Second sum: term

[Larry Trask]

----------------------------Original message----------------------------
Several weeks ago I asked for a term to denote the observation that a
linguistic change often results in simplification in one part of the
system but in a simultaneous complication in another part.  I
summarized the responses I received, and listed the proposed terms
(there is clearly no single term in regular use).  This query,
recall, was on behalf of the dictionary of historical linguistics I
was writing.
 
First, I inadvertently omitted Richard Janda from the list of those
who had responded; my apologies.  I received a couple more responses
to my original posting after I'd sent in my summary, and I've received
a number of further responses to my summary.  Here is the state of
play.
 
Most respondents seemed to favor the creation of a term for this
phenomenon, but not all: two people declared that they regarded the
observation as trivial and inconsequential and not worth naming.
One, however, took exactly the opposite view, seeing the lack of a
suitable term here as a serious gap in our terminology.
 
Many people expressed their opinions on the proposals included in my
first summary, but there was absolutely no consensus: a term that was
greatly preferred by one person was dismissed as unacceptable by
another.
 
A number of further proposals were put forward; here they are.
 
LOCAL OPTIMIZATION
ORGANIC COMPENSATION
SYSTEMIC COMPENSATION
LEVEL-BLINDNESS
ANTITELEOLOGY
SEE-SAW PRINCIPLE
SIDE-EFFECTS PRINCIPLE
 
One respondent particularly liked the idea of trying to borrow an
analogous term from evolutionary biology; I have been looking, but so
far I haven't uncovered such a term in biology (which is not to say
that none exists).
 
Another pointed out that the phenomenon in question is very familiar
in the field of systems engineering, though it appears to have no
recognized name there.
 
A third advised me to consult the writings of the Prague School
linguists, who were, of course, very much interested in phenomena of
this type, and especially Jakobson's celebrated paper on the history
of Russian vowels.  I dutifully did so, and found that Jakobson's
terminology, while certainly colorful, was decidedly warlike: he
speaks of "conflicts" and "struggles", and almost seems to see a
language as a battleground involving competing subsystems, each
struggling to get the upper hand.  But I could find no term for what
I was looking for.
 
There was also a suggestion that I consult the terminology of chess.
As it happens, I'm acquainted with that terminology, and I couldn't
find anything suitable.  I did note, somewhat wistfully, that the
wonderful chess term `Zugzwang' finds no linguistic use, but then I
can't imagine a language in Zugzwang: "OK; Kashubian is in great shape
today, but one phonological change and the whole language will fall
apart."
 
There were some comments on matters of policy.  A couple of
respondents thought it was entirely in order for my dictionary to
promulgate neologisms if these appeared to be useful, and even for me
to coin a neologism in the dictionary, if I thought that necessary.
Another respondent, however, urged caution here.  That respondent
considered it potentially dangerous and misleading to readers to
include neologisms, and urged me to include only terms with a
substantial history of published use, and to cite examples of use
from the literature.  I guess I'd better comment on this.
 
There are two classes of terms for which I am, in general, providing
no citations.  The first group consists of those terms which have been
in regular use for generations, such as `cognate' and `creole'.  These
are the terms which everybody uses freely and encounters regularly.
The second group consists of terms which are used in the older
literature but are now rare or obsolete.  Examples are things like
`proethnic' and `media aspirata', the first of which is obsolete and
the second of which is now generally confined to specialist handbooks.
Students reading the older literature will find these things used
without explanation, and will need to be able to look them up; I am
entering them, but marking them with a label like `rare' or `obsolete'
or `confined to older literature'.
 
Otherwise, my policy is to identify the first published source of a
term, whenever I can track that down, if the term is recent enough
that it cannot be safely regarded as established.  So, terms like
`exaptation', `metatypy' and `abrupt creolization' get explicit
citations.
 
Now: how far should I go in entering neologisms?  There is no simple
answer to this, because I can't predict the future of a recently
proposed term: maybe it will be taken up and flourish, or maybe it
will die on the pages of the article where it was proposed.
 
So I'm proceeding as follows.  At this stage, I am trying to include
every term which looks even potentially useful.  After writing my
definitions, I will certainly find that my typescript is too long -- I
have a strict length limit.  So then I'll prune the dictionary by
removing the terms that appear to me to be the most marginal.  This is
the only reasonable way of going about things: if I am ruthless to
begin with, and then find myself with unused space, my life will be
very difficult.
 
This length limit is a further good reason for omitting citations of
familiar terms.  Citations mean a corresponding absence of entries,
and, providing my definitions are good enough, I think we're better
off with more entries.
 
But *some* neologisms I am making a point of including.  For example,
`exaptation' has already gained some currency in the field, and
`metatypy' strikes me as an elegant and unambiguous term for a
concept which has been much discussed but which has previously been
given only a series of ad hoc and often cumbersome labels, such as
`extreme structural borrowing'.  Here I am using my judgement: in my
view, we need an established term for this, and `metatypy' is the
best I have seen, so I am going to promote it.
 
These policies doubtless won't please everybody, but they're the best
I can come up with.  Anyway, if you don't like them, I guess you're
free to write your own dictionary. ;-)
 
So, what's my decision on the missing term?  I am, as the objects of
major scandals are wont to say, considering my position.
 
My thanks to Richard Janda (again), Max Wheeler, Stefan Georg, Johanna
Nichols, Harold Koch, Rich Alderson, Eric Schiller, Richard Coates,
Mark Hale, Roger Wright, E. Bashir, Benji Wald, Phil Baldi, Paul
Lloyd, Donka Minkova, Stephen Schaufele, Robert Whiting, and Suzanne
Fleischmann.  I hope I haven't omitted anybody this time.
 
Larry Trask
COGS
University of Sussex
Brighton BN1 9QH
England
 
larryt at cogs.susx.ac.uk