules banning television advertising by the publishing and cinema.
where formally probably modifies ask rather than review. And in this table, an even smaller percentage of the examples are headlines of the [NP to V ... ] type, although that doesn't matter to the general point about scope.
The exact percentages would change a bit with more careful inspection, but the unchecked counts in these two tables give a strong indication that people are placing adverbs in a way that sensibly reflects their scope. When I was taken aback by the NYT headline "Prosecutors to Reportedly Drop Charges Against Bryant", the numbers were on my side.
I'm not sure what's going on with the minority cases of wide-scope "to reportedly" and narrow-scope "formally to". As Arnold Zwicky has discussed in a series of four posts here, there's a thin line between error and mere variation. In the case of "to reportedly", it's possible that we're not seeing either simple scope errors or the birth of a new scope dialect, but rather the erratic application of a genre-specific pattern of superfluous and/or misplaced modifiers, which Arnold discussed under the heading of Journalists' Alleged Hedges. And in the (very rare) cases of narrow-scope "formerly to" (such as the AP headline "Kerry formally to announce candidacy"), I suspect the influence of the pseudo-rule against splitting infinitives.
Posted by Mark Liberman at September 2, 2004 08:51 AM