Language Log

e e cummings and his iPod: Faith vs. WF again

I've been spinning out a series of postings on (several different kinds of) conflicts between faithfulness (Faith: roughly, stick to the original) and well-formedness (WF: roughly, make things fit your system).  Yet another case came up on the American Dialect Society mailing list back in December: must the first letter of a sentence always be capitalized?

Jim Smith asked on 12/14/06:


and Beverly Flanigan followed up with:


In this case, Faith says to preserve the details of expressions (including initial lower case), while WF says to make the spelling conform to a convention that demands initial upper case.


(Note that "well-formedness" here does not refer to some absolute sense of correctness, but only to conformity to some system -- a variety of language, a style sheet, whatever.)

I posted on Language Log a while back on some cases where points of mechanical style in writing are problematic when material is quoted: order of quotation marks and periods/commas, double vs. single quotes, general lowercasing vs. conventional capitalization, indications of emphasis, the serial comma.  Here, well-formedness generally trumps faithfulness; original schemes for such things are generally converted to the quoter's home scheme.

In many other details, WF will always win: except in very special circumstances, no one attempts to reproduce type fonts, line divisions, or many other details of the physical appearance of text.  Other details -- whether paragraphs are indented or flush left, whether a dash is indicated by one en-dash, two en-dashes in sequence, or one em-dash, and whether the interpolated material is solid with the surrounding text, or separated from it by a space, etc. -- are occasionally preserved, but usually not.

As for the serial comma in coordination, as far as I can tell, its use generally conforms to the quoter's preferred scheme, whether this is anti-serial, with no serial comma (the majority style), or serial (which is my own): anti-serialists quoting people generally remove the commas, while serialists put them in.  This is WF.  A notable exception to this is in titles, which are often quoted faithfully.  The New York Times, which is pretty relentlessly anti-serial, nevertheless seems generally to preserve the commas in titles (of books, in particular).

Now to initial capitalization.  There are (at least) three conventions at issue:


Expressions like "pH" and "iPod" present immediate challenges to (WF1a).  Unfaithful spellings like "PH" and "IPod" are unacceptable to many -- probably most -- people.  The question is the status of faithful spellings.  Some style manuals are resolute about (WF1a); faithful spellings are unacceptable, and so such expressions must be avoided as the first words of sentences  (you can, however, write "A pH..." or "The iPod...").  If both faithful and unfaithful spellings are unacceptable, we have a STALEMATE between Faith and WF, and the conflict must be avoided in one way or another.

The Chicago Manual of Style (15th ed.) allows unfaithful spellings but prefers avoidance: for "eBay", "iMac", and the like, it says (p. 366):


CMOS requires that things like

   r-lessness is widespread in the U.K.

be replaced by one of the following:

   R-lessness is widespread in the U.K.  [(WF1a) trumps Faith]

   The phenomenon of r-lessness is widespread in the U.K.  [avoidance]

Indeed, examples with (WF1a) trumping Faith are not hard to find.  From "Among Believers" by A. O. Scott, in the 9/11/05 NYT Magazine, p. 40, about a literary journal whose name is n+1:

 
But spellings in which Faith rules are also easy to find:


[Addendum 8/13/07: Marc Pelletier and Fernando Colina note a context where (WF1a) isn't really an option, and you have to use faithful spellings or avoid the issue: in writing about material in computer languages, since in most computer languages case is meaningful, so that "someFunction()" and "SomeFunction()" are not equivalent.]

If you take (WF1a) to be an inviolable constraint, then you're committed to (WF1b) as well, and you can't write things like the following:


Instead, you have to avoid the offending initial character in one way or another:


But, as we've seen, not everyone treats (WF1a) as inviolable, or at least as inviolable in all circumstances.  And some of these people have no problem with some or all of the examples that violate (WF1b).

On to personal names, like "e e cummings".  For these, (WF1a) and (WF2) are both in play, and there are four outcomes of the conflicts between them and Faith.  Since I happen to have collected some cites for "bell hooks" (and since, as many readers have now pointed out to me, the poet himself used conventional capitalization and punctuation in his name), I'll use her name (rather than cummings's/Cummings's) to illustrate the cases.


Winner-takes-all outcomes

Outcome 1.  (WF1a) and (WF2) are both inviolable; Faith loses everywhere.  We get things like the following:

  
plus sentence-internal occurrences of "Bell Hooks".

Outcome 2.  Faith wins over both (WF1a) and (WF2).  We get things like the following:

  
plus sentence-internal occurrences of "bell hooks".

Note the spelling in the url: http://en.wikipedia.org/wiki/Bell_hooks.  Wikipedia's software is committed to initial caps, no matter what those who maintain the site think; see Outcome 3 below.  As a result, the Wikipedia page is labeled "Bell hooks", with the entertaining warning:



Mixed outcomes

Another possibility is that Faith wins over (WF2) but not (WF1a), which is  inviolable.  There are two solutions:

Outcome 3.  (WF1a) straightforwardly wins over Faith: we get "Bell hooks" initially (as in the wiki page title below, and in most, but not all, of the rest of the page), but "bell hooks" internally (as in the rest of this wiki page).


On the ADS-L, Bill Mullins (12/14/06) brought up a case similar to "bell hooks":


(Mullins expressed considerable annoyance at people who go "to such great lengths to make their name flout normal conventions of spelling and capitalization".)  John Baker added that the woman in question opts for Outcome 3:


Outcome 4.  The conflict between (WF1a) and Faith is a stalemate: no sentence can begin with the woman's name, though "bell hooks" is fine internally.  This solution is not something you can search for, but I'm pretty sure that you could find people with this system.

I have no idea what the relative frequencies of these systems are, or even whether writers are consistent in the way they negotiate conflicts between Faith and WF in different cases.  (I suspect that I'm not consistent: I have no problem with sentence-initial "iPod" and the like, but I'm less happy with sentence-initial "bell hooks" and the like; and I'm not bothered by sentence-initial "(1)" or "/r/-lessness", but I tend to balk at sentence-initial "4,357".)  But it seems likely that all the logical possibilities are out there.

Having read this far, people often write me to ask which system of capitalization is the CORRECT one: what should they DO?  I generally refuse to be directive, on the grounds that whatever you do is likely to annoy someone or another.  As a general matter, conflicts between Faith and WF do not have clearly "correct" solutions -- because Faith and WF are both defensible principles.  The best you can do is use capitalization practices that satisfy you aesthetically and won't annoy too many people in your audience.  If you're writing for publication, of course, an editor or style sheet will probably make the decisions for you.

zwicky at-sign csli period stanford period edu

Posted by Arnold Zwicky at August 12, 2007 01:09 PM