[SGVLUG] OT our wonderful English language...

[Dustin Laurence]

On Fri, Feb 17, 2006 at 10:21:40AM -0800, Emerson, Tom wrote:
> > -----Original Message-----
> > Behalf Of Dustin Laurence
> > 
> > On Fri, Feb 17, 2006 at 08:57:35AM -0800, Emerson, Tom wrote:
> > > 
> > >    "... They conducted a nearly comprehensive survey of Greenland
> > 
> > This is the journal _Science_, so ... it means they're writing in academic
> > language, which is required to be more weasely-worded than most things.
> 
> Aye, there's the rub :)  They're FORCED to use an "oxy-moron" ;)

That's not an oxymoron; "nearly comprehensive" has a clear meaning.  It
would be an oxymoron if the words contradicted each other, such as a
"definitive initial survey."  If I had data on all but one glacier
outflow in all of Greenland, I think that would count as "nearly
comprehensive".  If was missing data on 40% of them, it wouldn't be any
kind of comprehensive, especially if (as would normally be the case) the
40% were not chosen statistically to minimize possible bias in the data.

> I suppose when "unqualified", the words "comprehensive study" could
> mean anything from "barely" to "completely", though most folks [lay
> people] would tend to think it more towards the "complete" end of the
> spectrum.

Nah, comprehensive means it has covered all relevant areas.  Are you
confusing it with "definitive" or "authoritative"?

Urk, dictionary.reference.com doesn't require completeness, which I
think of as part of the definition.

> ...However you point out that this is the Journal _Science_,
> where the readers would be unsatisfied with an unqualified claim, thus
> the writer has to choose some word to describe the level of
> completeness.  But what word to pick to describe the level of
> completeness?  Of course, few in the target audience would actually
> believe "completely", and in fact might attack the study out-of-hand
> if it did claim to be a complete study, so it has to be brought back a
> notch.

No need to try to invoke such shifty motives.  The readers of Science
will certainly accept an unqualified claim if it is sufficiently
supported.  The point was it was not quite *that* good, and so they said
so.

> > They are warning other researchers that there are minor points where
> > the study could be improved and they know it, however, they are
> > confident that the essential results will stand and that the results
> > are important enough to publish now (I and more importantly
> > _Science_ agree on that).
> 
> which is how I described what I think of when I hear of a "very
> comprehensive study" -- one in which further refinement of the
> techniques would not significantly change the results.

No, I don't think that's proper usage.  "Comprehensive" speaks to the
breadth of the study, in this case, that they had a nearly complete
picture of the ice flow over all of Greenland.  You are thinking of
something like authoritative or definitive.  A study could be
comprehensive but still be in question; perhaps all areas were studied
but the methodology wasn't the best.

It is perfectly meaningful to do an "initial comprehensive survey,"
which would mean that everything was looked at but not in great detail.
This would be appropriate, for example, in a paper that attempts to map
out which areas of a relatively unstudied field need the most immediate
attention by the research community.  Then typically would follow a
flurry of papers that look at individual areas in more detail,
eventually leading to either review papers which collect up many such
studies or bigger studies and eventually (the goal would be) to obtain a
picture which is both comprehensive and definitive.

Science is a group endeavor, but movies have made that mostly
incomprehensible to people. :-)

Also, I said they are *confident* that the results would stand.  To be
authoritative, that has to be pretty clear to everyone, and perhaps to
have withstood significant efforts to discredit it (or that nobody has
any reasonable idea how to do it).

> that too can be misconstrued -- is it so "very" comprehensive that to
> question the results would be deemed an insult?

If that is insulting, then the insulted person is not a scientist
(unless the questioner is an idiot, unable to accept the facts, or
resorting to ad hominem arguments).  Scientists don't get insulted when
their work is questioned intelligently, it's just that a good proportion
of the people who want to question it usually turn out not to even
understand the first thing about how science is done.  So they resort to
conspiracy theories about scientists behaving like movie mad scientists,
after which it's reasonable to take it personally. :-)

> are needed, making it a "very nearly comprehensive study", which, of
> course, is almost meaningless :)

That would be appropriate for, say, missing one month of data on the
smallest glacier outflow in Greenland over five years.  It isn't
*technically* complete, but it is "very nearly" so. :-)

Actually, if the data in question doesn't exist at all, I suppose you
could argue that no qualifier is needed; there is some ambiguity over
whether completeness would mean all the data that exists or all the data
that would have been needed to cover the problem. :-)

Dustin