Re: Re: stridden

[Arnold M. Zwicky]

On Mar 11, 2004, at 4:30 PM, i noted that:

> ... some analogies are
> hugely more likely than others.  the analogies normally tap into the
> morphological regularities of the language.

and went to consider the past participle of a verb with base form
"stride" and past tense form "strode":

> ,,,once you've registered "strode" as the past tense form, there are
> then
> four possible analogies giving a past participle:
>
> 1.  pple = past + n: stroden (cf. weave - wove - woven)
> 2.  pple = pres + n: striden (cf. take - took - taken)
> 3.  pple = ablaut past: strode (cf. find - found - found)
> 4.  pple = special stem + n: stridden (cf. drive - drove - driven)
>
> the first three of these have relatively few exemplars, and no exemplar
> with /ay/ in the present and /o/ in the past (like "stride" -
> "strode").  only the fourth has any legs, and it rests on just four
> verbs with any great frequency...

this last bit is significantly wrong; i should have listened to my own
observation (above) about the importance of morphological regularities.
  the fact is that though item-by-item analogy does occur, a  great many
of the forms described as "analogical" are in fact just regularizations
-- the extension of morphological regularities to new items.

one of the great regularities in english verb forms is that the past
tense and past participle are usually identical; this is true for all
completely regular verbs and for a great many that exhibit various
subregularities or irregularities.  it follows that once you've
registered "strode" as the past tense, the expectation will be that the
past participle is also "strode".  so analogy #3 above should have *a
lot* going for it, even if there are no items with this pattern that
are phonologically similar to "stride".  (phonological similarity is
what gives some life to the item-by-item analogy in #4.)

analogy #3 is the one favored by michael quinion and john baker in
recent postings to the list, and they provide examples from printed
works and google searches to show that this option has often been
taken, indeed more often taken than #4 -- results that really shouldn't
be surprising, given the past/pastparticiple generalization.

(quinion notes that fowler and garner both recommend "stridden".  this
is not exactly a  surprise, since prescriptive grammarians have a
tendency to value irregularities over regular forms, probably on the
belief that irregular forms are usually older -- this is by no means
always the case, by the way -- and that the more regular forms are
*simplifications* of the historical system, simplifications resulting
from the failure of speakers to learn all the complexities of their
language.  that is, they tend to see regular forms as errors of
ignorance.)

baker mentions also the possibility of "strid".  the only way i can see
for past participle "strid" to arise is if the *past* of "stride" has
been reshaped to "strid" (perhaps on analogy to "hide" - "hid"); then,
by the past/pastparticiple generalization, the past participle should
follow along (or be "stridden", continuing the analogy with "hide").  i
believe that past "strid" is attested (i'm away from my sources right
now).  i *know* that past "chid" for "chide" is attested (in john
newton's 1779 text for a hymn in the Sacred Harp shapenote book i sing
from).  i suspect that there are even attestations for "rid" as the
past of "ride" and "driv" as the past of "drive", and perhaps others of
this very minor pattern.  you'd expect such occurrences to be rare, of
course -- and past participles showing this pattern to be even rarer,
since past participle forms are a lot less frequent than past forms,
period.

but let me emphasize again that this discussion isn't about what the
past participle of "stride" *ought to be*.  all of the following
patterns are certainly attested:
   stride - strided - strided
   stride - strode - strode
   stride - strode - stridden
and other patterns might also be attested.  if you've learned that the
past tense form is "strode", then essentially you're faced with a
competition between "strode" (which follows the past/pastparticiple
generalization) and "stridden" (which is supported by items
phonologically similar to "stride").  in the absence of clear models
from other speakers -- the verb is rare, past participles are rare, and
there is variation in the few instances you might hear -- you might
well just avoid the form entirely.  which is what lots of people do.
if you settle on "strode", that's fine.  if you settle on "stridden",
that's fine too.  maybe some day one or the other of these forms will
catch on and spread throughout parts of the english-speaking world, but
that hasn't happened yet.  (maybe "stride" will get completely
regularized.  that sort of thing has been known to happen.)  at the
moment, there are (at least) two alternatives, and many people are
comfortable with neither of them.

arnold (zwicky at csli.stanford.edu)