According to this abstract, published in 2007, "the half-life of an irregular verb scales as the square root of its usage frequency: a verb that is 100 times less frequent regularizes 10 times as fast." The data that supports this study comes from 177 English irregular verbs, traced over a time span of 1200 years of evolution.

How much is known about verb regularization rates in other languages? Does the data collected so far, for the studied languages, show the same correlation?

  • 5
    not that it answers your question (yet!) but as part of my PhD I'm looking at this in Ancient Greek. Commented Oct 2, 2011 at 18:04
  • Great! Looking forward to it :-) Commented Oct 2, 2011 at 22:12
  • It's going to be a hard one to answer given that you're going to need a 1000 year written history of the language to do that - definitely narrows down the number of languages you can study this phenomenon in!
    – LaurenG
    Commented Nov 3, 2011 at 11:39
  • @LaurenG Fair enough. I edited the question to make it less demanding. Commented Nov 3, 2011 at 12:53
  • 2
    @James: That is very interesting! Which dialect(s) are you looking at? Have you found anything interesting yet? I have noticed this particularly in the aorist, where often two stems are possible: an (irregular) thematic stem, and a regularized sigmatic one. Cf. eipon / eleksa, and I believe the passive aorist has several options too (errêthên / elekhthên / elegên ?), which might show that the pseudo-sigmatic aorist is younger than the suppletive forms but older than the regular passive aorist. Or something.
    – Cerberus
    Commented Nov 3, 2011 at 13:53

1 Answer 1


German is a language which could probably match English as far as the historical record is concerned, and it is highly likely that studies on regularization rates have been conducted, though I don't have any highly specific references. A good place to start might be:

Fertig, David. Morphological Change Up Close. Two and a Half Centuries of Verbal Inflection in Nuremberg. 2000

Note, however, that frequency is not the only variable predictive of regularization rate. Another factor identified for English is phonological similarity: i.e. verbs like sing ~ sang will regularize more slowly because of the presence of phonologically similar alternations with other irregular verbs, e.g. ring ~ rang. A reference on this is:

Hare and Elman (1995), "Learning and Morphological change" Cognition 56, 61--98

Not the answer you're looking for? Browse other questions tagged or ask your own question.