Joe:
> > But introspecting (which I in my armchair always prefer to hard
> empirical
> > experimental psycholinguistic evidence...), less frequent irregulars are
> > often noticeably hard to remember, when there are rival candidates that
> > are equally plausible (but not equally correct).
> >
> >    SWING (I swang? I swung?)
> >    THRIVE (thrived? throve?)
> >    FIT (It fit me? It fitted me?)
> >
> > And for irregular plurals it also happens that one knows that the plural
> > is irregular but is unsure about what it is.
>
> Aha! So you're agreeing with me. Irregulars require more effort to
> access from memory.

I do agree, though my agreement is worthless (in proportion to my
ignorance).

My intuition is that the effort of accessing irregulars from memory is
proportional to their frequency, and that the effort of computing regular
forms comes someway midway along the scale of effort for accessing
irregulars.

> > BTW, I'd be interested to hear about whether these psycholinguistic
> > experiments also study plurals. Generalizing about (ir)regularity from
> > past tense forms alone is risky, because:
> >
> >   irregular plurals typically belong to infrequent words
> >   irregular pasts typically belong to frequent words
>
> Actually, louse~lice notwithstanding, most umlaut plurals tend to
> be pretty frequent: man~men, woman~women, foot~feet, tooth~teeth,
> mouse~mice, goose~geese (well, this last one not so much).
>
> Then, we have: child~children, which is very frequent; and we have
> a case of suppletion (person~people), which is also very frequent.
> An invariant animal plural such as fish~fish is pretty frequent
> and, where I come from, deer~deer is still used a lot.
>
> Othe native irregulars --ox~oxen and brother~brethren-- are pretty
> infrequent, however.

"Brethren" is defunct for everyone. I noticed a year or so ago that
for me the plural of OX is _oxes_, not _oxen_.

> Almost all the rest of irregular plurals are learned words, and
> therefore pretty infrequent.

Exactly. And for verbs, the picture is totally different. All irregulars
are native and nonlearned.

> So, what I'm trying to say is that the situation of nouns is not
> all that different from verbs.

I think it is. I think the facts I asserted before are valid. Also,
whereas the most frequent verbs are overwhelming irregular, this is
not so for nouns.

    frequent verbs are typically irregular
    frequent nouns are typically regular

    irregular verbs are typically frequent
    irregular nouns are typically (modal average) infrequent

--And.



%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%